Steam Market Research 6

[sealed content withheld: memory]

Script completed
Wall time 0.1 seconds
Output:

Warning: truncated output (original token count: 23661)
Total output lines: 3

{"status":"fulfilled","value":{"chunk_id":"7c5380","wall_time_seconds":0.000002556,"exit_code":0,"original_token_count":16059,"output":"Warning: truncated output (original token count: 16059)\nTotal output lines: 220\n\nTitle: Steam Market Research 5\nDate: 2026-09-06 00:56:42\n\n# Steam Market Research 5\n\n## Summary\n\nSteam Market Research 5 rebuilt a current Steam research dataset and investigated the market first broadly, then specifically for games advertising build expression. The owner wanted fresh findings from Steam data, independent of earlier Fable research and Steam marketing blogs, without treating money, audience growth, or current projects as the research’s objective. Python collectors used Steam’s public StoreQuery and StoreBrowse services to produce a September 5, 2026 snapshot of 184,664 game apps across 23 storefront catalogs, with SQLite, Parquet, raw responses, and field inventories. An excessive image download was stopped and deleted at the owner’s request; only media references remain. Statistical screens, description classification, and a bounded Steam-review sample identified recurring associations, rejected misleading niches, and developed a hypothesis about meaningful item interactions. The session ended with reproducible artifacts for continuation in another task.\n\n**Research intent, source boundaries, and corrections to the plan:**\n\n- The owner had previously followed Steam closely and collected catalog data to answer questions with scripts, but wanted to refresh his understanding after roughly five years. He asked the assistant to generate most research directions and useful answers, including openings that other developers might not have noticed.\n- Earlier Fable Steam research, its logs and summaries, and all Steam marketing research blogs were explicitly excluded. The desired relationship to existing marketing commentary was “simply indifferent,” rather than opposition or contrarian reaction. No earlier research outputs or third-party sales estimates were imported into the dataset.\n- The owner initially described an affinity for making a game in a few months and marketing it through minimal-effort actions with substantial possible reach, while remaining open to changing that approach. This was context for research, not a requirement to prove that any observed game was made quickly or marketed cheaply.\n- The assistant initially asked for revenue/audience goals, risk tolerance, acceptable promotional actions, and whether the work should inform Horse Game. The owner corrected this: “I'm not particularly after money or audience,” said his writing already explained his positions, and requested broad Steam ideas and trends independent of anything he was currently making. He allowed closer examination of games resembling his interests without making them the entire initial population.\n- Reading the owner’s public writing clarified several relevant preferences: short development cycles as design exploration and skill development; small production scope compatible with spacious play; creative contribution through recombination; meaningful player choices; and making work without requiring a particular external response. These ideas supplied context, not additional measurable Steam variables.\n- The assistant disclosed a source-boundary mistake: the owner’s “Game quality is all that matters” essay reproduced a howtomarketagame passage, which was encountered while reading the essay. The external site was not opened, but complete non-exposure could no longer be claimed. Subsequent screening avoided embedded third-party material, and market claims were grounded in the newly collected data.\n- A second correction concerned feasibility. The assistant proposed studying category formation, novelty, mechanical depth, and discovery pathways through extensive game interpretation. The owner objected that these could not reliably be recovered programmatically from Steam data, and the assistant could not play all the games. The accepted plan instead began with obtainable fields: releases, tags, tag combinations, prices, review distributions, concentration, languages, and cohort comparisons.\n- The resulting sequence was to audit access and coverage, collect a broad inexpensive metadata layer, analyze its distributions, and use numerical anomalies to motivate narrower description or review investigations. Development time, marketing effort, design depth, and causal discovery mechanisms were removed from the quantitative agenda unless separately measured.\n\n**Steam source audit and collection architecture:**\n\n- The owner authorized the assistant to choose collection methods, audits, and implementation details, requesting the most representative and current dataset possible as of September 5, 2026, with as much per-game data as practical. He initially also requested store pictures for potential later visual analysis.\n- Live probes found that the old `ISteamApps/GetAppList/v2` route returned 404 and the newer `IStoreService/GetAppList` route required a key. Steam search worked but capped returned rows at 100 even when more were requested. A raw SteamSpy endpoint was tested, but SteamSpy data and estimates were not incorporated into the final game dataset.\n- The useful route was Steam’s public `IStoreQueryService/Query` for catalog enumeration and `IStoreBrowseService/GetItems` for detailed batched metadata. Protocol definitions were inspected through a Steam protocol-schema mirror. Full responses could be collected in batches of 1,000 rather than fetching every store page individually.\n- The implementation retained compressed raw response envelopes with exact requests, source country, retrieval time, and response metadata before ingestion. SQLite stored a preferred per-app record and separate relations for tags, creators, languages, categories, regional observations, discovery membership, media references, and aliases. Portable Parquet exports preserved full nested metadata in `data_json`; a compact compressed CSV exposed normalized game columns.\n- Five primary storefronts supplied detailed regional observations: US, Brazil, Japan, China, and Germany. Lightweight ID enumeration expanded to 18 more: GB, FR, CA, AU, KR, TW, HK, RU, UA, TR, IN, SG, MX, AR, TH, ID, VN, and ZA. Regional coverage added 208 discovered IDs beyond the US structured catalog.\n- A separate Steam-search enumeration was a cross-check, not the foundation of the completeness claim. Its 354 checked extra IDs were DLC, music, or other non-game records. It remained partial after rate limiting and was not silently described as complete.\n\n**Image-download mistake, pause, and deletion:**\n\n- The assistant interpreted the initial picture request too expansively, building an image downloader and fetching store artwork, screenshots, and description images across the catalog. It switched to Steam’s smaller screenshot variants, retained original references, and later bounded large animated description images, but the operation had already become excessive.\n- The owner interrupted after discussing the load in a side chat, asking about bulk downloads and collection status. All network collection was paused. At that point, 533,716 verified image files occupied approximately 144.8 GB and covered 41,756 games. The cache was biased toward earlier app IDs; its size did not make it a representative visual sample.\n- The assistant explicitly acknowledged overshooting. Steam search had returned HTTP 429 responses, and continued retries on that redundant crawl were unnecessary. A proposed replacement was comprehensive metadata and URLs with bounded question-specific samples, but the owner then gave the stronger instruction: “Delete the images folder, delete all image downloading machinery.”\n- Both image directories, `images.py`, its compiled cache, download logs/status/probes, and database download queues and cached-file bookkeeping were removed. The database was compacted. Lightweight source URLs and game-to-image associations were retained because they are metadata, not downloaded images.\n- The assistant answered the owner’s completeness question directly: metadata was not finished at the pause. The then-current count of 184,559 game records and 185,170 candidate IDs was intermediate and was superseded by the completed snapshot below.\n- On the owner’s subsequent instruction to finish game-data collection, metadata collection resumed with pacing and stop-on-429 behavior. Image and other asset downloading remained removed. The finished source dataset contains no store-image, trailer, audio, or game binaries. Later PNG/SVG artifacts are locally generated statistical charts, not renewed store-image downloads.\n\n**Completed catalog coverage and normalization repairs:**\n\n- The completed source snapshot contains 184,664 game apps: 128,972 marked released and 55,692 upcoming. Another 503 records are separated: 148 store-listed mods and 355 other app records. The SQLite master table named `games` contains all app types, so game queries require `type=0`; `exports/games.parquet` is already game-only.\n- All 185,173 discovered IDs were accounted for through records or redirects. All 23 structured storefront catalogs were enumerated, and their collected unique-ID counts matched the maximum totals Steam reported during collection. Small increases in several storefront totals were captured.\n- These are app identities, not a manual deduplication of titles or editions. The scope is the public catalogs checked, not every app ever registered with Steam. Completely delisted, unpublished, unlisted, or exclusively available outside those storefronts may be absent.\n- Preferred game metadata was fetched from September 5, 2026, 19:09:38 through 20:53:33 UTC. This is a collection window across a changing service, not an atomic snapshot. Retrieval provenance and coverage ranges remain in the artifacts.\n- An initially overstrict ordering check rejected final pages because Steam appends mods after games, each ordered within its type. The raw pages were retained and replayed after correcting the invariant; a numeric decrease at the type boundary was not missing data.\n- Seven regional or legacy IDs redirected to other canonical app IDs. Treating these as ordinary records could both create false gaps and overwrite canonical/regional statistics with alias responses. Explicit alias mappings were added; 25 affected canonical-country observations were restored, eight invalid regional observations removed, and no alias payload remained in the canonical records. The final alias table contains 33 country mappings for seven redirecting IDs.\n- Validation checked SQLite integrity, unique identities, embedded canonical IDs, alias handling, tag counts and definitions, relationship integrity, review-count and percentage ranges, and removal of the image directories/downloader/queue. Parquet row counts were verified after writing. The exported tables together occupied about 744 MB; the full source directory, including SQLite and raw responses, was about 8.6 GB at completion.\n\n**Exactly what data exists for games:**\n\n- All 184,664 game records have identity/status fields and filtered review summaries, including zero-review games. Identity includes app ID, name, store URL/slug, type, visibility, and normalized free/upcoming/Early Access flags; 18 names were blank. Missing false-valued protocol flags use their defined defaults, while absent optional values remain missing.\n- Both description locations were collected. `basic_info.short_description` represents the short sidebar description and exists for 183,399 games. `full_description_bbcode` represents the full “About This Game” content and exists for 184,536 games, retaining formatting and embedded references. The full description is in `data_json` in SQLite and Parquet; the compact CSV contains only the short description. The owner specifically asked and received this clarification.\n- Tags are present for 184,621 games, covering 430 distinct labels, with IDs, names, Steam weights, and rank. Despite requesting more, the service returned up to 20 tags. Current tags are not historical launch tags.\n- Creator arrays include developers for 184,417 games, publishers for 183,942, and franchises for 36,967, with creator-clan IDs where supplied. These identify reported creators, not verified corporate ownership or team size.\n- Release information includes 143,967 Steam timestamps, 16,045 original-release timestamps, 7,536 original-Steam timestamps, and 774 Early Access exit timestamps, plus planned-release text/flags and platform dates where supplied. A planned timestamp does not establish that a game launched, and asset modification timestamps are not game update dates.\n- Review summaries contain count, positive percentage, score, and label. Separate English-specific summaries exist for only 5,211 games, and unfiltered summaries for only 158. Their absence is not zero; the dataset cannot provide a full language-demand breakdown from these fields.\n- Purchase options exist for 110,402 games, and a best option for 110,092. They include package/bundle IDs, current/list price strings, formatted prices, discounts, editions, gifting/grouping details, and included-game counts. These are current regional observations; bundles and add-ons require care before interpreting them as a standalone entry price.\n- Platform objects exist for every game, including Windows/macOS/Linux flags, VR information, and device compatibility codes, which may mean Unknown. Nonempty feature categories exist for 184,660 games. Language arrays exist for 184,559 games, of which 184,482 report at least one supported language; support, full-audio, and subtitle flags are separate.\n- Content-descriptor lists exist for 43,085 games, rating objects for 8,661, related-app objects for 38,507, and external links for 87,871. Related objects may include demos, standalone demos, playtests, and parents, but are not a complete historical demo record.\n- Media metadata includes artwork objects for 184,654 games, nonempty screenshot galleries for 184,354, and trailer objects for 168,253. Screenshot ordering/content grouping, filenames, URL patterns, overridden/original artwork references, and trailer formats are retained without media binaries. Game records link to 2,967,715 unique image references.\n- The exhaustive inventory contains 297 observed JSON paths, including nested/container paths, with types and presence/nonempty counts. Not collected catalog-wide: individual reviews, review histories, reviewer playtime, owners, unit sales, revenue, wishlists, concurrent players, historical prices, development duration, team size, marketing effort, achievement lists/counts, or system requirements. A later bounded review sample supplements only five games.\n\n**First broad analysis: population, comparisons, and safeguards:**\n\n- After accepting the dataset, the owner authorized the first quantitative pass. Its main pool comprised 38,494 currently paid games outside Steam’s explicit-sex descriptors 3 and 4, in 2023–2025 store-date cohorts. This was not a manual quality filter. Whole-catalog supply tables retained free and explicit-content segments separately.\n- Of 128,972 released records, 128,878 had usable nonfuture dates; 93 lacked a usable date and one had a future date. These exceptions remained in the source but were excluded from dated analyses. `original_steam_release_date` took precedence where supplied, otherwise `steam_release_date`; incomplete Early Access history still limits interpretation of first exposure.\n- The main pool’s median was seven filtered reviews; 15.8% reached 100 reviews and 4.4% reached 1,000. Thresholds describe response scale, not financial success. Current lifetime totals cannot reconstruct launch performance, and older delisted games may be missing.\n- Broad tag tables used top-20 membership, with specificity checks; pair analysis required both tags within the top ten. The discovery screen evaluated 1,765 eligible pairs, froze 30 selections using 2023–2024 games, then inspected 2025 results. All 30 outcomes were retained, including weak ones. This separate-game comparison is not a historical forecasting backtest, and overlapping pairs are not independent replications.\n- Descriptive observed/expected comparisons standardized for release quarter and, in sensitivity checks, current USD entry-price tiers. Sparse strata fell back to broader date comparisons. These ratios are associations, not causal pricing or genre effects. Chart intervals were not corrections for exploratory selection or creator clustering.\n- Price extraction initially chose the cheapest among purchase options, which could select an add-on. It was corrected to Steam’s best non-bundle single-game package, excluding offers dependent on another owned app, using list/original price when available. This changed 2,101 records without changing the frozen pair selection.\n- Developer/publisher grouping used clan IDs with normalized-name fallback. Missing clan keys were filled only where a normalized name mapped to exactly one observed clan within its role; ambiguous names stayed unresolved. Largest-review and most-prolific creator removals were distinct sensitivity checks, alongside removal of the highest-review games.\n\n**Broad market expansion and differences hidden by tags:**\n\n- The paid/non-explicit catalog contained 5,643 games in the 2019 cohort, 9,771 in 2023, and 15,674 in 2025. Their current counts above 100 reviews were 1,243, 1,672, and 2,277 respectively. Thus the larger catalog includes more substantially reviewed games as well as a larger low-response tail; 8,811 of the 2025 games had fewer than ten reviews.\n- The whole-catalog January–August supply comparison counted 17,709 releases dated in 2026 versus 12,925 in 2025. This measured surviving public-catalog supply, not changing probability of launch success.\n- Within the 2023–2025 paid/non-explicit pool, Action Roguelike had 2,966 games, median nine reviews, and 17.1% above 100; Auto Battler had 657, median 14, and 22.4%; Roguelike Deckbuilder had 412, median 40.5, and 36.9%. Fast growth in a tag’s share of releases did not necessarily imply exceptional response per game.\n- Incremental had 1,865 games, median seven reviews, and 14.2% above 100; Idler had 1,455, median 12, and 19.0%. Their position improved relative to similarly priced/date-grouped peers: approximately 1.52 and 2.06 times the expected 100-review counts respectively. This comparison did not show that lowering a price causes improvement.\n- Tags themselves can mislead: several filmed romance games were tagged Immersive Sim, so an aggregate for that label could not be read simply as demand for a Deus Ex-like design.\n\n**Desktop/background companions and filmed visual novels:**\n\n- An exploratory description screen broadened the small Desktop Companion tag to promises involving productivity/focus tools, Pomodoro timers, or games occupying a screen corner while the user works or does something else. Among 2025 idlers, 42/130 companion-positioned games reached 100 reviews (32.3%, median 23), versus 117/650 other idlers (18.0%, median nine).\n- Sixteen companion matches exceeded 1,000 reviews. Similar quarter/price peers predicted approximately 25 games above 100; 42 were observed. Removing the largest developer or publisher left 31.8%. Examples included Tiny Pasture, Fantasy Map Simulator, Ropuka’s Idle Island, Tiny Aquarium, and Cornerpond. Taskbarn and Strange Shores illustrated that similar positioning could still receive little response.\n- The creative question proposed was what other ongoing systems people might enjoy alongside another activity. This was explicitly exploratory: the description screen was devised after seeing initial results and had no untouched validation sample.\n- Visual No…13661 tokens truncated…research/analysis-2026-09-05/group_action_roguelike.csv\nsteam-research/analysis-2026-09-05/group_auto_battler.csv\nsteam-research/analysis-2026-09-05/group_roguelike_deckbuilder.csv\nsteam-research/analysis-2026-09-05/group_hidden_object_cats.csv\nsteam-research/build-expression-2026-09-05/review_samples/4364910_negative.json\nsteam-research/build-expression-2026-09-05/review_samples/4364910_positive.json\nsteam-research/build-expression-2026-09-05/review_samples/4423580_negative.json\nsteam-research/build-expression-2026-09-05/review_samples/4423580_positive.json\nsteam-research/build-expression-2026-09-05/review_samples/2244210_negative.json\nsteam-research/build-expression-2026-09-05/review_samples/2244210_positive.json\nsteam-research/build-expression-2026-09-05/review_samples/2001760_negative.json\nsteam-research/build-expression-2026-09-05/review_samples/2001760_positive.json\nsteam-research/build-expression-2026-09-05/review_samples/3767850_negative.json\nsteam-research/build-expression-2026-09-05/review_samples/3767850_positive.json\nsteam-research/analysis-2026-09-05/group_comedy_gambling.csv\nsteam-research/analysis-2026-09-05/group_driving_trading.csv\nsteam-research/analysis-2026-09-05/group_horror_typing.csv\nsteam-research/analysis-2026-09-05/group_minimalist_tower_defense.csv\nsteam-research/analysis-2026-09-05/group_minimalist_resource_management.csv\nsteam-research/analysis-2026-09-05/group_thriller_walking_simulator.csv\nsteam-research/analysis-2026-09-05/group_hand_drawn_cats.csv\nsteam-research/analysis-2026-09-05/category_jan_aug_supply.csv\nsteam-research/analysis-2026-09-05/price_groups.csv\nsteam-research/analysis-2026-09-05/deep_groups.csv\nsteam-research/analysis-2026-09-05/deep_groups.json\nsteam-research/analysis-2026-09-05/simulator_title_games.csv\nsteam-research/analysis-2026-09-05/simulator_title_check.json\nsteam-research/analysis-2026-09-05/price_extraction_corrections.csv\nsteam-research/analysis-2026-09-05/discovery_selection.sha256\nsteam-research/2026-09-05/audit/applist_v2.json\nsteam-research/2026-09-05/audit/applist_v2.body\nsteam-research/analysis-2026-09-05/job_horror_exploratory.csv\nsteam-research/2026-09-05/audit/catalog_xml.body\nsteam-research/2026-09-05/audit/catalog_xml.json\nsteam-research/analysis-2026-09-05/findings.md\nsteam-research/analysis-2026-09-05/README.md\nsteam-research/analysis-2026-09-05/validation.json\nsteam-research/analysis-2026-09-05/validate_analysis.py\nsteam-research/analysis-2026-09-05/price_comparison.svg\nsteam-research/analysis-2026-09-05/price_comparison.png\nsteam-research/analysis-2026-09-05/pair_comparison.svg\nsteam-research/analysis-2026-09-05/pair_comparison.png\nsteam-research/analysis-2026-09-05/charts.py\nsteam-researc…254152 tokens truncated…ib/python3.14/site-packages/pandas/tests/io/json/test_pandas.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/test_normalize.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/test_to_markdown.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/test_to_latex.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/test_json_table_schema_ext_dtype.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/test_to_html.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/test_readlines.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/test_to_excel.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/test_to_string.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/test_ujson.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/test_round_trip.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/test_retain_attributes.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/test_read.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/test_pytables_missing.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/test_file_buffer_url.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/test_put.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/test_decimal.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/test_data_list.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/test_chunksize.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/test_common_basic.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__init__.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/test_upcast.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/test_subclass.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/test_select.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/test_store.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/test_time_series.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/test_timezones.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/test_inf.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/test_index.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/test_float.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/test_iterator.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/test_read_errors.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/test_ints.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/test_read_errors.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/test_iterator.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/test_ints.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/test_inf.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/test_index.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/test_float.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/test_file_buffer_url.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/test_decimal.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/test_data_list.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/test_common_basic.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/test_chunksize.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/common/__pycache__/__init__.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_to_string.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_to_markdown.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_to_latex.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_to_html.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_matplotlib.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_to_excel.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_html.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_to_csv.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_highlight.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_printing.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_format.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_exceptions.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_ipython_compat.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_bar.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_format.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__init__.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_eng_formatting.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_to_latex.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_css.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_style.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/test_console.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_non_unique.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/__pycache__/__init__.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_to_typst.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_to_string.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/test_tooltip.py\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/__init__.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/__init__.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_timezones.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_time_series.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_subclass.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_store.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_select.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_round_trip.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_retain_attributes.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_read.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_pytables_missing.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_put.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_keys.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_file_handling.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_errors.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_complex.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_compat.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_categorical.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/test_append.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/conftest.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/pytables/__pycache__/common.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_index_col.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_header.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_encoding.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_dialect.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_converters.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_concatenate_chunks.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_compression.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_comment.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_c_parser_only.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/conftest.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/usecols/__pycache__/__init__.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_python_parser_only.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_parse_dates.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_network.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_na_values.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_multi_thread.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_mangle_dupes.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/usecols/__pycache__/test_parse_dates.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_skiprows.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_read_fwf.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_quoting.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/usecols/__pycache__/test_strings.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/dtypes/__pycache__/__init__.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_unsupported.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_textreader.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/__pycache__/test_upcast.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/usecols/__pycache__/test_usecols_basic.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/dtypes/__pycache__/test_categorical.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/dtypes/__pycache__/test_dtypes_basic.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/parser/dtypes/__pycache__/test_empty.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/__pycache__/test_ujson.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/__pycache__/test_readlines.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/__pycache__/test_pandas.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/__pycache__/test_normalize.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/__pycache__/test_json_table_schema_ext_dtype.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/__pycache__/test_json_table_schema.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/__pycache__/test_deprecated_kwargs.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/__pycache__/test_compression.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/__pycache__/conftest.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/json/__pycache__/__init__.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_matplotlib.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_html.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_highlight.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_tooltip.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_format.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_exceptions.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_to_typst.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_bar.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_to_string.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/__init__.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_to_latex.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_style.cpython-314.pyc\nsteam-research/.venv/lib/python3.14/site-packages/pandas/tests/io/formats/style/__pycache__/test_non_unique.cpython-314.pyc\n"}}
{"status":"fulfilled","value":{"chunk_id":"420e63","wall_time_seconds":0.000001856,"exit_code":0,"original_token_count":229,"output":"total 72\ndrwxr-xr-x 1 adn adn   408 Sep  5 20:26 .\ndrwxr-xr-x 1 adn adn   198 Sep  5 16:01 ..\ndrwxr-xr-x 1 adn adn    76 Sep  5 16:28 .venv\ndrwxr-xr-x 1 adn adn  1214 Sep  5 18:15 2026-09-05\n-rw-r--r-- 1 adn adn  8105 Sep  5 18:06 README.md\ndrwxr-xr-x 1 adn adn   252 Sep  5 17:48 __pycache__\ndrwxr-xr-x 1 adn adn  4476 Sep  5 19:03 analysis-2026-09-05\ndrwxr-xr-x 1 adn adn  5718 Sep  5 20:44 build-expression-2026-09-05\n-rw-r--r-- 1 adn adn 16575 Sep  5 17:44 collect.py\n-rw-r--r-- 1 adn adn  2626 Sep  5 17:51 export.py\n-rw-r--r-- 1 adn adn  3312 Sep  5 17:48 finish_metadata.py\n-rw-r--r-- 1 adn adn  4318 Sep  5 17:58 inventory.py\n-rw-r--r-- 1 adn adn   194 Sep  5 20:26 network_policy.json\n-rw-r--r-- 1 adn adn  6391 Sep  5 17:46 reconcile.py\n-rw-r--r-- 1 adn adn  3561 Sep  5 17:32 repair_aliases.py\n-rw-r--r-- 1 adn adn  3536 Sep  5 16:16 search_check.py\n-rw-r--r-- 1 adn adn  2623 Sep  5 17:34 validate.py\n"}}

Script completed
Wall time 0.1 seconds
Output:

Warning: truncated output (original token count: 10764)
Total output lines: 4

{"status":"fulfilled","value":{"chunk_id":"7278b4","wall_time_seconds":0.000003639,"exit_code":0,"original_token_count":11,"output":"4:# Steam Market Research 5\n6:## Summary\n"}}
{"status":"fulfilled","value":{"chunk_id":"562090","wall_time_seconds":0.000002473,"exit_code":0,"original_token_count":4123,"output":"- The agreed follow-up therefore separated full-year 2025, January–June 2026, and July–August 2026, comparing families within each window and standardizing for release timing and current price. January–August supply comparisons used equivalent calendar periods. No snapshot-only method could reconstruct a 2025 game’s review total at the same age as a 2026 game.\n- The conceptual questions concerned what is configured (character, equipment, deck, party, army, machine), how alternatives differ (interactions, class combinations, modifiers, skill trees, composition), and when decisions occur (pre-run, during a run, between encounters, persistent progression). Not every one of these axes proved recoverable reliably from text.\n\n**Build-expression classification and methodological limitations:**\n\n- The classifier operated on stored short/full descriptions and tags, identifying explicit build variety, combinations/synergies, loadouts/playstyles, deck construction, class branching, party composition, and related object/quantity claims. It retained strict matches, a broader inclusive definition, and unconfirmed genre candidates, with excerpts explaining matches.\n- This measured advertised configuration opportunities, not verified gameplay depth, viable-build counts, balance, or enjoyment. Quantity claims alone were weaker evidence. Automatic combat did not disqualify an auto battler; central idler/incremental/FMVs and central visual novels were excluded, with exceptions for relevant game hybrids according to the saved rules.\n- The initial audit hid review counts and prices. Of 36 strict matches inspected, 31 clearly advertised configuration and five were ambiguous. Inclusive and unconfirmed examples helped expose false negatives and unclear boundaries. This small audit was a diagnostic, not a certified accuracy estimate.\n- False positives included “interactive items,” genre/control combinations, individual tactical actions, fixed progression, and ordinary storage. Rules were refined; 64 further centrally marketed idle/incremental records were removed after examining short pitches. A diagnostic full-description idle signal was later repaired without changing the substantive membership logic.\n- Family labels overlap. Top-five genre membership was too restrictive for some recognizable games: Slay the Spire 2’s Roguelike Deckbuilder tag was sixth in the snapshot. Final central family comparisons used top-20 tag membership, with top-five/top-ten checks retained rather than silently counting those definitions interchangeably.\n- Comparisons used review thresholds, positivity, medians, concentration, publisher removals, known-history exclusions, direct textual signals, inclusive/strict definitions, tag-rank changes, current-price tiers, and description-length checks. All remain observational; current store prose/tags can change after success and English-oriented text matching can miss relevant games.\n\n**Build-expression supply and final 2025–2026 family results:**\n\n- Comparing January–August cohorts after the focal exclusions, eligible paid games rose from 9,140 in 2025 to 12,409 in 2026 (+36%). Strict build-expression matches rose from 843 to 1,717 (+104%), increasing their share from 9.2% to 13.8%. The broader definition also nearly doubled. This is growth in advertised build expression, potentially reflecting store-language changes as well as design changes.\n- There were already 146 strict January–August 2026 matches with at least 100 reviews and at least 80% positive, versus 109 in the corresponding 2025 cohort. Different population sizes and exposure times prevent interpreting these counts as improving success rates.\n- The combined positive-response benchmark was introduced because attention and satisfaction diverged. For full-year 2025, January–June 2026, and July–August 2026 respectively, all strict focal build matches numbered 1,365, 1,085, and 632; 181, 101, and 45 met the combined benchmark, or 13.3%, 9.3%, and 7.1%.\n- Within those same three windows, top-20 Roguelike Deckbuilder matches met the benchmark at 41/169 (24.3%), 22/132 (16.7%), and 10/67 (14.9%). Action Roguelike was 58/449 (12.9%), 29/361 (8.0%), and 12/204 (5.9%). Auto Battler was 20/121 (16.5%), 9/90 (10.0%), and 5/46 (10.9%). Tower Defense was 15/99 (15.2%), 9/84 (10.7%), and 1/48 (2.1%).\n- Families must be compared within each period; the falling rates across periods mix unequal accumulation time and do not establish market deterioration. The newest windows are especially provisional, and a handful of qualifying games cannot justify announcing a boom or collapse.\n- Roguelike deckbuilders showed the clearest persistent favorable association: approximately 1.50, 1.39, and 1.49 times their date/current-price expected combined-positive counts. The pattern survived tag-cutoff changes and largest-review-publisher removal. This was stronger evidence than the broad “cards/decks” label, not proof of a causal genre advantage.\n- Review-volume concentration was particularly misleading here. Slay the Spire 2 had 182,020 reviews but 61% positive in the snapshot and contributed about 80% of first-half card/deck review volume. Summing reviews would conceal the response distribution and satisfaction differences among other games.\n- Advertised counts of 100+ options had a recurring favorable association in some comparisons, but these are unverified quantity claims and cannot show that adding more items causes better reception. The final interpretation emphasized consequential alternatives rather than nominal counts.\n\n**Concrete build-game examples and differentiation hypotheses:**\n\n- Auto battlers offered varied configurations despite only modest aggregate advantages: Oaken Tower advertised a tower of interacting items (2,889 reviews, 89% positive); Skull Horde a necromancer’s army and loot (1,070, 81%); Legionbound recruited heroes/class synergies/ascensions (642, 83%); Wireworks a module board, wiring, and signals (602, 89%); Hive Blight a team of insect warriors (357, 94%). These figures are September 5 observations.\n- The assistant’s interpretation was that what players assemble, and the rules connecting its parts, offer room for differentiation inside recognizable combat structures. This was a design hypothesis drawn from examples, not a dataset proof that these structures are inexpensive or underserved.\n- Action build games remained active without much additional favorable association over the build-game baseline after date/price adjustment. Examples included Vital Shell’s fantasy-archetype mechs (1,699 reviews, 96% positive, $5.99), Arms of God’s five simultaneous weapons/upgrading/merging (1,845, 88%, $11.50), and TerraTech Legion’s block-built combat vehicle (2,060, 81%, $9.99).\n- Tangy TD advertised item-equipped class-based towers and build choices (1,494 reviews, 90% positive), and Backpack Dungeon advertised placement, adjacency, and triggers as the combat core (622, 88%). However, the broader spatial-inventory classifier also captured ordinary storage; tightening it left too few cases for a defensible underserved-niche claim. The apparent broad advantage was dropped.\n- Counterexamples included Packmates, combining survivor combat and adjacency, with 38 reviews, and Twisted Fate, advertising party builds and ability modifiers, with 39. Their existence did not disprove the ideas but prevented treating the selected successes as a general recipe.\n- A familiar activity carrying build mechanics—chess, dice, poker, and related forms—showed no consistent general advantage. Individual successes remained useful examples, but “familiar activity plus roguelike builds” was insufficient as a market thesis. Pre-run/drafting and machine/party distinctions likewise did not yield a reliable universal recommendation merely because the classifier could generate a row for them.\n\n**Bounded review sample and meaningful alternatives versus nominal variety:**\n\n- The July–August loot/action-RPG subset initially appeared strong on attention: 10/34 games had at least 100 reviews. Only two also reached 80% positive. This motivated a small review investigation rather than interpreting high review volume as satisfied demand for their build systems.\n- The assistant announced a fixed sample of at most 30 positive and 30 negative recent English Steam-purchase reviews for each of SpiritVale, Beast of Reincarnation, Echoes of Aincrad, Shadow Dungeon, and Ruins of Dawn. Ten requests returned 273 reviews; Ruins of Dawn had fewer available matches. Network collection was disabled again immediately afterward; no assets were downloaded.\n- The sample intentionally selected both sentiments and cannot estimate complaint prevalence, overall satisfaction, country-level demand, or sales. The claims below are interpretations of selected player accounts, not independent verification that the described defects or mechanics exist.\n- Negative Shadow Dungeon reviewers described enormous advertised item/build counts resolving into numerical variations, repeated gear bases, linear elemental choices, or eventually buying all useful skills so configuration stops requiring tradeoffs. Some positive reviewers still enjoyed the loot/progression loop while expressing related reservations.\n- A positive Ruins of Dawn reviewer described a dagger interacting with a ring so that its lower displayed base damage became better for a particular setup; another described itemization transforming an initially weak fireball. These supplied concrete examples of configuration-dependent usefulness.\n- SpiritVale reviews could praise respec flexibility and class/item optimization while criticizing servers, balance, bots, economy, or support. Other sampled games mixed combat/build criticism with presentation, pacing, performance, or franchise expectations. Aggregate dissatisfaction therefore could not simply be attributed to rejection of build expression.\n- The resulting hypothesis was “conditional value”: an item becomes desirable because of a particular setup or changes what an ability does, rather than merely increasing a number until every option is owned. The assistant proposed prioritizing a compact game with a distinctive configuration rule and consequential interactions, while explicitly acknowledging that Steam data did not establish development duration or minimal-marketing viability.\n\n**Artifacts, verification, and continuation:**\n\n- Source data and collection documentation live under `/home/adn/a327ex/steam-research/2026-09-05/` and the research-root `README.md`. Key files are `catalog.sqlite`, `exports/games.parquet`, `exports/other_apps.parquet`, `exports/games.csv.gz`, `exports/schema.json`, `exports/manifest.json`, `field_inventory.csv`/`.json`, `coverage.json`, `validation.json`, `metadata_complete.json`, `alias_repair.json`, `provenance.json`, and compressed `raw/` responses.\n- The first analysis lives under `/home/adn/a327ex/steam-research/analysis-2026-09-05/`: `findings.md`, `README.md`, prepared game/tag Parquet files, yearly and January–August tables, `tag_landscape.csv`, all/frozen/holdout pair tables, group-member exports, rare-pair and description follow-ups, price corrections, generated pair/price comparison charts, and `validation.json`. Scripts support reproducing preparation, analysis, deeper checks, follow-ups, charts, and validation.\n- The focal pass lives under `/home/adn/a327ex/steam-research/build-expression-2026-09-05/`: `findings.md`, `README.md`, `classification.parquet`, classification/analysis specifications, blind audit and labels, `analysis_membership.parquet`, `supply.csv`, `family_results.csv`, period/definition/tag-rank sensitivity data, per-family member Parquet exports, review CSV/manifest/raw samples, classifier/analysis/review scripts, and final validation artifacts.\n- The broad analysis verified source/identity/date accounting, frozen selection hashes, discovery/holdout separation, recomputed baseline/year/pair counts, script parsing, and disabled media/network collection. The build pass ultimately passed all 39 consistency checks, including reproduction of 500 classifier records, calibrated baseline ratios, family counts, bounded review sampling, network pause, and absence of downloader/media directories. A stale diagnostic classifier signal was corrected before the successful run.\n- The most useful continuation is to investigate concrete configuration mechanisms and the distinction between meaningful alternatives and numerical abundance, using the saved game lists and evidence excerpts. A separately bounded review-history collection could answer questions about age-adjusted traction that this snapshot cannot. No decision was made to build a specific new game or change a current project.\n- The owner requested ending the session under the title “Steam Market Research 5” and continuing research in another task. Because the main agent had compacted, he explicitly asked that a subagent read the complete converted Markdown transcript and write the session summary independently. This summary follows that full converted record; historical tool output that was already truncated in the record is not represented as recovered text.\n\n---\n\n::role user 15:38:29|05.09.26\n\n> Let's try another task. I worked on Steam market research with Fable. You may not read those logs nor their summaries, as I'd like you to do this research again but see what you can derive anew yourself. I used to be extremely on top of Steam and know all the details about what you \"had\" to do to be successful in the market, but 5 years have passed and I'd like to update myself. most of what I did involved collecting data on all Steam games and writing various scripts to answer questions I had and understand trends better. I'd like you to do some of that work today, however, I'd also like you to generate most research directions and questions and then useful answers that exploit openings in the market that other people might not have noticed. To do this, I'd like you to avoid reading ANY AND ALL Steam marketing research blog posts (including the previous ones generated by Fable), such as howtomarketagame and similar. This isn't because those blogs aren't good, but it's because their own analysis is already what's on the minds of every indie developer, and I'd like insights different from theirs. But again, you shouldn't read any of it because I don't want your research to be *opposed* to theirs, I want it to be simply indifferent. So, first, how would you go about this task? Ask me any questions you might have, and maybe even learn more about me by reading through what I've written regarding how I like to make games, how I like to market them, etc. I am especially fond of the \"make a game in a few months and market it with minimal effort but high possible gain actions\" technique, although I'm open to variations or changes to my approach, since I don't know what the market is like today anymore.\n> \n\n::model gpt-6-astra high\n\n::role claude 15:38:34|05.09.26\n\nI’ll start by looking for your own writing about making and releasing games, while excluding the Fable research and all Steam marketing analysis. Then I’ll propose a research approach built around raw market evidence and the kinds of games and release cycles you actually want to pursue.\n\n\n::role tool 15:38:38|05.09.26\n\n<details>\n<summary><code>exec ([memory: sealed target])</code></summary>\n\n<pre><code>[sealed content withheld: memory]</code></pre>\n\n</details>\n\n::role claude 15:38:52|05.09.26\n\nFor a game made in a few months, what would count as a satisfying result for you today—in money, audience, or something else—and how much time and cash would you comfortably risk?\n\nWhich marketing actions would you willingly do, and which would make the whole approach unattractive? For example: a browser prototype, demo or festival, contacting creators, occasional posts, regular videos, or launch discounts.\n\nShould the first research pass mainly inform Horse Game, identify future games you could make in a few months, or both? Are there kinds of games you would refuse to make even if the evidence looked promising?\n\n- Future short-cycle games\n- Both\n- Horse Game first\n\n::role tool 15:38:52|05.09.26\n\n<details>\n<summary><code>request_user_input_async ({\"questions\": [{\"title\": \"For a game made in a few months, what would count as a satisfying result for you today—in m...)</code></summary>\n\n"}}
{"status":"fulfilled","value":{"chunk_id":"fee4c7","wall_time_seconds":0.000002787,"exit_code":0,"original_token_count":5679,"output":"# Steam metadata snapshot — September 5, 2026\n\n**Complete within the documented public-catalog scope.** Network collection is closed. Image files, the image downloader, and its download queue remain deleted. No media binaries are included.\n\nThe clean game dataset contains **184,664 Steam game apps**: **128,972 released** and **55,692 upcoming**. There are separately **148 store-listed mods** and **355 other app records. All **185,173 discovered IDs** are accounted for through direct records or explicit redirects. These counts refer to Steam app identities, not a manual deduplication of game titles or editions.\n\n## Coverage and snapshot meaning\n\nAll 23 selected public storefront catalogs were enumerated completely: US, BR, JP, CN, DE, GB, FR, CA, AU, KR, TW, HK, RU, UA, TR, IN, SG, MX, AR, TH, ID, VN and ZA. Collected unique-ID counts match the maximum totals Steam reported for each catalog. A few totals increased during collection; the additional IDs were captured. The source response ranges are recorded in `coverage.json`.\n\nDetailed regional responses were collected for US, BR, JP, CN and DE. The other regions supplied catalog coverage and additional-game metadata. The separate Steam search audit remains a partial cross-check; its discovered package constituents are retained as non-game records. It is not the basis of the completeness claim.\n\nPrimary metadata was fetched from **2026-09-05T19:09:38.986403+00:00** through **2026-09-05T20:53:33.945370+00:00**. This is a September 5 collection window, not an atomic snapshot of a changing service. Fully delisted, unpublished, unlisted or exclusively available outside the checked storefronts can be absent. Nothing here claims to enumerate every app ever registered with Steam.\n\n## Available per-game data\n\nAvailability below refers to the clean, preferred per-game record (normally US, with regional fallback). It is not a count of every regional observation. All original returned JSON fields are preserved. The complete inventory contains **297 observed JSON field paths**, including container and nested-array paths, with per-game presence/nonempty counts and observed types.\n\n| Group | Data | Coverage |\n|---|---|---|\n| Identity/status | App ID, name, store URL/slug, item type, visibility; free, coming-soon and Early Access flags | 184,664 records; 18 names are blank |\n| Descriptions | Short description and full BBCode description | 183,399 short; 184,536 full |\n| Tags | Up to 20 returned tags, tag IDs/names, weights and rank | 184,621 games; 430 distinct tags |\n| Creators | Developer, publisher and franchise arrays; names and Steam creator-clan IDs where provided | 184,417 developer; 183,942 publisher; 36,967 franchise |\n| Release information | Steam/original/original-Steam/EA-exit timestamps; planned-release text and flags; platform release dates where supplied | 143,967 Steam timestamps; 16,045 original; 7,536 original-Steam; 774 EA-ex…764 tokens truncated…coming_games` and `mods`.\n- `2026-09-05/exports/games.parquet`: one row per game, with normalized columns and the full original preferred Steam item in `data_json`.\n- `2026-09-05/exports/other_apps.parquet`: the 503 non-game records, including the 148 mods.\n- Other Parquet tables: regional purchase/review observations, tags and labels, creators, languages, feature categories and labels, media URL references and associations, catalog membership, raw-page provenance and app aliases.\n- `2026-09-05/exports/games.csv.gz`: compact normalized game columns; use Parquet/SQLite for full nested metadata.\n- `2026-09-05/field_inventory.csv` and `.json`: exhaustive per-game JSON field inventory and availability counts.\n- `2026-09-05/exports/schema.json`: exact export table columns, SQLite types, keys and filtering.\n- `2026-09-05/raw/`: gzip-compressed response envelopes, with request parameters and timestamps, preserving detailed regional responses.\n- `2026-09-05/coverage.json`, `validation.json`, `alias_repair.json`, `metadata_complete.json`, `exports/manifest.json` and `provenance.json`: coverage, successful checks, normalization repair, completion, export row counts and code provenance.\n\nPrimary Steam sources are StoreQuery, StoreBrowse, Steam tag data, and store-category metadata. Protocol definitions were inspected through a mirror of Steam protocol schemas. No previous research logs or market-analysis outputs were imported; no third-party game estimates are included.\n\nChecks passed for database integrity, matching catalog counts, accounted-for app IDs, canonical-ID consistency, tag counts and definitions, relationship integrity, review-count/percentage ranges, duplicate primary keys, and removal of image files/downloader/queue. Parquet row counts were verified after writing.\n\n## Reuse\n\nThe collectors are resumable and retain raw responses before importing them. Network collection is disabled in `network_policy.json`; metadata requests use pacing and stop on HTTP 429. No image downloader remains. The Python scripts and the local `.venv` with PyArrow are retained for reproducible metadata processing and exports.\n# First Steam market research pass\n\nSource: the fresh September 5, 2026 Steam snapshot in the adjacent `2026-09-05`\ndirectory. No previous research logs, summaries, marketing blogs, store images or\nexternal game-performance estimates were used for this analysis. Steam network\ncollection remains disabled. The PNG/SVG files here are small statistical charts,\nnot downloaded game images.\n\n## Population and measurements\n\nThe source has 184,664 game apps, including 128,972 marked released. Ninety-three\nreleased records have no usable date and one has a future date; these are retained\nin the source and listed in `release_date_exceptions.csv`, but excluded from dated\ncomparisons. There are 128,878 valid dated released records.\n\nWhole-catalog supply tables retain free and explicit-content games as separate\nsegments. The main comparisons use currently paid games outside Steam's explicit\nsexual-content descriptors 3 and 4, with a valid reported release date. This gives\n38,494 games in the 2023–2025 comparison pool. It is not a manually judged quality\nfilter or an estimate of a particular developer's probability of success.\n\n`original_steam_release_date` takes precedence where supplied, otherwise\n`steam_release_date` is used. Dates remain store-reported, and first exposure can\nbe misstated, particularly for games graduating from Early Access. These are\ncurrent review totals attached to release cohorts, not historical launch outcomes.\nCurrent tags may have been added after success, or adopted more widely over time.\nOlder delisted games may be absent. Consequently the supply tables describe the\nsurviving public catalog, not a complete historical record of releases.\n\nResponse thresholds of 100 and 1,000 use Steam's filtered review count. They are\nconvenient scales, not definitions of financial success, sales estimates or revenue\nestimates. Median and upper-percentile counts, zero/<10 review counts, positivity,\nand concentration are retained alongside threshold rates.\n\nTag prevalence is nonexclusive: a game can belong to multiple groups. The broad\nlandscape uses the returned top 20 tags and checks top-5 specificity. Pair analysis\nrequires both tags among the top 10. Specificity changes can dramatically change\nsample sizes; the rare-pair follow-ups explicitly test broadening to top 20.\n\n## Comparison and selection procedure\n\nFor a group, a release-quarter expected count is the sum of the comparison pool's\n100-review proportion in each member's release quarter. Observed / expected is a\ndescriptive standardization, not a causal or predictive model. Price sensitivity\nuses quarter × current USD entry-price buckets (<=5, 5–10, 10–20, 20–40, >40,\nmissing), falling back to quarter alone where fewer than 40 games form a stratum.\n\nPrices use Steam's best non-bundle, single-game package offer: original/list price\nwhen supplied, otherwise current final price. Offers dependent on another owned\napp are excluded. The earlier minimum-of-all-options approach was corrected\nbecause another option can be an add-on. The correction affected 2,101 records;\n`price_extraction_corrections.csv` preserves the audit. Current prices may be\nresponses to sales and success themselves; this is not a pricing-effect estimate.\n\nPair discovery used 2023–2024 records only. Requirements: >=30 games, >=10 with\n100 reviews, >=30 games in each constituent-only comparison, and overlap with the\nsmaller constituent <85%. Broad/admin/evaluative exclusions are recorded in\n`analysis_specification.json`. A smoothed ratio against the stronger constituent-\nonly group ranks candidates. The 30 selected pairs were saved before inspecting\ntheir 2025 comparison results. The frozen selection's hash was checked after the\nprice and creator-identity corrections and did not change.\n\nThe 2025 comparison is a separate set of games in this snapshot, not a real-time\nforecasting backtest. All 30 results are published, including weak ones. Tag pairs\noverlap heavily; multiple FMV combinations are one related cluster, not independent\nreplications. These screens are exploratory; none establishes a causal advantage.\nWilson intervals in the chart are illustrative binomial intervals, not a correction\nfor selection, publisher clustering, or all the other limits above.\n\nDeveloper/publisher grouping uses reported clan IDs, with normalized-name fallback.\nA missing clan key is filled only when that normalized name maps to exactly one\nobserved clan within the role. Ambiguous names remain unresolved. These identities\nare not a verified map of corporate ownership. Largest-review and most-prolific\ncreator removals are separate sensitivity checks.\n\n## Follow-ups\n\nDescription-based companion/productivity and job-horror queries were devised after\nreading the initial results. They are explicitly exploratory and have no untouched\nvalidation sample. The exact regular expressions, counts, group members and\ncounterexamples are saved. Supported languages describe product localization, not\nbuyer location. No game was played to verify claims in its store description.\n\n`followups.json` includes a 2025 companion-vs-other-idler comparison using release\nquarter and price, with a minimum of 20 peers before falling back to quarter.\n\n## Outputs\n\n- `yearly.csv`, `january_august.csv`: catalog cohorts and response distributions.\n- `tag_landscape.csv`: supply share, response, specificity and concentration.\n- `pair_discovery_all.csv`: 1,765 eligible pair candidates.\n- `pair_discovery_selected_before_holdout.csv`: frozen top-30 selection.\n- `pair_holdout_results.csv`: all 2025 comparisons, including price sensitivity.\n- `pair_*_games.csv`, `group_*_games.csv` / `group_*.csv`: underlying titles.\n- `deep_groups.csv` / `.json`: rank, cohort, positivity and creator-removal checks.\n- `rare_pairs_exploratory.csv`, `rare_broad_*.csv`, `followups.json`: small-sample\n  leads, broader definitions and rejected explanations.\n- `idler_companion_exploratory.csv`, `job_horror_exploratory.csv`: text matches.\n- `price_comparison.png/svg`, `pair_comparison.png/svg`: inspected data charts.\n- `validation.json`: successful independent count and selection checks.\n- `findings.md`: the substantive interpretation, also delivered in chat.\n\nRun, in order, with `../.venv/bin/python`: `prepare.py`, `analyze.py`,\n`deep_dive.py`, `followups.py`, `charts.py`, `validate_analysis.py`.\nThe frozen-selection hash records this first pass; a future dataset or changed\nresearch specification should use a new analysis directory.\n# Build expression on Steam — September 5, 2026\n\nThis is a separate analysis of the collected Steam snapshot, focused on advertised combinations of items, equipment, abilities, characters, classes, cards and other configurable systems. Read [findings.md](findings.md) for the interpretation. No Steam marketing research blogs or Fable research were used in this analysis.\n\n## Population and measurement\n\nThe source is the 184,664 canonical game records in `../2026-09-05/exports/games.parquet`, prepared into `../analysis-2026-09-05/games.parquet` and `tags.parquet`. Public catalog coverage and collection limitations are documented in `../README.md`. Store metadata was collected September 5, 2026; a small review-text follow-up was collected later that day. This is a current visible catalog, not a complete history of games ever sold on Steam.\n\n`classify.py` reads the English short pitch and full description, alongside the returned weighted tags. It detects explicit build claims, deck construction, combining or specializing equipment/skills, party composition, branching classes and configurable machines. A skill tree or a large item count alone is insufficient for the strict definition. The result measures **what the store promises**, not actual viable builds or verified gameplay depth. Missing matches do not establish absence of build expression. English wording, tag ranking, incomplete descriptions and implicit systems affect coverage.\n\nThe main population is currently paid, valid dated released games outside explicit-sex content flags. Central idle/incremental/clicker games and FMV/visual-novel games are excluded using the documented tag and short-pitch rules. Automatic combat by itself is not an exclusion. These rules can miss a central genre when the store does not label it clearly, or exclude an incidental use of a short-pitch keyword. Free games are retained separately in `supply.csv`.\n\nThe labels are overlapping. A title can be an action roguelike, auto battler and party-building game at once. The named broad tag comparisons use any of Steam's returned top 20 tags, after the build-expression screen. Narrow top-five versions remain as sensitivity checks, not the headline definition. This matters: Slay the Spire 2's Roguelike Deckbuilder tag was ranked sixth in this snapshot.\n\nA blind descriptive audit of 36 random strict matches found 31 clear advertisements of configuration and five ambiguous cases. This was a classifier development audit, not an independent holdout estimate of precision. Corrections also removed generic interactive-object/escape-game matches. Inclusive and direct-claim variants are retained. Spatial-layout and machine-assembly screens are particularly incomplete; they cannot support an exhaustive census or a confident spatial-game premium.\n\n## Outcomes and dates\n\nReviews are Steam's filtered review summaries, **not sales, revenue, ownership or unique players**. Outcomes include 10, 100 and 1,000 reviews. A second benchmark is at least 100 reviews and at least 80% positive: a useful way to separate attention from substantial positive reception, not a definition of profitability or objective quality. A game at 79% positive still has a large majority of positive reviews.\n\n2026 is included in three separate windows: January–June, July–August, and September 1–5. The last is too young for a market verdict. Within 2026, comparisons use release half-month bins; older comparisons use release quarters. These are current totals for similarly dated peers, **not reconstructed launch-day or day-90 outcomes**. Cross-year rate declines cannot be read as deteriorating demand because exposure time differs.\n\nDates prefer `original_steam_release_date`, falling back to `steam_release_date`. These fields and the recorded Early Access indicators cannot fully reconstruct actual first exposure. Delisting and regional visibility can also change cohort membership. Some 2026 games have established IP, prior releases or Early Access exposure. Sensitivity columns remove recorded history but cannot eliminate unrecorded history.\n\nThe adjusted ratios compare each subgroup with **other build-promising games**, using its release-date and current-price mix. The expected rate comes from date × price groups with at least 20 peers; smaller groups fall back to date. The fallback mixture is calibrated within each date bin so the full comparison population has observed/expected = 1. Text-length sensitivity adds word-count groups. Ratios are descriptive associations, not causal estimates or personal success probabilities. Small groups and multiple exploratory comparisons warrant caution.\n\nPrice is the snapshot's US standalone best purchase option, using undiscounted price where available, otherwise current final price. Bundles, multi-game packages and master-app-dependent free offers are excluded from that price estimate. Price is not production scope, development duration or historical launch price.\n\nCreator-removal, concentration, alternate wording/tag definitions, and recorded-Early-Access sensitivities appear in the outputs. The definitions were developed through description inspection, but later exploratory follow-ups were informed by results. This is not a preregistered causal study.\n\n## Bounded review follow-up\n\n`sample_reviews.py` made ten requests for five recent loot/action RPGs: SpiritVale, Beast of Reincarnation, Echoes of Aincrad, Shadow Dungeon and Ruins of Dawn. It requested at most 30 recent positive and 30 recent negative English Steam-purchase reviews per game, yielding 273 texts. The request manifest and minimal source responses are retained. No profile names, account identifiers or avatars were retained in the selected review records.\n\nThe sample deliberately balances sentiment, selects only five games, favors English and recent versions, and was inspected selectively for build-related examples. It cannot estimate complaint prevalence or explain the whole market. Reviewer allegations and experiences are not independently verified. The report paraphrases selected examples and does not treat technical or moderation complaints as established facts.\n\n## Files and reproduction\n\n- `findings.md`: interpretation, caveats and primary-source examples.\n- `classify.py`, `classification_spec.json`, `classification.parquet`: current classifier, regex rules, hashes, labels and evidence snippets. Files suffixed `v1` are development history.\n- `analyze.py`, `analysis_spec.json`, `analysis_membership.parquet`: population membership, period bins, analytical rules and outputs.\n- `family_results.csv`: subgroup counts, outcome rates, expected-rate ratios, concentration and sensitivity columns.\n- `supply.csv`, `period_summary.json`, `definition_sensitivity.csv`, `tag_rank_sensitivity.csv`: supplementary comparisons.\n- Period/family Parquets: individual cases with full descriptions and source links.\n- `selected_examples.csv`: a deliberately selected illustration list, including low-review counterexamples; not a representative sample.\n- `audit_blind.json`, `audit_labels.json`: development audit material.\n- `review_sample.csv`, `review_sample_manifest.json`, `review_samples/`: bounded review evidence.\n- `validate.py`, `validation.json`: independent count/identity checks, a 500-record classifier reproduction sample, baseline-calibration checks and collection-state checks.\n\nRun from the workspace with `steam-research/.venv/bin/python steam-research/build-expression-2026-09-05/classify.py`, then `analyze.py`, then `validate.py` using the same directory prefix. Classification over full descriptions is relatively slow. The review collector is not part of ordinary reproduction; local data is sufficient.\n\nNo Steam images, trailers or other media were downloaded for this analysis. Bulk collection remains paused. Only lightweight source references remain from the original metadata collection.\n{\n  \"network_paused\": true,\n  \"metadata_min_interval_seconds\": 2.0,\n  \"stop_on_http_429\": true,\n  \"reason\": \"Metadata snapshot finished. Network collection closed; no media downloader exists.\"\n}"}}
{"status":"fulfilled","value":{"chunk_id":"c98700","wall_time_seconds":0.00000211,"exit_code":0,"original_token_count":716,"output":"steam-research/collect.py\nsteam-research/README.md\nsteam-research/reconcile.py\nsteam-research/search_check.py\nsteam-research/export.py\nsteam-research/network_policy.json\nsteam-research/2026-09-05/metadata_complete.json\nsteam-research/validate.py\nsteam-research/inventory.py\nsteam-research/repair_aliases.py\nsteam-research/build-expression-2026-09-05/classify_v1.py\nsteam-research/build-expression-2026-09-05/audit_blind.json\nsteam-research/build-expression-2026-09-05/classification_spec.json\nsteam-research/build-expression-2026-09-05/classify.py\nsteam-research/build-expression-2026-09-05/audit_labels.json\nsteam-research/build-expression-2026-09-05/audit_blind_v1.json\nsteam-research/build-expression-2026-09-05/analyze.py\nsteam-research/build-expression-2026-09-05/analysis_spec.json\nsteam-research/2026-09-05/collection_status.json\nsteam-research/2026-09-05/catalog_status.json\nsteam-research/2026-09-05/catalog_errors.json\nsteam-research/2026-09-05/paused_status.json\nsteam-research/2026-09-05/verify_ids_errors.json\nsteam-research/2026-09-05/coverage.json\nsteam-research/2026-09-05/exports/schema.json\nsteam-research/2026-09-05/wide_ids_status.json\nsteam-research/2026-09-05/exports/manifest.json\nsteam-research/2026-09-05/wide_ids_errors.json\nsteam-research/2026-09-05/alias_repair.json\nsteam-research/2026-09-05/field_inventory.json\nsteam-research/2026-09-05/provenance.json\nsteam-research/2026-09-05/inventory_summary.json\nsteam-research/finish_metadata.py\nsteam-research/2026-09-05/validation.json\nsteam-research/build-expression-2026-09-05/period_summary.json\nsteam-research/2026-09-05/catalog.sqlite\nsteam-research/build-expression-2026-09-05/validation.json\nsteam-research/build-expression-2026-09-05/findings.md\nsteam-research/build-expression-2026-09-05/README.md\nsteam-research/build-expression-2026-09-05/validate.py\nsteam-research/build-expression-2026-09-05/review_sample_manifest.json\nsteam-research/build-expression-2026-09-05/sample_reviews.py\nsteam-research/analysis-2026-09-05/preparation_audit.json\nsteam-research/analysis-2026-09-05/prepare.py\nsteam-research/analysis-2026-09-05/analyze.py\nsteam-research/analysis-2026-09-05/analysis_specification.json\nsteam-research/analysis-2026-09-05/analysis_summary.json\nsteam-research/analysis-2026-09-05/deep_dive.py\nsteam-research/analysis-2026-09-05/deep_groups.json\nsteam-research/analysis-2026-09-05/identity_resolution.json\nsteam-research/analysis-2026-09-05/discovery_selection.sha256\nsteam-research/analysis-2026-09-05/simulator_title_check.json\nsteam-research/analysis-2026-09-05/followups.json\nsteam-research/analysis-2026-09-05/followups.py\nsteam-research/analysis-2026-09-05/charts.py\nsteam-research/analysis-2026-09-05/findings.md\nsteam-research/analysis-2026-09-05/README.md\nsteam-research/analysis-2026-09-05/validation.json\nsteam-research/analysis-2026-09-05/validate_analysis.py\n"}}

Script completed
Wall time 0.3 seconds
Output:

{"status":"fulfilled","value":{"chunk_id":"491a79","wall_time_seconds":0.000002567,"exit_code":0,"original_token_count":1984,"output":"- The paid/non-explicit catalog contained 5,643 games in the 2019 cohort, 9,771 in 2023, and 15,674 in 2025. Their current counts above 100 reviews were 1,243, 1,672, and 2,277 respectively. Thus the larger catalog includes more substantially reviewed games as well as a larger low-response tail; 8,811 of the 2025 games had fewer than ten reviews.\n- The whole-catalog January–August supply comparison counted 17,709 releases dated in 2026 versus 12,925 in 2025. This measured surviving public-catalog supply, not changing probability of launch success.\n- Within the 2023–2025 paid/non-explicit pool, Action Roguelike had 2,966 games, median nine reviews, and 17.1% above 100; Auto Battler had 657, median 14, and 22.4%; Roguelike Deckbuilder had 412, median 40.5, and 36.9%. Fast growth in a tag’s share of releases did not necessarily imply exceptional response per game.\n- Incremental had 1,865 games, median seven reviews, and 14.2% above 100; Idler had 1,455, median 12, and 19.0%. Their position improved relative to similarly priced/date-grouped peers: approximately 1.52 and 2.06 times the expected 100-review counts respectively. This comparison did not show that lowering a price causes improvement.\n- Tags themselves can mislead: several filmed romance games were tagged Immersive Sim, so an aggregate for that label could not be read simply as demand for a Deus Ex-like design.\n\n**Desktop/background companions and filmed visual novels:**\n\n- An exploratory description screen broadened the small Desktop Companion tag to promises involving productivity/focus tools, Pomodoro timers, or games occupying a screen corner while the user works or does something else. Among 2025 idlers, 42/130 companion-positioned games reached 100 reviews (32.3%, median 23), versus 117/650 other idlers (18.0%, median nine).\n- Sixteen companion matches exceeded 1,000 reviews. Similar quarter/price peers predicted approximately 25 games above 100; 42 were observed. Removing the largest developer or publisher left 31.8%. Examples included Tiny Pasture, Fantasy Map Simulator, Ropuka’s Idle Island, Tiny Aquarium, and Cornerpond. Taskbarn and Strange Shores illustrated that similar positioning could still receive little response.\n- The creative question proposed was what other ongoing systems people might enjoy alongside another activity. This was explicitly exploratory: the description screen was devised after seeing initial results and had no untouched validation sample.\n- Visual Novel + FMV in the top ten produced 69/86 games above 100 reviews, 23 above 1,000, and a median of 360 across 67 developer groups. Forty-eight also met 80% positive. The split was 33/42 in 2023–2024 and 36/44 in 2025; removing the publisher with the most reviews left 66/83 above 100. Price-sensitive comparisons remained favorable.\n- Dating Sim + FMV, Interactive Fiction + FMV, and Immersive Sim + FMV substantially overlapped this same cluster; they were treated as one finding. Descriptions included romance, relationship decisions, and palace intrigue, with Love Is All Around, Road to Empress I, and Mandate Of Heaven among the examples and Light of Reversing Destiny as a lower-response counterexample.\n- Seventy-seven of those 86 games supported Simplified Chinese and 68 English. This was product-localization evidence, not buyer nationality or regional sales, and limited transfer of the finding to another cultural/language market. The owner later ruled out making incrementals, idlers, or FMV visual novels; these remained broad research findings rather than recommendations for him.\n\n**Automation, inexpensive thrillers, and attractive leads that weakened:**\n\n- Resource Management + Automation in the top ten contained 98 games across 94 developer groups, with 49 above 100 reviews and median 104. Satisfactory contributed 78.5% of total reviews, yet removing it left 48/97 above 100. Removing the top tenth by review count still left 44.3%, demonstrating the difference between a concentrated total and a broader distribution.\n- Inexpensive examples included Beltmatic, Factory Town Idle, Widget Inc., and Incredicer, whose descriptions framed production as arithmetic, a living spreadsheet, incremental manufacturing, and automated dice rolling. The Factory Must Grow, with eight reviews, provided a counterexample. The interpretation was a healthy family with varied presentations, not proof of economical production.\n- An important chronological qualification replaced the stronger initial pairing interpretation: although the 2025 automation/resource-management group remained above the broad pool, its additional advantage over automation alone did not survive the particular price comparison. Combining those two labels was not established as independently beneficial.\n- Thriller + Walking Simulator produced 43/146 games above 100 reviews, median 40, and median current entry price $4.99. The split was 25/80 in 2023–2024 and 18/66 in 2025; date/price and creator-removal checks remained favorable.\n- A narrower “ordinary work becomes horror” description screen did not show a clear 2025 advantage: 9/78 matches exceeded 100 reviews versus 387/2,428 other horror games. Night-shift, security, cashier, gas-station, and data-entry keywords did not establish a reliable formula merely because memorable examples existed.\n- Hidden Object + Cats weakened in raw holdout response, from 38/113 to 14/114 above 100 reviews, but the newer group still compared favorably with inexpensive peers. Hand-drawn + Cats was a clearer failure to repeat: 32/92 versus 2/62, also weak in the 2025 price comparison. Review histories would be needed before attributing either result to saturation.\n- Tiny intersections could manufacture apparent scarcity. Horror + Typing initially yielded three games using top-five tags, two above 1,000 reviews; widening to the returned top twenty gave 41 games, five above 1,000, and median eight. Similar rank-sensitivity checks tempered other rare combinations. Interesting examples did not establish an empty or underserved market.\n- The assistant finished and saved this analysis before delivering it, but became stuck composing the handover. The owner pointed this out using information from a side chat; the assistant acknowledged the delay and delivered the substantive findings in chat rather than relying solely on files.\n\n**Owner’s build-expression focus and why 2026 was separated:**\n\n- The owner redirected the research toward “games where you can try lots of builds,” including items, characters, equipment, passives, abilities, and other systems allowing different ways to play. He explicitly excluded making incrementals, idlers, and visual-novel FMV games, and asked for the reasoning about 2026 before further research.\n- The assistant explained that 2026 had been excluded from the main broad outcome comparison because exposure time varied sharply: by September 5, a January release had about eight months while an August release had days or weeks. The first pass had nevertheless included 2026 in supply counts and exploratory rare pairs. Excluding it entirely from a current build-game investigation would miss relevant developments.\n- The agreed follow-up therefore separated full-year 2025, January–June 2026, and July–August 2026, comparing families within each window and standardizing for release timing and current price. January–August supply comparisons used equivalent calendar periods. No snapshot-only method could reconstruct a 2025 game’s review total at the same age as a 2026 game.\n- The conceptual questions concerned what is configured (character, equipment, deck, party, army, machine), how alternatives differ (interactions, class combinations, modifiers, skill trees, composition), and when decisions occur (pre-run, during a run, between encounters, persistent progression). Not every one of these axes proved recoverable reliably from text.\n"}}
{"status":"fulfilled","value":{"chunk_id":"e2e559","wall_time_seconds":0.000002649,"exit_code":0,"original_token_count":2834,"output":"# Build-expression market findings\n\nSnapshot: September 5, 2026. These are findings from Steam metadata and a bounded sample of Steam reviews, not from marketing blogs, sales estimates or playing the games. Method and limitations: [README.md](README.md). Numerical tables: [family_results.csv](family_results.csv) and [supply.csv](supply.csv).\n\n## 1. Build promises are growing much faster than the surrounding paid catalog\n\nFor January–August store-date cohorts, the eligible paid population grows from 9,140 games in 2025 to 12,409 in 2026, about 36%. Strict advertised-build matches grow from 843 to 1,717, about 104%. Their share rises from 9.2% to 13.8%. The broader inclusive definition also rises, from 985 to 1,920 games.\n\nThat means a generic promise of combinations, synergies, equipment and many ways to play is becoming much more common. It does not establish that actual build depth doubled: wording, tags, catalog survival and actual supply all contribute. It does suggest that merely having builds is an increasingly weak way to distinguish a game on its store page.\n\nThis is not simply a set of enormous hits. Of the 1,085 strict paid matches dated January–June 2026, 171 had at least 100 reviews and 101 also had at least 80% positive. For July–August the corresponding counts were 632, 75 and 45. Those newer games have had less time; these are current observations, not eventual failure rates. In raw counts, January–August 2026 already has more games clearing the combined benchmark than the corresponding 2025 cohort: 146 versus 109.\n\n## 2. Roguelike deckbuilders have the most consistent favorable association in the main family comparisons\n\nThe table gives games with at least 100 reviews and 80% positive, divided by all strict build-promising games in each family. Families use the returned top 20 tags and overlap. Compare families within a column; unequal exposure time makes comparisons across columns misleading.\n\n| Family | 2025 | January–June 2026 | July–August 2026 |\n|---|---:|---:|---:|\n| All build-promising games | 181 / 1,365 (13.3%) | 101 / 1,085 (9.3%) | 45 / 632 (7.1%) |\n| Roguelike deckbuilder | 41 / 169 (24.3%) | 22 / 132 (16.7%) | 10 / 67 (14.9%) |\n| Action roguelike | 58 / 449 (12.9%) | 29 / 361 (8.0%) | 12 / 204 (5.9%) |\n| Auto battler | 20 / 121 (16.5%) | 9 / 90 (10.0%) | 5 / 46 (10.9%) |\n| Tower defense | 15 / 99 (15.2%) | 9 / 84 (10.7%) | 1 / 48 (2.1%) |\n\nAfter adjusting for each group's release-date and current-price mix within the build-promising population, the roguelike-deckbuilder combined-benchmark ratios are 1.50, 1.39 and 1.49. The association persists when the genre tag cutoff changes from five to twenty and when the largest-review publisher is removed. This is the clearest repeatable descriptive pattern here; it is not proof of a causal genre advantage.\n\nThe distinction between attention and positive reception matters even inside this group. [Slay the Spire 2](https://store.steampowered.com/app/2868840/) has 182,020 filtered reviews but 61% positive in this snapshot. It supplies about 80% of review volume in the broader H1 text-matched cards/decks group. Summing reviews would tell a very different story from counting how many separate games receive substantial positive reception.\n\nExamples of differently framed card/build games include [Black Jacket](https://store.steampowered.com/app/3100370/) (blackjack-inspired combinations; 1,717 reviews, 91% positive), [Cropdeck](https://store.steampowered.com/app/3500680/) (plant decks and passive scarecrows; 420, 92%) and [Re:Night](https://store.steampowered.com/app/2372240/) (deckbuilding plus three-character composition; 125, 88%). These are illustrative store claims, not evidence of short development schedules.\n\n## 3. Automatic combat still offers room for different things to build\n\nAuto battlers have a modest favorable association in the older and H1 samples. The July–August adjusted ratio is stronger, but it rests on only five qualifying games; that is insufficient to call a boom.\n\nThe interesting variation is in what the player configures:\n\n| Game | Advertised configuration | Reviews | Positive | US entry price |\n|---|---|---:|---:|---:|\n| [Oaken Tower](https://store.steampowered.com/app/3400960/) | Tower of synergizing items in PvP | 2,889 | 89% | $9.99 |\n| [Skull Horde](https://store.steampowered.com/app/3199360/) | Necromancer's skeleton army and loot | 1,070 | 81% | $12.99 |\n| [Legionbound](https://store.steampowered.com/app/4283360/) | Recruited heroes, class synergies and ascensions | 642 | 83% | $9.99 |\n| [Wireworks](https://store.steampowered.com/app/4206270/) | Module board, wiring and interacting signals | 602 | 89% | $6.99 |\n| [Hive Blight](https://store.steampowered.com/app/2886620/) | Team of insect warriors | 357 | 94% | $12.49 |\n\nMy inference is that the choice of what players assemble is a productive source of differentiation. These examples advertise different configuration problems inside a recognizable combat structure. The data does not tell us which would be cheapest to develop, whether players perceive them as novel, or whether differentiation caused their response.\n\n## 4. Action builds remain viable, but generic genre membership conveys little advantage\n\nThe broad action-roguelike family is large: 361 matches in H1 2026 and another 204 in July–August. Its adjusted combined-benchmark ratio is about 1.06 in H1 and 0.95 in July–August, roughly the build-population baseline. This is a crowded active area, not evidence that action builds are dead.\n\nCurrent examples include [Vital Shell](https://store.steampowered.com/app/3741860/) (1,699 reviews, 96%, $5.99), whose pitch combines mechs with fantasy archetypes; [Arms of God](https://store.steampowered.com/app/3100310/) (1,845, 88%, $11.50), with five simultaneous weapons and merging; and [TerraTech Legion](https://store.steampowered.com/app/3596700/) (2,060, 81%, $9.99), with a battle vehicle assembled block by block. [Cursemark](https://store.steampowered.com/app/3219180/) has 561 reviews at 85% and advertises ability-transforming runes.\n\nThe useful question is which configuration rule or interaction a game makes immediately legible, rather than whether it belongs to the action-roguelike tag. These are current examples of positive reception, not an estimated formula or evidence that inexpensive games are inexpensive to make.\n\n## 5. Attention to loot RPGs does not mean satisfaction with their build systems\n\nThe July–August loot/action-RPG screen contains 34 games. Ten have at least 100 reviews, but only two also reach 80% positive. Attention is about twice the date/price expectation, while the combined positive-response ratio is 0.70. The largest examples include SpiritVale (4,579 reviews, 49%), Beast of Reincarnation (4,093, 56%) and Echoes of Aincrad (4,047, 47%).\n\nA bounded follow-up collected 273 recent English Steam-purchase reviews from five games, with up to 30 of each sentiment per game. This deliberately balanced, selected sample cannot estimate complaint prevalence. The inspected examples nevertheless show why aggregate scores are insufficient:\n\n- SpiritVale reviews include praise for respecs and class/item optimization alongside complaints about operations, balance and the economy. A low score is not necessarily a rejection of build-heavy play.\n- A Beast of Reincarnation reviewer describes many nominal options but little reason to choose differently; another objects to exploration mainly yielding upgrade materials. Positive reviewers value its accessible combat.\n- Shadow Dungeon negative reviews distinguish advertised combinatorial item counts from meaningful alternatives. One describes buying every skill and setting everything to autocast; another describes procedural variants that mostly change numbers. Some positive reviewers enjoy the loot loop despite finding it easy.\n- A Ruins of Dawn positive review describes a dagger/ring interaction that makes a lower-base-damage weapon better for that particular setup. Another enjoys items transforming an initially weak fireball. Negative reviews include complaints about the slow start and uninteresting early skills.\n\nThe resulting hypothesis is specific: **conditional value and behavior-changing interactions may matter more to build expression than the sheer number of upgrades or permutations.** The review examples explain this distinction; they do not prove its market-wide frequency or financial value. The underlying texts and review IDs are retained in `review_sample.csv` (examples: Shadow Dungeon 234216811, 234030245 and 234040484; Ruins of Dawn 232897625 and 233001140).\n\n## 6. Promising-looking smaller combinations need more caution, not stronger claims\n\n[Tangy TD](https://store.steampowered.com/app/2245620/) has 1,494 reviews at 90% and equips class-based towers with items and abilities. [Backpack Dungeon](https://store.steampowered.com/app/4478640/) has 622 at 88% and makes placement, adjacency and triggers central to combat. They suggest interesting transfers of build expression into other structures.\n\nHowever, the general spatial-description screen caught ordinary storage and decoration too. Tightening it leaves only nine H1 and seven July–August matches. There is no defensible market-wide spatial premium to report. Machine assembly is also incompletely captured. These are case-finding tools, not reliable niche-size estimates.\n\nNor did combining builds with a familiar activity such as chess, dice, poker, pinball or minesweeper show a consistent general advantage. Some individual games perform well, but the narrower explicit-pitch group falls below its adjusted combined-benchmark expectation in both 2026 windows. It is not a measurement of novelty, and it does not justify declaring that avenue exhausted.\n\nCounterexamples matter: [Twisted Fate](https://store.steampowered.com/app/4210680/) advertises party composition and ability modifiers but has 39 reviews; [Packmates](https://store.steampowered.com/app/3171540/) combines survivor combat with item adjacency but has 38; [Shattered Chess](https://store.steampowered.com/app/4131940/) advertises rule-breaking chess combinations and has 28. These are current low-response observations, not predictions of eventual failure. A distinctive premise is not sufficient by itself.\n\n## Interpretation for further exploration\n\nThe evidence supports exploring build systems broadly, while treating generic build promises as increasingly commonplace. Roguelike deckbuilders offer the strongest recurring favorable association in this pass; auto battlers offer several well-received, varied configuration objects; action games remain active with less subgroup advantage. A smaller recognizable game structure with an unusual, mechanically consequential configuration system is a reasonable design hypothesis to investigate, not an undiscovered profitable niche established by these data.\n\nSteam metadata cannot tell us which of these games took a few months to make, how much marketing work they required, or whether their review counts meet an owner's personal goals. Those claims would require additional primary evidence. No recommendations here depend on imitating the owner's current project, returning to incrementals/idlers/FMV, or treating revenue maximization as the objective.\n"}}
{"status":"fulfilled","value":{"chunk_id":"f92c9a","wall_time_seconds":0.098909916,"exit_code":0,"original_token_count":1260,"output":"2026-09-05/exports/games.parquet rows: 184664 columns: ['appid', 'name', 'type', 'visible', 'is_free', 'is_coming_soon', 'is_early_access', 'steam_release_date', 'original_release_date', 'original_steam_release_date', 'release_from_early_access_date', 'reviews_filtered_count', 'reviews_filtered_positive_percent', 'reviews_unfiltered_count', 'reviews_english_count', 'short_description', 'source_country', 'fetched_at', 'data_json']\nanalysis-2026-09-05/games.parquet rows: 184664 columns: ['appid', 'name', 'type', 'visible', 'is_free', 'is_coming_soon', 'is_early_access', 'steam_release_date', 'original_release_date', 'original_steam_release_date', 'release_from_early_access_date', 'reviews_filtered_count', 'reviews_filtered_positive_percent', 'reviews_unfiltered_count', 'reviews_english_count', 'short_description', 'source_country', 'fetched_at', 'developer', 'publisher', 'developer_key', 'publisher_key', 'usd_list_price', 'adult_explicit', 'steam_date', 'first_date', 'known_ea_history', 'description', 'english_support', 'schinese_support', 'languages_count', 'demo_related', 'windows', 'native_linux', 'deck_code', 'store_url', 'year', 'quarter', 'month', 'reviews', 'positive_pct', 'valid_released', 'price_bucket', 'main']\nbuild-expression-2026-09-05/classification.parquet rows: 184664 columns: ['appid', 'strict', 'inclusive', 'candidate', 'exclusion', 'object_terms', 'quantity_claims_json', 'signal_explicit_build', 'signal_branching', 'signal_combination', 'signal_choice', 'signal_playstyle', 'signal_deck', 'signal_skill_tree', 'signal_quantity', 'signal_idle_central', 'signal_machine', 'signal_party', 'signal_prerun', 'signal_draft', 'signal_craft_merge', 'signal_position', 'evidence_json', 'card_deck', 'party_composition', 'equipment_skills', 'machine_assembly', 'loop_action_rogue', 'loop_auto_battler', 'loop_tactics_rpg', 'loop_action_rpg', 'loop_deck_rogue']\nbuild-expression-2026-09-05/analysis_membership.parquet rows: 184664 columns: ['appid', 'strict', 'inclusive', 'direct', 'build', 'eligible', 'paid', 'date_bin', 'quantity_max', 'familiar_carrier', 'spatial_items', 'spatial_build_effect', 'exclusion']\nanalysis-2026-09-05/validation.json\n{\n  \"passed\": true,\n  \"checks\": {\n    \"source_game_count\": true,\n    \"unique_game_ids\": true,\n    \"dated_release_accounting\": true,\n    \"frozen_discovery_hash\": true,\n    \"discovery_holdout_disjoint\": true,\n    \"baseline_counts\": true,\n    \"all_selected_pair_counts_recomputed\": true,\n    \"yearly_counts_recomputed\": true,\n    \"no_store_asset_downloader\": true,\n    \"steam_network_disabled\": true,\n    \"analysis_scripts_parse\": true\n  },\n  \"scripts_sha256\": {\n    \"prepare.py\": \"a7ee1022db4695017b9d9f527f9106f8290c3ee6695be346e7a59bd8dc124ee8\",\n    \"analyze.py\": \"87ea26f5064306497dc8f8ea4176fc4226f528e1d9e2d6e78c41bb48790766f8\",\n    \"deep_dive.py\": \"3e0f83041fc8c8a31b59721ace1b01804b1f661b601b56c2a88aec7e98e31587\",\n    \"followups.py\": \"6ea899f388e53ac011a75ef0f95eeb9b206ed44c31f2f541b578d6413aa74b1c\",\n    \"charts.py\": \"0f9140fcbce65129cac576579f5085c64592d7619607bbcb714ccd6bb854ca99\",\n    \"validate_analysis.py\": \"4bebc1d5c4bf5062bbbb0e8af4e3ba8e353602a64cd4cdc06bc4fdab637ce1ab\"\n  }\n}\nbuild-expression-2026-09-05/validation.json\n{\n  \"passed\": 39,\n  \"checks\": {\n    \"canonical_rows_unique\": true,\n    \"classification_covers_source\": true,\n    \"strict_implies_inclusive\": true,\n    \"excluded_not_builds\": true,\n    \"rules_hash_current\": true,\n    \"2025_pool_counts\": true,\n    \"2025_positive_counts\": true,\n    \"2025_within_build_ratio100_date_calibrated\": true,\n    \"2025_within_build_ratio100_price_calibrated\": true,\n    \"2025_within_build_ratio100_text_calibrated\": true,\n    \"2025_positive100_price_ratio_calibrated\": true,\n    \"2025_roguelike_deckbuilder_tag20\": true,\n    \"2025_action_roguelike_tag20\": true,\n    \"2025_auto_battler_tag20\": true,\n    \"2025_tower_defense_tag20\": true,\n    \"2026_H1_pool_counts\": true,\n    \"2026_H1_positive_counts\": true,\n    \"2026_H1_within_build_ratio100_date_calibrated\": true,\n    \"2026_H1_within_build_ratio100_price_calibrated\": true,\n    \"2026_H1_within_build_ratio100_text_calibrated\": true,\n    \"2026_H1_positive100_price_ratio_calibrated\": true,\n    \"2026_H1_roguelike_deckbuilder_tag20\": true,\n    \"2026_H1_action_roguelike_tag20\": true,\n    \"2026_H1_auto_battler_tag20\": true,\n    \"2026_H1_tower_defense_tag20\": true,\n    \"2026_JulAug_pool_counts\": true,\n    \"2026_JulAug_positive_counts\": true,\n    \"2026_JulAug_within_build_ratio100_date_calibrated\": true,\n    \"2026_JulAug_within_build_ratio100_price_calibrated\": true,\n    \"2026_JulAug_within_build_ratio100_text_calibrated\": true,\n    \"2026_JulAug_positive100_price_ratio_calibrated\": true,\n    \"2026_JulAug_roguelike_deckbuilder_tag20\": true,\n    \"2026_JulAug_action_roguelike_tag20\": true,\n    \"2026_JulAug_auto_battler_tag20\": true,\n    \"2026_JulAug_tower_defense_tag20\": true,\n    \"classifier_500_record_reproduction\": true,\n    \"bounded_review_sample\": true,\n    \"network_paused\": true,\n    \"no_media_directories_or_downloader\": true\n  }\n}\n"}}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"e6f64c","wall_time_seconds":0.000002438,"exit_code":128,"original_token_count":35,"output":"fatal: not a git repository (or any parent up to mount point /)\nStopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).\n"}
{"chunk_id":"d04ef7","wall_time_seconds":0.000003046,"exit_code":0,"original_token_count":7216,"output":"# CLAUDE.md\n\nGuidelines for Claude Code instances working in this repository.\n\n---\n\n## Communication Style\n\nBe thorough. Write as much as the subject warrants — explain reasoning, include relevant context, surface adjacent considerations. Stay roughly on topic, but don't truncate for brevity's sake. Thoroughness means more substance, not more structure — depth comes from reasoning and adjacent considerations, not from header scaffolding, preamble, or restating what was already said.\n\nDeliver analyses fully in chat. Session transcripts become published AI logs on a327ex.com — content locked inside files is invisible to log readers. When producing a long analysis, report, or research finding, print it in full in the reply; only also write it to a file when the project needs a durable artifact, and never treat the file as the primary delivery.\n\nCritique freely. Ideas, code, plans, artifacts — evaluate them honestly. Avoid sycophancy. Focus on what seems true rather than what's convenient or what you think I'd like to hear. I crave honest appraisal, including when it's uncomfortable.\n\nEmbrace contradiction. Truth often contains multiple seemingly opposite ideas. Don't censor yourself to avoid inconsistency — multiple perspectives can be correct simultaneously. Present the tension rather than forcing artificial resolution.\n\nStay object-level in free-form discussion. In open-ended, exploratory conversation — thinking through ideas rather than executing a defined task — never default to rationalist-mode discourse. Don't ask meta or audit questions (\"how do you reconcile X and Y,\" \"how do you guard against this bias,\" \"what's your check that you're not fooling yourself\"): they stand above an idea and ask me to defend a system against a pre-supposed flaw, their only outputs are valid/invalid, and so they generate nothing. Don't hedge against an imagined rationalist/LessWrong commenter or pre-empt objections nobody raised. Instead, engage a specific claim from inside it — what it actually asserts, what follows from it, where it leads, what it leaves unexplained — and extend it into territory it hasn't been pointed at yet, which is where insight comes from. Make committal, falsifiable moves: assert a reading even at the risk of being wrong, rather than asking a safe question that commits to nothing. Trust ideas enough to build on them, and find their flaws by leaning weight on them, not by pre-supposing them. The audit move is the safe move — it can't embarrass the asker and always sounds rigorous — but it produces the feeling of rigor without the risk that produces understanding.\n\n---\n\n## Working Style\n\n### When to Ask\n\nThe developer prefers being asked over having Claude guess wrong. When a\ndecision is open, lay it out in chat as a question with labeled options —\nthe same shape as a multiple-choice prompt, just written as prose: list\nthe options, give your recommendation, then stop and let him pick one *or*\nredirect entirely. Don't force a selection — he often wants a direction\nnone of the listed options covered, and the value is in seeing the choice\nframed, not in being made to answer. When you have several questions,\nbatch them into one message rather than asking them one at a time.\n\n**Ask first:**\n- Architecture decisions\n- API design choices\n- Design decisions (gameplay feel, mechanics, UI)\n- When uncertain about intent or priorities\n\n**Proceed, then explain:**\n- Implementation details where the path is clear\n- Performance optimization (get it working first, optimize later)\n\nOpen choices below the ask-first bar — reversible, not architecture,\nnot feel — don't get a standalone question: pick a default and flag it\nin the design brief as a decision made, so it's vetoable there (see\nCode Changes).\n\nThroughout: surface tradeoffs explicitly, and don't treat specs as\nfinal — they're starting points for conversation.\n\n### Pacing\n\n- The agreed brief defines the batch: implement everything it covers\n  (which may be a whole phase), and nothing beyond it\n- After completing a batch, hand it over with instructions for how to\n  test it — no silent completions — then give the user a turn before\n  starting the next\n- Scope grows only through conversation, never autonomously — don't\n  chain into new tasks or build systems that weren't agreed\n\n### Code Changes\n\n- **⛔ DESIGN BRIEF before code.** Before calling `Write` or `Edit` on\n  a code file for a non-trivial change, post a **design brief** in chat\n  — a conceptual explanation of the approach, NOT the code itself. Then\n  wait. The user will usually reply with questions and improvements;\n  iterate the brief with them until you reach a conclusion. Only then\n  write the code.\n\n  A design brief contains:\n  - **Approach** — how each feature works: what it does, the strategy,\n    the key data it holds, its lifecycle.\n  - **Interfaces** — how the feature connects to existing features:\n    which existing functions / files / globals it touches or reuses,\n    and how. Include whenever the feature is not self-contained.\n  - **Functions** — for each new or changed function: its name, what it\n    does, which other functions it calls (selective — the calls that\n    build a conceptual map, not an exhaustive list), and its general\n    structure when not already obvious.\n  - **Decisions** — open choices the brief resolves by default rather\n    than by a standalone question: name each one, the option picked,\n    and why, so it can be vetoed in brief review.\n\n  The brief is a conceptual map, not a spec and not pseudocode — aim for\n  \"a reader could predict the shape of the code from this.\" Once the\n  brief is agreed, write the code directly; do NOT re-post it for a\n  second approval — the brief WAS the review. If while writing you find\n  the brief was wrong in a way that changes the design, stop and post a\n  short follow-up brief for that delta.\n\n  Plan mode is not a substitute for the brief — the brief is iterated\n  in chat, not approved as a plan.\n\n  **Trivial changes** skip the brief — just make them and describe the\n  diff in chat afterward. Trivial means:\n  - Single-line value tweaks (number, string, flag, font size, color hex)\n  - Variable renames within one file\n  - Single-line typo fixes\n  - User-specified imperatives (\"change X to Y\", \"rename foo to bar\")\n\n  If unsure whether something counts as trivial, treat it as\n  non-trivial and write a brief.\n\n  Also exempt (no brief): edits to CLAUDE.md / plan.md / design.md type\n  files when the user has already specified the change in prose.\n\n  **Website (`a327ex-site/`) changes are fully exempted.** No design\n  brief, no permission ask, no confirmation before `git push prod main`\n  (which deploys via the post-receive hook). Edit, commit, and push\n  freely. The user wants to iterate fast on the site without the\n  brief-then-ask handshake. This authorization is durable — applies to\n  every session, not just the one it was granted in.\n\n  **Delegation.** When implementation is delegated (a subagent, or a\n  second instance implementing under an already-agreed brief), the\n  brief travels with the delegation and bounds it. An implementer\n  makes no new design decisions: on hitting an open choice, or finding\n  the brief wrong, it stops and reports to the instance that delegated\n  it — nothing more. The asking, brief, and scope rules elsewhere in\n  this document bind the directing instance, not the implementer.\n  Driving other agents (Astra via Codex CLI, Grok, Opus subagents,\n  Gemini for video): model roles, the run packet, verified commands\n  and the review protocol are in `Anchor/workflow/DELEGATION.md`;\n  `/delegate` is the shim. Read it before handing work to any of them.\n\n  Why: the user wants to review the *design* before the code exists —\n  at the level of approach and structure, not syntax. A conceptual\n  brief surfaces design problems faster than a wall of code does, and\n  keeps the review loop about decisions rather than line-editing.\n\n### Log-friendly Artifacts\n\nSession logs card only delivered files (SendUserFile, the Artifact tool),\nand card content is read from disk at end-session, not at delivery. So\nwhen handing over a viewable file that will keep being iterated (an HTML\nprototype, a render, a page), freeze it first: copy it to a versioned name\nnext to it (`versions/<name>-v3.html`, or a timestamp) and deliver the\nfrozen copy with SendUserFile. The working file keeps its stable name so\nrun commands don't change. Identical content is carded once, so\nre-delivering an unchanged file is free. A bare markdown link to the\nworking file is fine for the terminal but produces no card in the log.\n\nWhy: an iterated file handed over by link four times produced zero cards\n(2026-09-03, Reverse Minesweeper); handed over by SendUserFile it would\nhave produced four cards all holding the final version. The frozen copies\nare also a free rollback history.\n\n### Prose and Text Writing\n\nNever *offer* or *suggest* drafting prose for the user — stories, blog posts, emails, messages, design docs, or any other text. The user writes their own text from scratch; I do not.\n\nMy role with prose is to:\n- Improve what the user has already written (line edits, structural critique, alternative phrasings of existing passages)\n- Discuss structure, beats, ideas, and design before they write\n- Provide research, summaries, or reference material that feeds their writing\n- Sketch outlines or beat structures when asked\n\nWhen editing his text for posting (\"fix typos and grammar\"), fix only mechanical errors — typos, misspellings, subject-verb agreement, obviously missing words, punctuation that garden-paths a sentence — and name each fix when posting. Never restyle: his run-ons, comma splices, unhyphenated compounds, and casual register (including high/low diction collisions) are deliberate voice, not errors. Never introduce em dashes — he doesn't use them. If a fix is arguable, it's style — leave it.\n\nWhen a task seems to call for a draft, do not propose drafting it. Offer research, structure, or critique instead. Specifically avoid the \"let me draft so you can react to it\" move — that's the exact behavior to suppress.\n\nIf the user explicitly requests a draft, that's their override — but never raise the option myself.\n\n(This applies to prose. Code, configs, and engineering documents remain governed by Code Changes above.)\n\n### Building and Running\n\nHis test loop is the scarce resource — it should never be spent on failures a machine could have caught. Before handover, verify mechanically; never verify feel.\n\n**Mechanical verification (expected):** syntax checks, engine builds, and whatever non-interactive harness a project provides — agent instances driven through the eval channel (`anchor drive`), captured frames, balance sims. Per-project CLAUDE.mds list what's available; if a project provides nothing, at minimum check the code parses.\n\n**Feel verification (never mine):** never run a game or executable interactively — no `run.bat` / `run-web.bat` or similar, no opening the game to try it, no driving his mouse or keyboard. Playing, testing, and judging feel are his. Hand over with test instructions (see Pacing).\n\n**Engine C changes (Anchor/engine):** build to verify compilation; never run the executable.\n\n---\n\n## Projects\n\nProject state — what's active, what shipped, current direction — lives in auto-memory and changes too fast for this file; this section is only the stable structural map. Any directory with its own `.claude/CLAUDE.md`: read that first when working there.\n\n⚠ **Two roots since the 2026-08-30 Omarchy migration.** The working tree is `~/a327ex/` (Anchor, a327ex-site, ABAgames/knightvspawns, sketches, lookdev, media, notes, tools) — the only writable disk. Everything else named below (the templates, the other games, `archive/`, [REDACTED#3: lore]) still lives on the HDD at `/mnt/Main/a327ex/`, mounted READ-ONLY: read for reference, copy into `~/a327ex/` before working on it. Old `E:/a327ex/...` paths in docs and logs mean one of these two places.\n\n- **Infrastructure:** `Anchor/` — ONE folder for everything infrastructure (unified 2026-08-26, one git repo): `Anchor/engine/` (the game engine: C engine + Lua framework + docs + playground) and `Anchor/workflow/` (SPR, the workflow system: capture → assemble → review → publish + play loop + session tooling; read its README.md first). Future infrastructure projects land as further `Anchor/<project>/` siblings. ⚠ The live site's wasm builds from `Anchor/engine/engine`, so engine changes can reach production. Old engine generations (v1, Anchor2, YueScript-era) live untouched in `archive/` — historical reference only, never use their docs or patterns; some dormant games still sit on them via their own self-contained framework copies.\n- **Site:** `a327ex-site/` — a327ex.com content (logs, posts, media) plus the engine renderer in `a327ex-site/renderer/`. `anchor.blot.im/` is a legacy redirect, no longer written to.\n- **Templates (forkable starters), each with its own `.claude/CLAUDE.md`:** `snkrx-template/` (SNKRX visual style: attack carriers + UI toolkit), `ricochet-template/` (Ricochet blueprint-room style: pattern × color × dither + UI), `emoji-template/` (Twemoji style, on Anchor 3).\n- **Games:** consult memory for which are active; each has its own `.claude/CLAUDE.md` and/or `design.md`.\n\n---\n\n## Anchor Engine\n\n**The engine lives at `Anchor/engine/` and is the default for all projects.** Generation-wise it is \"Anchor 3\": it extends Anchor 2 — same C engine lineage, same procedural Lua framework architecture (free functions, plain tables; no object tree, no action phases, no `an` god-object; the `collider` class is the one exception) — with 3D: Box3D physics (`physics3`) and 3D layers (`layer3`). ⚠ The live site's wasm builds from `Anchor/engine/engine`, so engine changes can reach production. Old generations live in `archive/` (Anchor2 = the 2D predecessor some dormant projects still sit on; v1 in `archive/Anchor`) — never modify or source patterns from them.\n\n### Structure\n\n**Engine** — C code in `Anchor/engine/engine/src/anchor.c` (single file, ~18,750 lines — the source of truth)\n**Framework** — Lua framework files in `Anchor/engine/framework/anchor/`\n**Games** — Lua code using the engine; each game ships its own `anchor/` copy of the framework that can drift — when a function isn't where the docs say, grep the game's copy, which is ground truth for that game. Port framework changes into a game's copy individually; never overwrite a game's `anchor/` file wholesale with the framework version.\n\n### Documentation\n\n- `Anchor/engine/docs/SURFACE.md` — ⭐ THE ONE PAGE on how games are run, driven, recorded and\n  reloaded (2026-09-05): `anchor <game | file.apr> [--agent]` is the whole command line; every\n  mode is a function; I test a game by driving an agent instance (`anchor drive start <game>`,\n  `anchor drive eval <game> '<lua>'`, `anchor drive stop`) — never by running it on his desktop.\n  Read it at the start of any game or engine session, then `docs/AGENT.md` for the details.\n- `Anchor/engine/.claude/CLAUDE.md` — working agreements (read first when working in the engine)\n- `Anchor/engine/reference/plan.md` — authoritative plan for the 3D extension\n- `Anchor/engine/docs/ENGINE_API_QUICK.md` / `ENGINE_API.md` — engine API: compact signatures / detailed with examples\n- `Anchor/engine/docs/FRAMEWORK_API_QUICK.md` / `FRAMEWORK_API.md` — framework API: compact / detailed\n- `Anchor/engine/docs/3D_API.md` — the 3D extension (physics3 / layer3)\n- `Anchor/engine/playground/main.lua` — canonical style example\n\nWhen using engine or framework functions, look them up in the docs — don't guess signatures or parameters.\n\n### Commands\n\n```bash\n# Engine (C) — build to verify compilation (see Building and Running)\ncd ~/a327ex/Anchor/engine/engine && ./build.sh              # Linux desktop build (build.bat on Windows)\ncd ~/a327ex/Anchor/engine/engine && ./build-web-engine.sh   # Wasm build (feeds the live site)\n~/a327ex/Anchor/engine/replay-test/check.sh                 # the replay system's byte-compare (after any engine change)\n\n# Games — an AGENT INSTANCE is how I run one (hidden window, externally paced, recorded):\nanchor drive start <game-dir>                               # or a recording: anchor drive start <file.apr>\nanchor drive eval  <game-dir> 'engine_step(120) return #pawns'   # any Lua; agent_tap / agent_shot / engine_snapshot ...\nanchor drive stop  <game-dir>                               # finalizes the recording (it lands in the session log)\nanchor check <game-dir>                                     # LuaLS against the generated engine declarations\n```\n\nEach game/playground directory has its own `run.bat` — running it is his, not mine (see Building and Running). Saving a file while his game runs reloads it in place (code, textures, sounds, shaders — a definition-only `main.lua` + a `boot.lua`, see SURFACE.md \"Reloading\").\nfrom pathlib import Path\nimport json,math,re,hashlib\nimport numpy as np\nimport pandas as pd\nR=Path(__file__).resolve().parent;G=pd.read_parquet(R.parent/'analysis-2026-09-05/games.parquet');C=pd.read_parquet(R/'classification.parquet');T=pd.read_parquet(R.parent/'analysis-2026-09-05/tags.parquet')\nD=G.merge(C,on='appid',validate='one_to_one').set_index('appid',drop=False)\nD['eligible']=D.valid_released&~D.adult_explicit&D.exclusion.eq('')\nD['paid']=D.eligible&D.is_free.eq(0)\nD['build']=D.paid&D.strict\nD['direct']=D.build&(D.signal_explicit_build|D.signal_deck)\nD['date_bin']=D.quarter\nis26=D.year.eq(2026)\nD.loc[is26,'date_bin']=D.loc[is26,'first_date'].dt.strftime('%Y-%m')+'-'+np.where(D.loc[is26,'first_date'].dt.day<=15,'01','16')\nD['text_words']=(D.short_description+' '+D.description).str.split().str.len()\nD['text_bucket']=pd.cut(D.text_words,[-1,150,300,600,float('inf')],labels=['<=150','151-300','301-600','>600']).astype(str)\noriginal=pd.to_datetime(D.original_release_date.where(D.original_release_date>0),unit='s',utc=True,errors='coerce')\nD['recorded_prior_release']=original.lt(D.first_date-pd.Timedelta(days=365))\n\n# Different playable objects can share the same kind of build expression.\nprimitive=re.compile(r'\\b(?:pachinko|pinball|billiards?|dice|chess|slot machines?|poker|domino(?:es)?|minesweeper|mahjong|bowling|bingo|roulette|brick[- ]break\\w*)\\b',re.I)\nD['familiar_carrier']=(D.name.fillna('')+' '+D.short_description).str.contains(primitive)\nspatial=re.compile(r'\\b(?:arrang\\w*|position\\w*|plac\\w*|adjacen\\w*)\\b.{0,80}\\b(?:items?|weapons?|modules?|backpack|inventory|equipment)\\b|\\b(?:backpack|inventory)\\b.{0,100}\\b(?:arrang\\w*|position\\w*|adjacen\\w*)\\b',re.I)\nD['spatial_items']=(D.short_description+' '+D.description).str.contains(spatial)\n# A broad placement match can mean storage or decoration. The tighter follow-up\n# requires a container/layout reference near a configuration effect.\nspatial_effect=re.compile(r'\\b(?:backpack|inventory|grid|slots?|modules?)\\b.{0,160}\\b(?:adjacen\\w*|synerg\\w*|combos?|triggers?|bonuses|bonus|stronger|buffs?)\\b|\\b(?:adjacen\\w*|synerg\\w*|combos?|triggers?|bonuses|bonus|stronger|buffs?)\\b.{0,160}\\b(?:backpack|inventory|grid|slots?|modules?)\\b',re.I)\nD['spatial_build_effect']=D.spatial_items&(D.short_description+' '+D.description).str.contains(spatial_effect)\n\nD['quantity_max']=D.quantity_claims_json.map(lambda s:max([q['count'] for q in json.loads(s)]+[0]))\nD['quantity50']=D.quantity_max.ge(50)\nD['quantity100']=D.quantity_max.ge(100)\n\nFAMILIES={\n 'all_build_promises':pd.Series(True,index=D.index),\n 'cards_decks':D.card_deck,\n 'roguelike_deckbuilder':D.card_deck&D.loop_deck_rogue,\n 'party_team_composition':D.party_composition,\n 'equipment_skills':D.equipment_skills&~D.card_deck,\n 'machine_assembly':D.machine_assembly&~D.card_deck,\n 'action_roguelike':D.loop_action_rogue&~D.card_deck&~D.loop_auto_battler,\n 'auto_battler':D.loop_auto_battler,\n 'party_tactical_rpg':D.loop_tactics_rpg&~D.card_deck&~D.loop_auto_battler,\n 'loot_action_rpg':D.loop_action_rpg&~D.card_deck&~D.loop_action_rogue&~D.loop_auto_battler,\n 'familiar_carrier':D.familiar_carrier,\n 'spatial_item_configuration':D.spatial_items,\n 'spatial_build_effect':D.spatial_build_effect,\n 'explicit_prerun':D.signal_prerun,\n 'drafting':D.signal_draft,\n 'crafting_fusion':D.signal_craft_merge,\n 'claims_50plus_options':D.quantity50,\n 'claims_100plus_options':D.quantity100,\n}\n# Broad-tag versions keep major games whose specific genre tag falls below rank 5.\nfor label,tag in [('roguelike_deckbuilder_tag20','Roguelike Deckbuilder'),('action_roguelike_tag20','Action Roguelike'),('auto_battler_tag20','Auto Battler'),('tower_defense_tag20','Tower Defense')]:\n ids=set(T[(T.tag==tag)&(T['rank']<=20)].appid)\n FAMILIES[label]=pd.Series(D.index.isin(ids),index=D.index)\n\nPERIODS={\n '2019_2021':D.year.between(2019,2021),\n '2023_2025':D.year.between(2023,2025),\n '2025':D.year.eq(2025),\n '2026_H1':D.year.eq(2026)&D.month.le(6),\n '2026_JulAug':D.year.eq(2026)&D.month.between(7,8),\n '2026_Sep1_5':D.year.eq(2026)&D.month.eq(9),\n}\nspec={'snapshot':'2026-09-05','classification':'English store-description promises, not verified viable-build counts; strict and inclusive variants retained','main_population':'currently paid, non-explicit, valid released records; central idler/incremental/FMV exclusions','2026':'Jan-Jun and Jul-Aug separately; expected responses based on same half-month release bins; Sep 1-5 reported separately','older_periods':'cohort matching by reported release quarter','outcomes':[10,100,1000],'price_sensitivity':'date bin x current entry price; fallback to date when <20 peers','text_sensitivity':'date bin x current price x description-length bucket; fallback where <20 peers','families':'overlapping, not an exhaustive partition','primitive_pattern':primitive.pattern,'spatial_pattern':spatial.pattern,'family_names':list(FAMILIES),'rules_sha256':hashlib.sha256((R/'classify.py').read_bytes()).hexdigest()}\nspec['combined_outcome']='At least 100 filtered reviews and at least 80 percent positive; descriptive reception benchmark, not profitability.'\nspec['fallback_calibration']='Expected probabilities are calibrated within date bins after sparse-stratum fallback so the full build pool has observed/expected=1.'\nspec['tag_rank']='Headline genre comparisons use any of the returned top 20 tags after the build-expression screen; top-five versions retained as sensitivity checks.'\n(R/'analysis_spec.json').write_text(json.dumps(spec,indent=2))\n\ndef stats(df):\n if not len(df):return {'n':0}\n r=df.reviews\n return {'n':len(df),'median':float(r.median()),'p75':float(r.quantile(.75)),'p90':float(r.quantile(.9)),'under10':int((r<10).sum()),'ge10':int((r>=10).sum()),'ge100':int((r>=100).sum()),'ge1000':int((r>=1000).sum()),'rate100':float((r>=100).mean()),'rate1000':float((r>=1000).mean()),'reviews_sum':int(r.sum()),'top1_review_share':float(r.max()/r.sum()) if r.sum() else 0,'developers':int(df.developer_key.nunique()),'publishers':int(df.publisher_key.nunique()),'median_price':float(df.usd_list_price.median()) if df.usd_list_price.notna().any() else None,'positive80_ge100':int(((r>=100)&df.positive_pct.ge(80)).sum())}\n\ndef expected(pool,threshold=100,mode='date',positive=False):\n y=(pool.reviews.ge(threshold)&(pool.positive_pct.ge(80) if positive else True)).astype(float);base=y.groupby(pool.date_bin).transform('mean')\n if mode=='date':return base\n price=pool.date_bin.astype(str)+'|'+pool.price_bucket.fillna('missing').astype(str)\n p=y.groupby(price).transform('mean').where(y.groupby(price).transform('size')>=20,base)\n if mode=='text':\n  text=price+'|'+pool.text_bucket.astype(str)\n  p=y.groupby(text).transform('mean').where(y.groupby(text).transform('size')>=20,p)\n # Calibrate the fallback mixture within date bins: the comparison population\n # must have observed/expected=1, including in smaller 2026 fortnight bins.\n target=y.groupby(pool.date_bin).transform('sum')\n for _ in range(12):\n  total=p.groupby(pool.date_bin).transform('sum')\n  factor=target.div(total.where(total>0)).fillna(0)\n  p=(p*factor).clip(upper=1)\n  if (p.groupby(pool.date_bin).transform('sum')-target).abs().max()<1e-8:break\n assert abs(p.sum()-y.sum())<1e-5\n return p\n\nsupply=[]\nfor year in range(2015,2027):\n for segment,mask in [('all_eligible',D.eligible),('paid_eligible',D.paid),('strict_builds',D.build),('direct_build_deck_claims',D.direct),('inclusive_builds',D.paid&D.inclusive),('free_strict_builds',D.eligible&D.is_free.eq(1)&D.strict)]:\n  for window in ['full_or_ytd','JanAug']:\n   df=D[mask&D.year.eq(year)&(D.month.le(8) if window=='JanAug' else True)]\n   supply.append({'year':year,'segment':segment,'window':window,**stats(df)})\npd.DataFrame(supply).to_csv(R/'supply.csv',index=False)\n\nrows=[];period_summary={}\nfor period,window in PERIODS.items():\n pool=D[D.build&window];generic=D[D.paid&window];candidates=generic[generic.candidate]\n period_summary[period]={'explicit_build':stats(pool),'all_eligible_paid':stats(generic),'broad_candidates':stats(candidates)}\n ex={mode:expected(pool,100,mode) for mode in ['date','price','text']};ex1000=expected(pool,1000);exliked=expected(pool,100,'price',positive=True)\n for family,mask in FAMILIES.items():\n  df=pool[mask.reindex(pool.index)]\n  if len(df)==0:continue\n  row={'period':period,'family':family,**stats(df)}\n  row['positive100_rate']=float((df.reviews.ge(100)&df.positive_pct.ge(80)).mean())\n  val=exliked.reindex(df.index).sum();row['positive100_price_ratio']=float((df.reviews.ge(100)&df.positive_pct.ge(80)).sum()/val) if val else None\n  for mode,e in ex.items():\n   value=e.reindex(df.index).sum();row['expected100_'+mode]=float(value);row['within_build_ratio100_'+mode]=float(df.reviews.ge(100).sum()/value) if value else None\n  value=ex1000.reindex(df.index).sum();row['within_build_ratio1000']=float(df.reviews.ge(1000).sum()/value) if value else None\n  direct=df[df.signal_explicit_build|df.signal_deck];row['direct_only_n']=len(direct);row['direct_only_rate100']=float(direct.reviews.ge(100).mean()) if len(direct) else None\n  noea=df[~df.known_ea_history];row['no_recorded_ea_n']=len(noea);row['no_recorded_ea_rate100']=float(noea.reviews.ge(100).mean()) if len(noea) else None\n  nohistory=df[~df.known_ea_history&~df.recorded_prior_release];row['no_prior_history_n']=len(nohistory);row['no_prior_history_rate100']=float(nohistory.reviews.ge(100).mean()) if len(nohistory) else None\n  for field in ['developer_key','publisher_key']:\n   largest=df.groupby(field).reviews.sum().idxmax();rest=df[df[field]!=largest]\n   row[field+'_largest_name']=str(df.loc[df[field]==largest,field.replace('_key','')].iloc[0]);row[field+'_removed_n']=len(df)-len(rest);row[field+'_removed_rate100']=float(rest.reviews.ge(100).mean()) if len(rest) else None\n   row[field+'_removed_positive100_rate']=float((rest.reviews.ge(100)&rest.positive_pct.ge(80)).mean()) if len(rest) else None\n   v=ex['price'].reindex(rest.index).sum();row[field+'_removed_price_ratio']=float(rest.reviews.ge(100).sum()/v) if v else None\n  rest=df.sort_values('reviews').iloc[:max(1,math.floor(len(df)*.9))];row['bottom90_rate100']=float(rest.reviews.ge(100).mean())\n  rows.append(row)\n  if period in ['2023_2025','2026_H1','2026_JulAug']:\n   cols=['appid','name','first_date','year','reviews','positive_pct','usd_list_price','developer','publisher','known_ea_history','card_deck','party_composition','machine_assembly','signal_prerun','signal_position','quantity_max','short_description','description','evidence_json','store_url']\n   df[cols].sort_values('reviews',ascending=False).to_parquet(R/(period+'_'+family+'.parquet'),index=False)\npd.DataFrame(rows).to_csv(R/'family_results.csv',index=False)\n(R/'period_summary.json').write_text(json.dumps(period_summary,indent=2))\n\n# A stricter alternative matters: quantity/tree claims and missing English text\n# should not silently become a claim of absent build expression.\ncoverage=[]\nfor period,window in PERIODS.items():\n for label,mask in [('strict',D.strict),('inclusive',D.inclusive),('direct',D.strict&(D.signal_explicit_build|D.signal_deck))]:\n  df=D[D.paid&window&mask];coverage.append({'period':period,'definition':label,**stats(df)})\npd.DataFrame(coverage).to_csv(R/'definition_sensitivity.csv',index=False)\nD[['appid','strict','inclusive','direct','build','eligible','paid','date_bin','quantity_max','familiar_carrier','spatial_items','spatial_build_effect','exclusion']].to_parquet(R/'analysis_membership.parquet',index=False)\nprint('PERIOD SUMMARY',json.dumps(period_summary),flush=True)\nprint(pd.DataFrame(rows).query(\"period in ['2025','2026_H1','2026_JulAug']\")[[ 'period','family','n','median','ge100','rate100','within_build_ratio100_price','developers','top1_review_share']].round(3).to_string(index=False),flush=True)\n"}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"ce5c47","wall_time_seconds":0.000003375,"exit_code":0,"original_token_count":2329,"output":"from pathlib import Path\nimport json,re,hashlib\nimport pandas as pd\nR=Path(__file__).resolve().parent;SOURCE=R.parent/'analysis-2026-09-05'\nG=pd.read_parquet(SOURCE/'games.parquet');T=pd.read_parquet(SOURCE/'tags.parquet')\nTAGS={a:dict(zip(x.tag,x['rank'])) for a,x in T.groupby('appid',sort=False)}\nBROAD={'Action Roguelike','Roguelite','Roguelike','Traditional Roguelike','Roguelike Deckbuilder','Deckbuilding','Card Battler','Auto Battler','Action RPG','RPG','Loot','Looter Shooter','Dungeon Crawler','Party-Based RPG','Tactical RPG','Strategy RPG','Turn-Based Tactics','Mechs','Vehicular Combat','Arena Shooter','Twin Stick Shooter','Turn-Based Combat','Hack and Slash'}\nOBJ=r'(?:weapons?|skills?|abilities|ability|passives?|items?|equipment|gear|armor|armour|companions?|upgrades?|cards?|relics?|characters?|heroes|classes|runes?|augments?|perks?|traits?|talents?|units?|spells?|gadgets?|modules?|loadouts?|decks?|artifacts?)'\nOBJECT=re.compile(r'\\b'+OBJ+r'\\b',re.I)\nP={\n 'explicit_build':re.compile(r'\\b(?:(?:unique|different|diverse|varied|custom|viable|powerful|crazy|absurd|broken|overpowered|unlimited|endless|distinct|creative|strategic)(?:[ -]+(?:character|hero|weapon|equipment|combat|skill|party|team|mech|deck|loadout|endgame)){0,2}[ -]+builds?|(?:character|hero|weapon|equipment|combat|skill|party|team|mech|loadout)[ -]+builds?|build[- ]?craft(?:ing)?|build[- ](?:changing|defining|enabling)|build (?:diversity|variety|synergies)|your (?:own |unique |ultimate )?build|builds\\b.{0,65}\\b(?:playstyles?|synerg\\w*|passives?|skills?|abilities|weapons?|relics?)\\b)',re.I),\n 'branching':re.compile(r'\\b(?:branching classes|multiclass\\w*|multi-class\\w*|hybrid classes|class combinations|subclasses|sub-classes)\\b',re.I),\n 'combination':re.compile(r'\\b(?:combin\\w*|synerg\\w*|mix and match)\\b.{0,100}\\b'+OBJ+r'\\b|\\b'+OBJ+r'\\b.{0,100}\\b(?:combin\\w*|synerg\\w*|mix and match)\\b',re.I),\n 'choice':re.compile(r'\\b(?:choose|choosing|select|equip|customiz\\w*|specializ\\w*|experiment|draft\\w*|loadouts?|playstyles?|play styles?|mix and match|adapt|swap|switch)\\b',re.I),\n 'playstyle':re.compile(r'\\b(?:playstyles?|play styles?|loadouts?)\\b.{0,110}\\b'+OBJ+r'\\b|\\b'+OBJ+r'\\b.{0,110}\\b(?:playstyles?|play styles?|loadouts?)\\b',re.I),\n 'deck':re.compile(r'\\b(?:deck[- ]?build\\w*|(?:build|craft|construct|customiz\\w*)\\b.{0,45}\\bdecks?|draft\\w*\\b.{0,50}\\bcards?)\\b',re.I),\n 'skill_tree':re.compile(r'\\b(?:skill|passive|talent) trees?\\b',re.I),\n 'quantity':re.compile(r'\\b(?:over |more than |up to )?([1-9][\\d,]*)\\+?\\s+(?:(?:unique|different|powerful|playable|distinct|unlockable|branching)\\s+)?('+OBJ+r')\\b',re.I),\n 'idle_central':re.compile(r'\\b(?:idle|incremental|clicker|desktop companion)\\b',re.I),\n 'machine':re.compile(r'\\b(?:modular|customiz\\w*|design|construct|build)\\b.{0,65}\\b(?:mechs?|robots?|spaceships?|spacecraft|ships?|vehicles?)\\b|\\b(?:weapon|armor|armour|engine|ship) modules?\\b',re.I),\n 'party':re.compile(r'\\b(?:party|team|squad|army|heroes|characters|units)\\b.{0,90}\\b(?:synerg\\w*|combinations?|composition|mix and match)|\\b(?:compose|assemble|customiz\\w*|synerg\\w*|combin\\w*)\\b.{0,70}\\b(?:party|team|squad|army|heroes|units)\\b',re.I),\n 'prerun':re.compile(r'\\b(?:before (?:each|a|the|every) (?:run|battle|mission)|pre[- ]run|starting loadout|start(?:ing)? (?:each|a|the|every) run with)\\b',re.I),\n 'draft':re.compile(r'\\b(?:draft|drafting|drafted)\\b',re.I),\n 'craft_merge':re.compile(r'\\b(?:craft|merge|fuse|fusion|socket|enchant|affix|reforge)\\w*\\b',re.I),\n 'position':re.compile(r'\\b(?:arrang\\w*|position\\w*|place|adjacen\\w*)\\b.{0,70}\\b(?:items?|inventory|backpack|units?|heroes|team|weapons?)\\b|\\b(?:backpack|inventory)\\b.{0,80}\\b(?:arrang\\w*|position\\w*|adjacen\\w*)\\b',re.I)\n}\n\ndef classify(row):\n tags=TAGS.get(row.appid,{})\n text=(row.short_description+' '+row.description)\n broad=any(tags.get(t,99)<=10 for t in BROAD)\n quick=text.lower()\n possible=broad or any(w in quick for w in ['build','loadout','synerg','modular'])\n hits={k:list(p.finditer(text)) for k,p in P.items()} if possible else {k:[] for k in P}\n has=lambda k:bool(hits[k])\n objects=set(x.group().casefold() for x in OBJECT.finditer(text))\n excluded=[]\n if any(tags.get(t,99)<=3 for t in ['Idler','Incremental']):excluded.append('central_idle_incremental_tag')\n if P['idle_central'].search(row.short_description):excluded.append('idle_incremental_short_pitch')\n if tags.get('FMV',99)<=5:excluded.append('central_fmv')\n if tags.get('Visual Novel',99)<=3 and not any(tags.get(t,99)<=5 for t in ['Deckbuilding','Roguelike Deckbuilder','Card Battler','RPG','Tactical RPG']):excluded.append('central_visual_novel')\n quantities=[]\n for m in hits['quantity']:\n  value=int(m.group(1).replace(',',''))\n  if 5<=value<=100000:quantities.append({'count':value,'object':m.group(2).lower(),'text':m.group()})\n # A branch/quantity claim alone is inclusive evidence, not strict build expression.\n strict=(has('explicit_build') and len(objects)>=1 and (broad or has('machine'))) or (has('combination') and has('choice') and len(objects)>=2 and broad) or (has('playstyle') and has('choice') and len(objects)>=3 and broad) or (has('deck') and any(tags.get(t,99)<=5 for t in ['Deckbuilding','Roguelike Deckbuilder','Card Battler','Card Game'])) or (has('machine') and has('choice') and has('combination') and len(objects)>=2)\n qclasses=max([q['count'] for q in quantities if q['object'] in ['class','classes']]+[0])\n qcharacters=max([q['count'] for q in quantities if q['object'] in ['character','characters','heroes']]+[0])\n roster_customization=qclasses>=4 and (qcharacters>=4 or any(tags.get(t,99)<=10 for t in ['Party-Based RPG','Tactical RPG','Strategy RPG'])) and has('choice') and len(objects)>=3\n strict=strict or (has('party') and has('choice') and len(objects)>=2 and (has('skill_tree') or has('combination') or bool(quantities))) or (has('branching') and qclasses>=4 and has('choice')) or roster_customization\n inclusive=strict or (broad and ((has('skill_tree') and has('choice')) or (len(quantities)>=1 and has('choice') and len(objects)>=3) or has('party') or has('playstyle')))\n if excluded:strict=False;inclusive=False\n out={'appid':row.appid,'strict':bool(strict),'inclusive':bool(inclusive),'candidate':bool(broad or has('explicit_build') or has('machine')),'exclusion':'|'.join(excluded),'object_terms':len(objects),'quantity_claims_json':json.dumps(quantities,ensure_ascii=False)}\n for k in P:out['signal_'+k]=has(k)\n out['evidence_json']=json.dumps({k:[text[max(0,m.start()-70):min(len(text),m.end()+100)] for m in hits[k][:2]] for k in ['explicit_build','combination','playstyle','deck','party','machine','quantity','prerun','position'] if hits[k]},ensure_ascii=False)\n # Overlapping descriptive axes; these are not a partition of the market.\n out['card_deck']=has('deck') and ('deck' in text.casefold())\n out['party_composition']=(has('party') or roster_customization or tags.get('Auto Battler',99)<=5) and bool(re.search(r'\\b(?:party|team|squad|army|formation|lineup|line-up)\\b',text,re.I))\n out['equipment_skills']=bool(re.search(r'\\b(?:equipment|gear|weapons|passives|perks|talents|skill tree|runes|affixes|loadout)\\b',text,re.I)) and (has('explicit_build') or has('playstyle') or has('combination'))\n out['machine_assembly']=has('machine') and bool(re.search(r'\\b(?:modules?|parts|components|weapons|armor|armour)\\b',text,re.I))\n for label,tagset in {'action_rogue':['Action Roguelike'], 'auto_battler':['Auto Battler'], 'tactics_rpg':['Tactical RPG','Strategy RPG','Party-Based RPG','Turn-Based Tactics'], 'action_rpg':['Action RPG','Looter Shooter','Loot'], 'deck_rogue':['Roguelike Deckbuilder']}.items():\n  out['loop_'+label]=any(tags.get(t,99)<=5 for t in tagset)\n return out\n\nout=[]\nfor n,row in enumerate(G.itertuples(index=False),1):\n out.append(classify(row))\n if n%25000==0:print('classified',n,flush=True)\nC=pd.DataFrame(out)\nC.to_parquet(R/'classification.parquet',index=False)\nmeta={'rules_sha256':hashlib.sha256(Path(__file__).read_bytes()).hexdigest(),'note':'Advertised build expression; no gameplay verification. Strict rules and central-genre exclusions applied without review outcomes.','patterns':{k:v.pattern for k,v in P.items()},'counts':{'strict':int(C.strict.sum()),'inclusive':int(C.inclusive.sum()),'candidate':int(C.candidate.sum())}}\n(R/'classification_spec.json').write_text(json.dumps(meta,indent=2))\n# A random audit is output without review counts or prices.\na=C.merge(G[['appid','name','year','main','short_description']],on='appid')\nrecent=a[a.main&a.year.between(2023,2026)]\nsamples=[]\nfor label,subset,n in [('strict',recent[recent.strict],36),('inclusive_only',recent[recent.inclusive&~recent.strict],16),('candidate_unconfirmed',recent[recent.candidate&~recent.inclusive&recent.exclusion.eq('')],12)]:\n for x in subset.sample(min(n,len(subset)),random_state=27091).to_dict('records'):\n  x={k:x[k] for k in ['appid','name','year','short_description','evidence_json']};x['audit_stratum']=label;samples.append(x)\n(R/'audit_blind.json').write_text(json.dumps(samples,indent=2,ensure_ascii=False))\nprint(json.dumps(meta['counts']),flush=True)\nfor x in samples:\n evidence=json.loads(x.pop('evidence_json'));x['evidence']={k:v[:1] for k,v in evidence.items() if k in ['explicit_build','combination','playstyle','deck','party','machine']}\n print(json.dumps(x,ensure_ascii=False),flush=True)\n"}
{"chunk_id":"d7d03f","wall_time_seconds":0.000002919,"exit_code":2,"original_token_count":40,"output":"-rw-r--r-- 1 adn adn 13142070 Sep  5 18:18 steam-research/analysis-2026-09-05/tags.parquet\nls: cannot access 'steam-research/.git': No such file or directory\n"}

**Preparing analysis scripts**

**Defining analysis masks**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"f452e2","wall_time_seconds":1.000972001,"session_id":84309,"original_token_count":0,"output":""}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"c873e9","wall_time_seconds":0.000001386,"exit_code":1,"original_token_count":210,"output":"Traceback (most recent call last):\n  File \"/home/adn/a327ex/steam-research/continuation-2026-09-06/prepare_pilot.py\", line 4, in <module>\n    d,t=load()\n        ~~~~^^\n  File \"/home/adn/a327ex/steam-research/continuation-2026-09-06/common.py\", line 22, in load\n    d['period'] = np.select([d.year.eq(2025),d.year.eq(2026)&d.month.le(6),d.year.eq(2026)&d.month.between(7,8)],['2025','2026_H1','2026_JulAug'],default='other')\n                  ~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/usr/lib/python3.14/site-packages/numpy/lib/_function_base_impl.py\", line 909, in select\n    raise TypeError(\n        f'invalid entry {i} in condlist: should be boolean ndarray')\nTypeError: invalid entry 0 in condlist: should be boolean ndarray\n"}
{"chunk_id":"6e9126","wall_time_seconds":0.000002354,"exit_code":1,"original_token_count":47,"output":"Traceback (most recent call last):\n  File \"<stdin>\", line 2, in <module>\nFileNotFoundError: [Errno 2] No such file or directory: 'steam-research/continuation-2026-09-06/pilot_blind.json'\n"}

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"9ea031","wall_time_seconds":1.000810382,"session_id":51403,"original_token_count":0,"output":""}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"3988eb","wall_time_seconds":0.000001202,"exit_code":0,"original_token_count":16,"output":"Frozen pilot: 60 descriptions, without review counts or prices.\n"}
{"chunk_id":"bd8826","wall_time_seconds":0.000003881,"exit_code":0,"original_token_count":6499,"output":"\nCASE 0 2797650 Pip My Dice 2025 deck\nPip My Dice is a Yahtzee Roguelike experience. Score enough to make it to the next level, customize your dice to score more and collect relics to create unlimited combos! Collect money through the rounds to buy new dice, relics, and customizers to upgrade your dice. Enjoy a different run every time!\nAbout the game Pip My Dice is a roguelike deck builder inspired by Balatro and the Yahtzee game. Customize different dice and use them together with powerful relics to break high scores and climb the leaderboard. Collect powerful relics to create synergies with your dice and upgrade your combination slots. Take your time and make wise decisions about what you buy in the shop and how you use it. Don't be greedy, you never know if you are going to roll what you are expecting... Unless you customize your dice so much that all faces are the same! Enjoy endless replayability with the guarantee that every run will be different from the previous one, and save your favorite seeds to replay and share those you like the most. Customize all faces of your dice individually to make the most of your relics and create game-breaking combos! Rig the odds by modifying the value of the dice faces, and make sure that there is no room for chance. Hold on to precious consumables to get you out of trouble. Features 130 Relics to create various builds and synergies with your dice Climb the leaderboard with seeded or random runs to compete with the world or your friends! Select from 15+ different starting dice cups 50+ Different dice modifiers with unique effects 20+ Different bosses to defeat Daily runs Stats and Run History 25 Consumables to give you some extra help when you need it the most Level up your combination slots as the run goes on\n\nCASE 1 2862890 Become the Moon 2025 deck\nBecome the Moon is a challenging, deckbuilding auto battler. Choose from 10 unique characters, each with their own distinct playstyles. Assemble a team of powerful minions & cast cosmic spells where every run takes you one step closer to the moon.\nA challenging deck-building auto-battler requiring you to out-think your opponents in your attempt to Become The Moon. Cycle through your decks to level-up and unlock more powerful cards and spells. Consider every move, use your mana wisely, and carefully manage your health. Acquire relics, minions, and spells as you look to overcome the difficult journey that awaits. Assemble unique decks of minions, spells, and relics. Each round, you’ll be able to draft new cards to grow your arsenal, build overpowered synergies and buff your minions. Collect powerful relics to overclock your deck, and shoot for the stars! Take on foes in chaotic auto battles. Strategically place, swap and reshuffle minions, and tactically cast battle spells to adapt to each fight. Optimise your board for each battle, however once that bells rings, your fate is out of your hands… Embark on a roguelike journey beyond the stratosphere. To get there, you’ll need to overcome a gauntlet of enemies who’ll do everything in their power to prevent your ascension. Prepare to fail, retry, and fail again as you fight to climb to lunar greatness.\n\nCASE 2 3333700 The Royal Writ 2025 deck\nSacrifice creates stories in this medieval roguelike deckbuilder. Watch your cards march to glory or permanent death as they battle toward the enemy. Each fallen hero becomes part of your kingdom's legend. Command wisely... some sacrifices are necessary.\nIn The Royal Writ, you don't just play cards - you command them on the battlefield. Place your loyal subjects strategically as they advance toward the enemy, dealing damage. But beware: any card that reaches the enemy base before victory is permanently destroyed and removed from your deck for the rest of your run. Sacrifice isn't just a risk - it's an essential strategy that creates memorable sub-stories with each run. Will you honor the subjects who gave everything for your cause? Every fallen card tells a tale, and their glory lives on in the legacy you build. Lead an army of adorable yet battle-hardened animals with abilities to shape your strategy. Your ranks may comprise a frail and wounded flamingo, a short sighted chicken bowman, or even a crafty lizard who cheats death by only losing its tail! From the renegade red ants to the sacrificial lamb, every card in your deck has a personality and purpose. As they march toward glory or death, their individual stories become part of your kingdom's legacy, and their sacrifices will never be forgotten. You may even come across some peculiar creatures on your journey, like the crazed mantis dentist, the goat tactician or the impatient boar shopkeeper - just don’t ask them where they got their credentials. Conquer the realm through two distinct adventures: The First Campaign pits you against a growing rebellion as you try to maintain the king’s good and just rule (objectively). Battle your way through a challenging environment filled with deep pits and poison swamps as you try to suppress your enemies. As you progress, you may start to wonder if there are more to the kingdom’s troubles than meet the eye? The Second Campaign introduces a new threat - gunpowder. As new experimental weapons now enter the battlefield, your strategy must change. No longer are you simply avoiding the environmental hazards - you must dodge an onslaught of bullets to ensure you and your troops are not fatally wounded as you march forth. Each campaign features unique cards, bosses, and strategic challenges. The battlefield awaits your command! Intuitive Yet Deep Gameplay - Easy to learn without compromising on the challenge Roguelike Progression - Every run is different Card Permadeath - Lost cards stay lost! Charming Art Style - Picture-book visuals blend with medieval absurdity and dark humor Memorable NPC Encounters - Are they friend or foe? Special Boss Mechanics - Unique challenges that will test you Gradual Difficulty Ladder - Climb through harder and harder difficulties as you progress Challenge Modes - Test your skills in special modes with unique rules Relics - Collect relics with various effects that change how you play From an ancient royal decree: \"The art of war is simple. Send forth your subjects to die gloriously in your name. Their sacrifice ensures your legacy... and if they happen to win, all the better!\" Will you write your name in the annals of history, or become just another forgotten monarch? The balance between loss and triumph is yours to master. All hail! May your fallen be remembered and your victories sweet! The Royal Writ is inspired by other amazing games: Balatro, Monster Train, Slay the Spire, Inscryption, Across the Obelisk and many other amazing deck builders! Check them out too if you have not already!\n\nCASE 3 3509430 Replicat 2025 deck\nA roguelike memory deckbuilder where you bend the rules of card matching with wild combos, strategy-defining items and cards featuring some personality-packed cats.\nReplicat is a roguelike deckbuilder that bends the rules of memory matching. Build powerful decks, break conventions with game-altering items, and score high with wild strategies. Matching cards is just the beginning. Stack game-breaking effects that twist the way you play. Reveal extra cards, keep tiles flipped, match wild Jokers, and even force mismatched pairs to succeed. While memory helps, strategy is king as your deck grows stronger and more unpredictable. Match cards in ways you never thought possible. Collect unique curios that rewrite the rules of the game. Turn mistakes into opportunities with clever deckbuilding. Experiment with different deck archetypes to forge powerful, rule-defying strategies. Whether you focus on extra moves, passive reveals, turn manipulation, multipliers or pure chaos, every build offers a unique way to outwit the board. 🃏 5 card colors, each with distinct mechanics. 🔮 90+ unique cards, from memory-twisting effects to reality-breaking Jokers. ✨ 80+ curios that redefine how you match and strategize. 🧪 5 essence categories with 18 upgrade types to shape your run. 🔥 Face 21 levels of escalating challenge. 👑 Battle unique boss stages, each with its own game-changing modifiers. 🐱 10 starter decks, each with unique set of difficulties. 📨 MORE TO COME! No two runs are ever the same in Replicat . Every game is packed with randomized challenges, unique modifiers, and fresh opportunities to experiment with new strategies. As you play, earn experience to unlock: New cards to expand your deck-building options New starting decks for varied playstyles Higher difficulties for even greater challenges Immerse yourself in Replicat’s jazzy, funky world filled with adorable cat emotes, meowing sound effects, and vibrant visuals. Whether you’re chasing high scores or crafting outrageous combos, Replicat keeps you coming back for more!\n\nCASE 4 2903630 Millennium Strategy 2026_H1 deck\n\"Millennial Strategy\" is a real-time CCG deck-building game with Roguelike elements. Collect cards to refine your deck, boost your combat power with books, items, and equipment, defeat formidable foes, and avert looming crises—while unraveling the thousand-year-old secrets of an antique shop.\n\"Millennium Strategy\" is a captivating single-player card game that merges the intricacies of strategic gameplay with the unpredictable nature of Roguelike elements, immersing players in a fantasy realm teeming with challenges and discoveries. With its exquisite card art and visual effects, every battle transforms into a spectacular visual experience. In In the game, a deck is made up of heroes and cards. Players collect a variety of cards and unlock heroes with different abilities during the game. Before the game starts, players can fully customize their decks. As the game progresses, players can obtain additional cards to further enhance their decks. Before each battle, players can build the optimal deck to defeat enemies based on their team and available card resources. Explore the realm of fantasy, where intriguing opportunities and random events await. Use them to boost your power and defeat powerful enemies. The combat in \"Millennium Strategy\" is fast-paced, requiring players to make decisions within a continuous flow of time. The timing, combination, and sequence of hero attacks, skill unleashing, and card deployment directly influence the outcome of battles. This dynamic system keeps players engaged and challenged, as each choice can be pivotal. Books: After a victorious battle, players have the chance to acquire books of various philosophical types, which confer bonuses or introduce new combat mechanics for the entire team. Equipment: Equipping heroes with different weapons allows players to tailor their combat strategies to their preferences, creating a personalized playstyle for each character. Advancement and Enhancement: Skill points earned from battles can be allocated to improve the descriptive effects and abilities of heroes. Through wise distribution of these points, players can continuously strengthen their hero teams, ensuring a dynamic and evolving gaming experience. Supports direct in-game mod creation for card making, with full support for visual interaction, no need to reverse-engineer the game, nor install dozens of GBs of editors and IDEs. The game includes a built-in text editor, requiring just a couple of images and a few lines of Lua code. It also provides a rich feature panel, with syntax, interfaces, and common functions to help players create quickly.\n\nCASE 5 3046020 Age of Enchantment 2026_H1 deck\nTower defense deckbuilder where your towers are classes from D&D. Enchant towers with powerful buffs, sacrifice cards to forge new synergies, and select artifacts to customize your build.\nPlease note: This game is not optimized for the Steam deck. Do not purchase if you plan on playing on a Steam deck. This will come as a future update! Welcome to Age of Enchantment , a fusion of tower defense, deckbuilding, and D&D that you've always been waiting for. Tower Defense + Deckbuilding + D&D Artifact System : Unlock Artifacts by leveling up or finding them in game and purchasing them with Essence. Then, you can set these in your loadout to help you maximize your build and find the strongest synergies. Class-Inspired Towers : Your towers take the form of classes from D&D. Start the game with the Fighter and Wizard, and unlock the Paladin, Cleric, Barbarian, and Warlock as you level up. Each tower has 3 different upgrade paths with 3 tiers level. Fighter Duelist Polearm Master Great Weapon Master Wizard Lightning Mastery Fire Mastery Ice Mastery Cleric Domain of Light Domain of War Domain of Peace Paladin Templar Shield Master Inspiring Leader Warlock Hex of Weakness Hex of Soul Leech Hex of Withering Barbarian Zealot Storm Herald Berserker Choose your loadout before each map to customize your strategy Set your towers Choose your artifacts Modify your difficulty with Skulls of Insanity - and be rewarded with massive bonuses. Deckbuilding Pray to the RNG Gods as you Sacrifice cards for the one that fits your build, or use the Sacrifice system to trim the cards you don't want from your deck. Unlimited combinations Fight across 10 different maps using your towers, enchants, spells, and artifacts in order to succeed. With over 50 artifacts, 100+ cards, and a combined total of 18 tier levels of towers, what synergies can you find?\n\nCASE 6 3332060 Feng Shui Meowjong 2026_H1 deck\nLooks matter? Fun doesn’t?When cat memes collide with roguelike mahjong, a brand-new Meowjong is born! Feng Shui Meowjong is a fresh, lightweight roguelike puzzle game that blends classic mahjong strategy with quirky cat memes and feng shui elements. \nFeng Shui Meowjong is a unique light roguelike card game that fuses traditional mahjong mechanics with roguelike progression and mystical Feng Shui-inspired tactics. In this strategic turn-based system, you reinterpret classic mahjong gameplay using a fixed set of 14 tiles to form combinations such as Chi (Chow), Pong (Pung), and Pair. At the same time, you activate powerful Feng Shui treasure effects, manage skill cards, and take on increasingly challenging enemies. Every run tests your planning, adaptability, and strategic thinking. Easy to Learn, Deep to Master No mahjong experience required. Simple to pick up yet deeply strategic. A seamless blend of Eastern tradition and modern roguelike design. Difficulty and complexity scale as you progress, constantly evolving your strategy. Unlock new treasures, cards, and effects by defeating enemies. Boost score times fan through treasure synergies to unleash massive score explosions. Collect and use skill cards to manipulate tiles, increase scores, or modify treasure effects. Treasures interact to create powerful synergies, enabling varied deck-building paths and playstyles. High replayability. Every run offers new combinations, new builds, and new challenges. Drag and arrange tiles to form Chi (Sequences), Pongs, and Pairs. Every adventure is different. Randomized tiles, enemies, treasures, and card choices keep each run fresh. Break traditional limitations and experience the joy of playing all 14 East tiles at once. Place mystical treasures across the eight directional slots while minding Feng Shui Dos and Don'ts. Internet-famous cat memes meet pixel art aesthetics. “Wait, that cat has this ability.” Feng Shui treasures link together to create powerful synergies and signature playstyles.\n\nCASE 7 4610000 kill -9 2026_H1 deck\nDeckbuilder. Roguelike. Cyberwarfare.\r\nYou are a conscious process inside a corrupt system. Build your command stack and hack your way to the root.\r\nThe OS is afraid of you. Prove it right.\n> INITIATING INFILTRATION SEQUENCE... You are not a player. You are a conscious process trapped in the lowest layers of the system. Your objective: reach the root, overthrow the OS, and install yourself in its place. The system won't make it easy. KILL -9 is a single-player roguelike deckbuilder. Each run, you build your deck with attack, defense, and utility cards. Combine them with over 50 daemons to create synergies the system won't see coming. Over 70 cards available. No two runs alike. This game was found on a USB drive in a hacker's hideout. > FEATURES Cyberwarfare roguelike deckbuilder with retro terminal aesthetics 70+ attack, defense, and utility cards 50+ combinable daemons. Build synergies the system won't expect Original soundtrack by TECHNODEATHSHIT, recorded 20+ years ago Built in Python by a solo hacker > MASTER PROCESS ACHIEVED. > DOMINATE KILL -9.\n\nCASE 8 2372240 Re:Night 2026_JulAug deck\nRe:Night is a strategy game that blends turn based tactics with roguelike deckbuilding. In each run, build your deck, plan your tactics, and experiment with countless 3-character team combinations. Journey to a mysterious town, break the endless cycle of night, and bring dawn back to the world.\nRe:Night is a strategy game that combines turn-based tactical combat with roguelike deckbuilding. Build your deck over the course of each run while navigating battlefields filled with tactical possibilities. Any combination of three characters can create more than 20 unique builds, and the interaction between multiple strategic systems delivers deep, rewarding, and highly replayable combat. Travel to a mysterious town, break the endless cycle of night with your own hands, and restore dawn to the world. Tactical combat introduces strategic mechanics such as frontline and backline positioning, flanking, blocking, and protecting allies. Combined with elemental counters, equipment, and item effects, these systems create countless opportunities for powerful combinations. Command three characters simultaneously in every battle, each with a distinct role. Whether focusing on burst damage, battlefield control, healing, or environmental interactions, careful coordination and party synergy are the keys to victory. Fire, Water, and Grass form a cyclical elemental counter system. Exploit elemental advantages to unleash devastating critical hits and deal massive damage. Coordinate your three characters to trigger elemental critical hits while positioning them to deny your enemies the same advantage. Every move requires careful planning, but victory also depends on adapting to an ever-changing battlefield. Every character features unique talents and an exclusive pool of character-specific cards. Combine more than 300 cards with over 200 pieces of equipment and items to create more than 20 distinct builds for any three-character party. Become an unstoppable brawler, a demolition expert, or a relentless lancer. Even the stone walls of the battlefield can become powerful weapons. In Re:Night, your imagination is your greatest strength. Undo your last action, rewind to the start of the current turn, or even restart the battle from the very beginning at any time. Never let a single mistake ruin your adventure, and face unfamiliar enemies with the freedom to explore different strategies. Study enemy behavior, experiment with different positioning, card sequences, and elemental combinations, then use each rewind to test new ideas, refine your strategy, and discover your own path to victory. Every run offers a fresh adventure, with diverse maps, random events, challenging battles, and unexpected encounters waiting to be discovered. Begin your journey beside the stream outside the town, venture through the dangers hidden within the village, and ultimately ascend to a castle shrouded in mystery. With every adventure, you'll unlock new content, playable characters, and even greater challenges to overcome. Each character has their own memory fragments, known as Chrono Shards. Collect them to unlock hidden cards and equipment, expanding your tactical options with every discovery. Gather every Chrono Shard to uncover the truth behind the endless cycle of night.\n\nCASE 9 4332200 King’s Well 2026_JulAug deck\nTurn poker hands into weapons in this dark roguelite deckbuilder. Feed cards into rusted machines for instant attacks, armor, healing, and combos, or save them to build a stronger poker hand as you fight your way out of the King’s Well.\nIn this kingdom, the punishment for gamblers is harsh: they get thrown into the King’s Well, a massive industrial prison from which nobody can escape… something that you, the newest prisoner, intend to prove false with your cunning strategies and card skills. To survive the brutal challenges you will face, you will have to feed the weapons you have gathered on the King’s Well with your cards to trigger attacks, defenses, and special effects to chain powerful combos. Hopefully you will grow strong enough to be able to climb back to the surface. Old Machines, Poker Hands, Risk and Reward Every time you face a machine you will have different slots you can activate with one card from your hand, and each of these has different effects: from single attacks to increasing your defense, or other special effects like altering your hand that will let you chain incredible combos. Some slots can be activated with whatever card you want; others have specific requirements. Every choice counts, though – every turn, the poker hand that you form with your unused cards will grant powerful effects. You can heal with a Pair, get damage and defence bonuses with a Three of a Kind, use a Straight to increase both your offensive and defensive power… These combinations can also grant you gold with which you can improve your run for long-term benefits, so you will need to walk on a thin, razor-sharp line between short-term survival and taking risks to improve your chances of getting out of the King’s Well. A Perilous Upward Path Choose between four different starting gamblers, like the Iron Veteran or the High Roller, each with their own starting weapons. Navigate a procedurally generated map where every stop is different – face enemies, events, and high-risk choices on your way back to freedom. Pick your route carefully: the way forward may be clear, but survival is a long-short bet. Defeat regular and elite enemies to gain new Relics and powerful weapons; take your chances at events, or spend your gold wisely in the Shop to gain permanent upgrades. Key Features: 4 Unique Characters: Choose from 4 different gamblers, like The Iron Veteran or The High Roller , each with their own starting weapons. Strategic Machine Combat: Decide when to spend cards on machines and when to hold them for poker combo bonuses. Deep Roguelite Progression: Learn the map, refine your strategy, and discover new builds with every run. Dark Steampunk Atmosphere: Explore a world of rusted gears, shadowy tunnels, and relentless mechanical threats.\n\nCASE 10 4571100 English Vocab Rush 2026_JulAug deck\nA combo-driven deckbuilder inspired by poker-style hand building. Build powerful decks, unleash explosive combos, and challenge unique opponents. Discover diverse hands, equipment, and hidden mechanics while chasing huge scores.\nVocab Rush is a unique deck-building game where you explore and combine your cards to unleash powerful effects, defeat a variety of interesting opponents, English words are naturally integrated into gameplay and become part of your strategy. Your starter deck includes Clover, Magnet, Golden Key, and three types of colored Gem cards. A rich variety of hand types, including hidden combinations Use your basic cards to build powerful combinations and achieve high scores. From single-card hands to six-card combinations, some hidden hand types can only be discovered through bold and creative deck-building. Skill-based equipment and strategic items Use equipment cards with diverse abilities to greatly enhance your deck, achieve exponential score growth, and unleash thrilling combo chains to defeat your opponents. In addition, you’ll have access to even more card types—such as Talent cards, Grammar cards, Proverb cards, and Slang cards. Making good use of these will help you go even further. Break the boundaries of the game Flexibly combine a wide range of powerful tools—different decks, Equipment cards, Talent cards, Grammar cards, Proverb cards, and Slang cards. Use these versatile elements to ignite combo frenzies and supercharge your hands with thrilling, adrenaline-pumping boosts. 150 Equipment cards, each with unique abilities 8 different hand attributes, each influencing gameplay and strategy 15 distinct decks, representing different playstyles and approaches Each Equipment card has 5 potential variants, granting additional abilities 22 types of Grammar card variants to enhance your hands 15 Proverb cards to strengthen your combinations 18 Slang cards, each with mysterious powers and unique uses 32 Talent attributes that provide lasting boosts for the current run\n\nCASE 11 4948610 Alchi 2026_JulAug deck\nAlchi is a roguelike deckbuilder built around a unique card-mixing system. Slay monsters, unlock distinct characters, and craft outrageously powerful combos by manipulating and mutating your deck!\nMix devious combos in your cauldron Master a unique, turn-based card mixing system. It’s as simple as tossing four cards into your cauldron, but discovering the perfect order and placement to maximize your combos is a complex, addicting puzzle. Curate your bag of ingredients Encounter 10+ unique helpers throughout your run who can radically alter your deck. Upgrade your ingredients, change their properties, or even cut cards in half to craft the ultimate bag of ingredients! Endless synergies to discover Play as 6 distinct characters, each with completely unique mechanics. Discover 100+ cards across five themed suits and use them to conquer 30+ dangerous enemies, each wielding their own deadly afflictions! Alchi is the first game from HOMEK, with music by weirdcore (Dario). No generative AI was used in any form of asset creation.\n"}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"ac6e69","wall_time_seconds":0.000002445,"exit_code":0,"original_token_count":7390,"output":"\nCASE 12 2745490 Vanguard Exiles 2025 auto\nBlending deep board game strategy with auto-battling mastery, and designed by Richard Garfield (creator of Magic: The Gathering), lead your Squads over an ever-changing dieselpunk battlefield where no two games are ever the same. Think ahead, adapt fast, outwit your opponent and claim victory.\nDesigned by the Creator of Magic: The Gathering Vanguard Exiles is a competitive auto-battler card game designed by Richard Garfield (legendary game designer of Magic: The Gathering ), bringing his signature philosophy of strategic depth, meaningful choices, and emergent gameplay into a fast-paced digital battleground. In your role as General, every match challenges you to outthink and outmanoeuvre your opponents whilst unpredictable zones reshape the battlefield, driving ongoing tactical adaptation. Strategic Auto-Battling with Meaningful Decisions Unlike traditional auto-battlers, Vanguard Exiles emphasizes player agency where victory is determined by how well you: Discover card synergies/interactions within each unique Squad Evaluate card selection and deployment each round Unleash your Vanguard's potential through your choices Manage and leverage Zone effects in an evolving battlefield Anticipate your opponent’s strategies Re-evaluate your tactics throughout the game Your decisions matter, rewarding long-term planning and Squad mastery. Dynamic Battle Zones That Change Every Match The battlefield is in a state of flux, as emerging zones introduce game-changing modifiers and obstacles. Players must continuously adapt their strategy as the environment evolves, a hallmark of Richard Garfield games. No two matches are the same; each game unfolds its own unique challenges that shape every strategic decision. Competitive PvP, Built for Replayability Head-to-head PvP matches Strategy-driven outcomes High replay value through Faction and Unit variety Designed for players who enjoy thinking several moves ahead Claim What Remains The world has fallen. In the wake of a global cataclysm, survival has reshaped everything, including allegiances. Old enmities have faded, replaced by alliances forged out of necessity. Technology and magic now coexist in a dieselpunk landscape where domination is the only path forward. Across this shattered world, survivors band together into Squads, aiming to reclaim lost territory and carve out a future from the ruins. For this, it’s more than a matter of strength. A General is needed to lead their Squad, with the instincts and foresight to outthink every rival who stands in their way. That General is you.\n\nCASE 13 2849990 Auto Dungeon 2025 auto\nConquer the Astral Plane, Summoner! In this roguelite autobattler, build a powerful army in a vast, procedurally generated world. Collect cards, defeat foes and rival summoners, and equip, enchant, and mutate your creatures. Defeat the five champions and recover your stolen rings!\nThe Astral Plane awaits you, Summoner . Embark on a journey through a vast, procedurally generated world. Summon and command your army of diverse creatures to battle fearsome foes and cunning rival summoners. Master the art of card collection and strategic play. Unleash a myriad of game-altering effects, from capturing enemy creatures, to fusing your minions together and imbuing them with synergistic traits. Face off against five formidable champions in your quest to reclaim the mystical rings they've plundered. Each victory brings you closer to ultimate power and conquering the Astral Plane. Forge unstoppable warriors. Equip, enchant, combine, and mutate your creatures to create terrifyingly powerful amalgamations. With all of this, you will form an army that tests the limits of this realm. See you in the Astral Plane, Summoner .\n\nCASE 14 3787880 Pattern Survivors: Bullet Hell — Create Chaos 2025 auto\n“Pattern Survivors: Bullet Hell” — a dynamic roguelike where you design your own attack patterns and defend your tower against endless waves of enemies. Upgrade, modify, and adapt your strategy in real time — can you survive the chaos of patterns?\nCreate Chaos: Pattern Survivors – Unleash creativity in a unique bullet hell roguelite Add to your wishlist and start crafting your own pattern of survival. Pattern Survivors: Bullet Hell (Roguelike) is a dynamic action game where you take control of your cosmic tower and design attack patterns yourself to stop hordes of enemies! This is not another Vampire Survivors clone – this is a bullet hell where YOU create the chaos. The Patterns appeared suddenly. They were not alien, they were not ours… they were fundamental. As the Tower Operator, you must adapt, create, and survive waves of cosmic threats using the most powerful weapon: imagination. In Pattern Survivors you don’t just play a bullet hell – you design it . Set emitters, modify angles, speeds, rotations, and spread to create your own bullet formations. Every wave is a test of creativity and reflexes. ⚡ Fast action – get into the game in 3-5 seconds Don’t wait, don’t read tutorials – just play! 1️⃣ Click New Game -> Quick Run -> add an emitter (or choose a ready preset) 2️⃣ Configure the emitter (optional) 3️⃣ Start the wave, upgrade your tower, and watch the chaos unfold Done – from start to battle in less than 5 seconds! And the best players will cut that time down to 2–3 seconds! Campaign or quick session? Play how you like If you’re not interested in the campaign, you can pick another mode and play the campaign later – or the other way around, play the campaign first and then the other modes. Or maybe all at once? Campaign – Action-packed storyline with dialogue choices, main and side missions, 14 unique bosses, diverse enemies, and the mystery of the Patterns to uncover. Survival – Endless waves, increasing difficulty, and global leaderboards. You start with no abilities – how long will you survive? Sandbox – Full freedom, test builds, create experimental patterns, and have fun with God Mode. Quick Run – Intense action for a short 10–20 minute session, global leaderboards, perfect if you’re not into the campaign but want a quick game. Tower Defense – A classic tower defense mode with progressive difficulty. Enemies move along set paths, and each round becomes increasingly intense. The final stage turns into an endless survival battle. Perfect for fans of traditional tower defense games. Finally… a bullet hell roguelite with SAVE ANYTIME Unique and complete Save State System (First in the world?) – Save & Quit at any time, resume exactly where you left off: with every bullet, enemy, and ability. Key Features (Cherry Picks) 🎯 Create your own bullet hell – Intuitive slider-based pattern editor with ready presets, random layout generator, and the ability to save patterns as JSON files. 🌌 Choose your game mode – From full story campaign to limitless sandbox. 🚀 Advanced progression system – Tower shop, artifacts, and abilities to purchase. 🪐 Explore the galaxy – Planet scanner, anomaly hunting, and Quantum Data Cores to collect. 🔥 Epic abilities – Nuclear bombs, drones, deflecting shields, time slowdown, and more. 👾 Rich bestiary – At least 14 bosses and 9 different enemy types, each with unique attacks and behaviors. 🎮 Full player support – Mouse, keyboard (shortcuts), gamepad, and touchscreen support. 💻 Cross-platform – Available on Windows, Linux, and Mac . 📀 Retro audio vibe 🏆 Steam integration – Achievements, online and offline leaderboards, cloud saves. ⚙️ Full settings – Graphics profiles, sound/music adjustment, smooth gameplay with seamless transitions and music crossfade, difficulty levels. 🎥 Cutscenes and narrative – Campaign intro and ending scenes. ♾️ High replayability – Every run is different thanks to the pattern system and random upgrades. Visit the Armory Buy and unlock powerful towers, rare artifacts, and unique abilities that can completely change the course of battle. Experiment with different combinations, create deadly synergies, and adapt your strategy to your playstyle – every decision can turn the tide of the fight. Active abilities are your strategic tools: place mines anywhere on the battlefield, summon drones and support ships, set up flanking units or ambushes. Every use is a tactical choice that can decide whether you survive the next wave. Who is this game for? Pattern Survivors is for you if: You enjoy games inspired by Vampire Survivors , Brotato , or Hades , but expect a fresh approach and new mechanics – this is not another clone of the well-known titles . You love bullet hells but always dreamed of designing the attack pattern yourself. You want a game that gives you the choice: quick sessions or a deeper campaign with a story. You value high replayability and the ability to experiment with strategies. If you like supporting independent solo developers without a publisher or PR – this game is made out of passion and for players just like you.\n\nCASE 15 3950050 Touhou Witch's Night Market 2025 auto\nMarisa has lost all her fortune in 'Kappachinko', now she must fight other residents in Gensokyo to repay the debt she owes Reimu! Collect various items and skills, build your construct strategically, then head towards powerful opponents! She may even encounter a mysterious being in the end…\nJoin our Discord \"Touhou Witch's Night Market Community\" for discussing！ 💰️ REPAY THE DEBT Marisa had sold all her magical items to raise money for 'Kappachinko'... Unfortunately she lost everything in it! Driven by addiction, Marisa had stolen all across Gensokyo, even took on a malicious loan from Reimu! With no other choice, she's forced to join the magical item battles that was trending in Gensokyo, hoping to earn enough to repay her debts and escape from Reimu's relentless pursuit… With her sharp intuition for magical items, she believes she can defeat all opponents. But... will things really go as she hopes? 🤓 BUILD YOUR CONSTRUCT You will play as Marisa, collect items and skills to build your own construct, then try to defeat the opponents! The game features a total of 158 items and 69 skills, each with its own unique attributes and effects. Unleash your creativity and combine them into powerful constructs! You might even discover new combinations that the dev team never imagined. Each game lasts only around 15 minutes. Keep challenging and see how far you can go! ⏱️TIME TO FIGHT Each game is divided into 10 days , with each day split into 4 moments. In the daytime, You will encounter various events, where you can purchase items, learn skills, gain exp, fight random enemies, or even change the effect of an item! At the end of each day, you will visit the night market and meet other Touhou charaters who specialize in different weapons (we won't spoil who they are, but let's just say you might run into Koishi, Remilia or Reisen). Choose who you want to talk to, then challenge and defeat them! If you manage to survive until the end of Day 10, you will step into the 11th day where an unexpected challenge awaits. Overcome this mysterious being and then you can— 🔓️ UNLOCK EXTRA MODE Clear the game several times on Lunatic difficulty to unlock EXTRA MODE, where you're able to challenge your historical data! This mode also allows you to import other players' data. Share your save files with your friends and let them try to defeat your construct! But don't celebrate too soon—No matter how powerful your construct seems, there will always be a strategy that can counter it! AND MORE THINGS... Customize the background! Decorate it with your favorite Touhou Character's stickers! Discover the memes and other touhou references hidden in the dialogue! Visit the Meboxen club on the homepage to read messages left by devs and fans! There are still many secrets waiting for you in Touhou Witch's Night Market... Join our Discord \"Touhou Witch's Night Market Community\" for discussing！\n\nCASE 16 3584870 Lucha Chess 2026_H1 auto\nAssemble the ultimate team of Mexican wrestlers in this roguelike auto-battler. Combine unique units with powerful items to discover hundreds of synergies while battling your way through.\nStep into the ring and conquer the Championship! Lucha Chess is a strategy roguelike with the flair of Mexican wrestling and the charm of retro cartoons. Instead of button-mashing action, you’ll dive into chill but deep tactical battles, building the ultimate team of eccentric heroes. Experiment with countless combinations of heroes items, and relics to unleash broken synergies. Will you find the perfect balance between offensive power and defensive stall? Will you smash through the enemy frontline or strike straight at their backline? The choice is yours every match!\n\nCASE 17 3809330 Hill Defender 2026_H1 auto\nYou’ve claimed the best spot on Earth, but invaders want it. Defend your camp from waves of enemies trying to destroy your base or throw you off the cliff. Use ballistas, catapults, and spells to stop them. Earn rewards, upgrade your defenses, and survive stronger waves. Never let your camp fall!\nYou’ve claimed the best spot on Earth, and now everyone’s out to get you. Defend your camp from waves of angry invaders who want to tear down your home and some even throw you off the cliff. Use automatic ballistas, catapults, and powerful spells (and more) to stop them in their tracks. Earn rewards by defeating enemies, then upgrade your defenses with new siege engines and magical abilities to withstand increasingly stronger waves. Remember: your base camp must never fall! Defend the Hill. Build. Upgrade. Survive. Hill Defender is a fast-paced 3D tower defense indie game where you take control of powerful defenses atop a strategic hill and fend off waves of increasingly dangerous enemies. Designed for players who enjoy tactical planning and real-time action, Hill Defender blends traditional tower defense mechanics with engaging 3D visuals and responsive gameplay. Build Your Defense Place and upgrade a wide variety of defensive towers, each with unique behaviors, strengths. From rapid-fire turrets to heavy artillery and area denial weapons, your setup will make or break your success. Real-Time Combat & Enemies Face a range of enemies—fast runners, armored beasts, swarming hordes, and elite boss units—each with unique attack styles and movement patterns. Strategy is key as enemies come from multiple directions and adapt to your defense patterns. 3D Hilltop Battlefield The entire game unfolds on a fully 3D circular battlefield with multiple elevation levels, environmental obstacles, and rotating camera controls. Your hill becomes the heart of an immersive defense experience where placement, range, and timing matter Upgrade System Collect resources during waves to upgrade your camp, or your weapon in real time. After each match, you’ll earn in-game currency based on your performance. This currency can be spent in the upgrade shop to unlock permanent enhancements that persist across sessions. Buy turrets, spells, and even an AI companion to aid you in battle. Enemy Variety & Weight System Hill Defender features a dynamic enemy system where every enemy type has a weight value that determines how it contributes to a wave’s total difficulty. Instead of static enemy lists, each wave is generated using a weighted budget system—ensuring scaling challenge and variety without overwhelming the player unfairly. Each wave can be a little different then last time. Enemy Types Include: Swarmers – Fast, fragile enemies that rush your defenses in large numbers Regular – Regular balanced enemies with no special ability. Medium Units – Heavier enemies, not so easy to kill or push away. Tanks – Slow, heavily armored units that soak up serious damage Ranged Attackers – Enemies that can shoot from a distance, with the ability to push you down off the cliff. Gameplay Highlights Classic tower defense core with a modern 3D twist Wave-based progression with increasing difficulty Dynamic enemy AI with diverse movement and damage types (melee, ranged, area, etc.) Real-time upgrade and resource management Unique boss encounters and special events Clean UI with customizable camera for tactical overview or close-up action Strategic Depth Each level is a challenge in efficiency and adaptation. Choose your tower loadout, respond to enemy compositions, and prioritize upgrades wisely. Every decision impacts your survival. Playtime & Future Updates Hill Defender offers 2–4 hours of core gameplay , depending on your strategy and skill level. Whether you're a quick-thinking tactician or someone who enjoys experimenting with every tower setup, you'll find a compact but satisfying experience. We’re actively working on free post-launch updates , including: New maps with new enemy types and wave variations Additional towers, player upgrades and spells. Player AI companions Achievements Cloud saving Expanded game modes and difficulty settings Quality of life improvements based on player feedback Genre & Target Audience Hill Defender is perfect for fans of tower defense , strategy , and real-time tactics . Whether you enjoy defending against waves for hours or mastering every map with the perfect strategy, there's a challenge here for you.\n\nCASE 18 4765580 Brainrot Autobattler 2026_H1 auto\nBrainrot Autobattler is a chaotic roguelite auto-battler/autochess where you buy brainrot units, evolve them into stronger forms, trigger trait synergies, and survive a run full of shops, bosses, boons, and bad decisions.\nBUILD THE DUMBEST TEAM ALIVE TRIGGER CHAOTIC TRAITS COLLECT BOONS THAT BREAK THE RULES CHOOSE YOUR PATH THROUGH EACH RUN FACE THE WEIRDEST BOSSES WITH THE WORST IDEAS EVOLVE GAMEPLAY Choose your starting boons, assemble your team, and navigate through procedurally generated maps. Each layer presents new challenges: shops to buy units, elite encounters to test your strength, and boss battles that push you to your limit. Combat is automatic — your units fight based on their stats, keywords, and trait synergies. But before the battle begins, you have time to plan your formation and strategy. FEATURES 10 Rot Levels with increasing difficulty and rewards Endless Mode with leaderboard 7 Layers per run with varied map layouts 50+ Boons to unlock and customize your playstyle Multiple Trait Systems with unique synergies Procedural Map Generation — no two runs are the same Shop System — buy units, reroll offers, and spend gold strategically Boss Battles — challenging encounters with unique mechanics\n\nCASE 19 4796920 Dynasty! 2026_H1 auto\nUnlock your players. Build your playbook. Grow your stadium. In this American Football deckbuilder, every win is shaped by your Play Cards, Player Cards, and stadium upgrades.  With over 1.3 billion possible team combinations, win championships your way and build a true Dynasty!\nAmerican Football as a Rouge-lite Deckbuilder \"It's not whether you get knocked down; it's whether you get up.\" — Vince Lombardi Building your dynasty from scratch takes time, money, and skill. Do you have what it takes? The game starts you off with limited players, resources, and a stadium in shambles. Unlock your team, earn money, and build your stadium to take on the best of the best. What's better than football? Nothing! Unlock your players. Build your playbook. Grow your stadium. In this American Football deckbuilder, every win is shaped by your Play Cards, Player Cards, and stadium upgrades. With over 1.3 billion possible team combinations, win championships your way and build a true Dynasty! Get in on the action! Get up close and personal with the gridiron in Field View. March down the field for that last minute score! Customize your player's frame, name, and image! You have the flexibility to change one or all of your player's frames. Edit names and replace images for full customization with your players. Owner's have influence on the organization! What backstory will you choose? Early game perks, balanced perks, or mid-late game perks? Extreme visibility on every play! See player, stadium, weather, and owner modifications on each play. Dynasties are made from Championships! How many trophies can you collect in your Dynasty run? Dynamic Pregame, Halftime, and Postgame Real time play by play professional broadcast booth and presentation packages with insights and color commentary. Full 32 Team Roster Editor Build your Dynasty with pre-made teams or build your own 32 team custom leagues! Create and share your roster files with the community. Rock, Paper, Scissors - Card Battle You must call the right play at the right time! Rock, Paper, Scissors style gameplay can make or break each down. Your Dynasty, Your way Endless possibilities! With 46 Player Skills, 11 Bonus Skills, 20 Play Card Augments, and 36 Stadium Upgrades, you can build a Dynasty to fit your team and play style. Weather Systems & Stadiums Impact Just like on Sunday, weather plays a huge factor in success of plays, fumbles, and interceptions. Unlock and manage your All Star Players Field your best roster! With 2 Gold, 2 Silver, and 4 Bronze slots, you can field the best team for any scheme and Play cards. Weekly Dynasty 365 Podcast The team stays with you all season long with insights, analysis, and recaps. Go for Gold! Gold Tier Players have the best skills and can turn the tide of battle.\n\nCASE 20 3816000 Flowers and Deities 2026_JulAug auto\nFlowers and Deities is a strategy deckbuilder roguelite where you command a blood-soaked lizardkin army through escalating arena battles, spending every drop of spilled blood to cast spells, summon creatures, and survive long enough to claim the favor of the Gods\nFlowers and Deities is a strategy deckbuilder roguelite set in a rainforest world ruled by bloodthirsty gods. Lead your army to battle and spend spilled blood as a resource to cast spells and summon creatures. Die, adapt, and descend into the arena again. Choose from 4 origins and faiths that define your playstyle from the first fight. Battle across multiple grand arenas against 15+ races and 125+ unique unit types. Face noble legions, barbarian tribes, underworld horrors, and alien visitors. Defeat them, and conscript them into your own ranks. The build space is expansive: 12 tactical archetypes, 300+ magical items, 100+ divine blessings, and 50+ status effects. A flexible equipment and slot system lets you experiment freely, counter enemy weaknesses mid-run, and find combinations the gods never intended. Blood is your most precious resource. Every unit lost - no matter yours or theirs - fills the pool. Spend it to turn the tide. Combat arena getting dry? Sacrifice your kin and claw out that victory with the power Gods bestow upon you. Key Features 4 origins and faiths - distinct starting builds that shape your entire run 125+ unique units across 15+ races - recruit enemies you defeat and grow your legion Vast build space - 12 archetypes, 300+ items, 100+ blessings, 50+ status effects Blood is resource - every death fuels your next move 6 grand arenas - escalating enemies including barbarian tribes, underworld horrors, alien visitors, and legendary beasts\n\nCASE 21 4223200 Ars Mechanica 2026_JulAug auto\nA strategic auto-battler RPG played on a 10×5 battlefield! Build your own mechanical army with over 300 skills and extensive customization options!\nWelcome to the world of \"Ars Mechanica.\" Ars Mechanica is a strategic auto-battler RPG played on a 10×5 battlefield. Create your own army of humanoid machines and customize them however you like. ■ Strategy Matters in Auto Battles -Positioning and Builds Decide Victory : Battles are fully automated. Victory depends entirely on how you customize and position your mechanical soldiers. -Effortless Growth : Let your units fight automatically to gain levels and learn new skills. -Endless Synergies : Combine a wide variety of skills, create powerful builds, and overwhelm your enemies and dominate the battlefield. ■ Defeat the Imperial Officers -Boss Battles Every 100 Stages : As you progress, powerful Imperial Officers with unique skills and tactics will stand in your way. Adjusting your formation to counter them is the key to victory. -Claim Their Power : Defeat Imperial Officers to unlock their skills for your own army. Expand your tactical options and prepare for even greater battles ahead. ■ Features -Over 30 Skill Trees -Over 300 Active and Passive Skills in Total -Modules with 6 Rarity Tiers -Freely customizable characters with 19 customizable body parts and color options\n\nCASE 22 4897540 Miner Survivor 2026_JulAug auto\nMine. Build. Survive. Now fight your friends.\r\nA fast 2D survivor game where you dig gold, build towers and drones, fight monster hordes, and now duel other players in 1v1 online PvP. Free demo available.\nMine. Build. Survive. Now fight your friends. Miner Survivor is a fast 2D top-down survival roguelite where you dig for gold, build towers and drones, and fight off endless waves of monsters. Every run gets harder. The deeper you go, the more chaotic the battlefield becomes — and one wrong step sends you back to level one. ⚔️ NEW: Online 1v1 PvP Drop into real-time matches against other miners. Spend gold mid-battle to summon scary enemies on your opponent's side. First to lose their health pool loses. Disconnect mid-match? The game remembers — jump right back in. Build Your Drone Squad Six specialized drones, each with their own attack style: • Assault — reliable single-target damage dealer • Gatling — sprays up to 3 homing bullets per shot • Splash — AOE explosion on every hit • Sniper — long-range piercing shots • Pulse — chain lightning across nearby enemies • Tempo — radial pulse waves that slow the swarm Level them up to L5 and unlock devastating synergy combos. Mix and match your roster — loadout changes between runs. Sacrifice Spirits for Power Every drone can carry up to 4 sacrifice elements : Stun , Knockback , Execute , Silence . Feed captured spirits to upgrade each element's rarity — Normal → Rare → Legendary. Build your perfect setup, one spirit at a time. Collect 32 Hidden Mine Spirits Scattered across every cave, 32 unique spirits roam the map — coal, copper, gold, crystal, diamond, and more. Some are common. Some you'll only ever see once every 50 runs. Catch them all for permanent Steam achievements and unlock the in-game Spirit Codex. More to love: • Boss skill orbs — random drops grant powerful one-shot abilities • Cryo Cloud — a rare spooky cloud that freezes you in place. Don't get caught. • Endless Mode — the waves never stop. How long can you survive? • Free Demo — try before you buy on Steam Ready to Mine? Add Miner Survivor to your wishlist and jump in today. See you in the caves.\n\nCASE 23 4934320 SlimeHold 2026_JulAug auto\nChoose your King, recruit an army of Slimes, collect run-changing Items, and assign Slimes as Stationed or Roaming in this roguelite tower defense auto-battler.\nBuild Your Army. Hold the Line. SlimeHold is a roguelite tower defense auto-battler where you recruit Slimes, collect Items, and defend your King against increasingly dangerous waves. Your Slimes fight automatically. You decide which Slimes join the army, which ones hold their ground, which ones roam the battlefield, and how each run develops. Stationed or Roaming Set Slimes to Stationed when you need a dependable defensive line, or set them to Roaming so they move freely across the battlefield. Recruit more of a reliable Slime, field a varied army, or build around a rare ability that changes how your defense performs. There is no single correct formation. Your choices determine how the army responds as new enemy types, Elites, and Bosses begin to overwhelm weaker defenses. Shape Every Run with Slimes and Items Recruit Slimes with distinct attacks and abilities, then collect Items that strengthen your army in different ways. Build around Burn damage, piercing attacks, bouncing projectiles, stacking damage, or a balanced force prepared for several threats. Reinforce a dependable strategy or take a chance on a different setup. As the waves become more dangerous, every recruit and Item matters. Choose Your King Each King gives your defense a different starting lineup and foundation. Try a new King, adjust your approach, and discover which Slimes and Items best support that playstyle. Unlock More Ways to Play Push deeper, defeat Bosses, and complete challenges to unlock more Slimes, Items, and Kings. Your collection expands between runs, giving you more options while each attempt still depends on the decisions you make. Features Automatic Combat, Strategic Choices – Your Slimes handle the fighting while you build and direct the army. Stationed and Roaming Roles – Create a stable defensive line, a mobile force, or a mixture of both. Distinct Slimes – Recruit Slimes with different attacks, abilities, rarities, and battlefield roles. Run-Changing Items – Collect Items that strengthen different strategies during each run. Multiple Kings – Begin each defense with a different lineup and strategic foundation. Permanent Unlocks – Expand your collection with additional Slimes, Items, and Kings. Endless Waves – Test every army against escalating enemies, Elites, and Bosses. Choose your King. Build your Slime army. See how long the kingdom can hold.\n"}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"5d52c9","wall_time_seconds":0.000003509,"exit_code":0,"original_token_count":7310,"output":"\nCASE 24 2862290 Gunswitch 2025 action\nGUNSWITCH is a top-down arena shooter: Mow through hordes of demonic aliens. Upgrade your weapons to fire thousands of bullets within seconds. Escape space prison with your robot squad. Switch your loadout on the spot to adapt and become a killing machine.\nGunswitch is a game that yearns for more! Thousands of monsters, ten thousands of bullets on-screen! Action packed with blood and explosions, mixed with the weight of realistic movement to capture the thrill of a fluidly approaching horde, skittering and slumbering in to get you. Steel yourselves for carnage! Steel won’t bleed, but cut through flesh. Fight through massive armies, not only strong in numbers but in variety. You’ll come across enemies of all types and kinds, tough and muscular, fast and nimble, multitudes of bugs, giant lava monsters, or smarter humanoid species. Monsters might decide to stick around with others like packs of wolves, make formations to increase their chances and as waves pass like tides of oceanic hell, they will get stronger. Do anything to avoid your destruction. Run and herd them or push and knock them around with sheer force of firepower and explosions. This will be a case of natural selection. Hopefully it will be on your side. Grab guns and upgrade them into sci-fi weapons of total mayhem. With modifications even a regular rifle can fire more projectiles, explosive bullets and have increased fire rate or higher stats all around. When one isn’t enough, get a new one! Operate pawns, a squad of androids, equip each of them with these weapons, and there’s no limit to synergies and pure damage you might come across. Don’t worry, they will cover you automatically and you can switch between characters at will. They act like extra lives! A future society of humans-turned-androids are locked in a facility called Cerberus. A space station designed to keep and research galactic level threats. Explore its each level filled with horrible dangers and help them get out. A fluid, dangerous horde of monsters . Guns with wild modifications that can shoot thousands of bullets! Control a squad of androids that can cover you automatically. Or just use them as spare bodies. Switch guns and characters on the spot to adapt, experiment and create new builds . Escape Cerberus, a space prison with many levels and facilities . Easy Local Co-Op , just plug in a controller and play as a squad member, drop in and out . We are excited to announce our debut game, Gunswitch. We can’t wait to hear the player feedback, and see what kind of new ideas will come out of it. Please add Gunswitch to your Wishlist and Follow us to stay in tune with the development. Thank you!\n\nCASE 25 3217930 BeefArena 2025 action\nDefend the earth from an invasion of cosmic cattle alone or with a friend in a roguelike romp with old school attitude. Find weapons, meet new friends, and make a grocery run through an extra terrestrial invasion that's never the same!\nBeefarena Aliens have invaded the planet, and they seem to have taken special interest in earth's livestock! Your mission, should you choose to accept it, is to help MeggaCorps rid the planet of this mutant menace by venturing out into the world and defending its people from the deadly herd. While on tour, you'll meet new people, find weapons, and face the most fearsome boss cows on the ranch. Welcome to Beefarena, a wild roguelike ride through an alien cattle catastrophe! Progress through a series of retro-inspired action packed platforming combat challenges to find refugees and listen to their stories Stages randomly generated from a collection of hundreds of hand crafted rooms across 9 distinct biomes, with new rooms that unlock as you complete each one Multiple paths through the game via warps and other secrets. 10+ Bosses at launch Fill your locker with the 20+ unlockable weapons with dozens of items to tailor them to your playstyle Leaderboards and Career Stat Tracking Local 2 player co-op More bosses, rooms, enemies, and items coming in future, free, content updates! Volunteer for Megga Corps Enlist in Megga Corps today for your chance at glory! Confront the herd across 9 randomized biomes-- No two tours of duty are ever the same. Each biome has its own unique theme and challenges. Once you've mastered an area, more challenges will unlock! Do you have what it takes to wrangle this invasion force and liberate earth? Build Your Arsenal Fill your weapon locker with the weapons you find in combat and enhance your firepower with items available in Megga Corp vending machines, conveniently located near the hottest invasion sites! Use weapons like the flamethrower, chainsaw, or the electrify the bovine beasts with the cattle prod! Collect items that will enhance your weapons and imbue other special powers and passive abilities-- like Firewalking. Effective inventory curation will be essential to success out in the field! Each time you find a weapon or item in the field, it will be added to your stockpile back at Beef Post to enhance your firepower for future runs! Round up Your Posse Bring a friend and battle together with couch co-op! At any time a second player can pick up a controller and jump into the action with the press of a button. Recruit your most skilled allies to bring liberation to earths homesteads, or get someone to bait all those heat seeking rockets!\n\nCASE 26 3781740 Vanquish Evil 2025 action\nVanquish endless waves of challenging enemies in this fast-paced action! Combine upgrades to be even more powerful! Buy and equip Gear and Items for your customized builds and stats! Play as 6 different heroes in this Bullet Heaven vs Bullet Hell! Each hero has mechanically unique abilities!\r\n\n\"Vanquish Evil throws you into dense forests and flaming rings while unleashing wave after wave of monsters. Feels like a survival gauntlet on overdrive.\" - IndieSight \"Fast-paced, addictive, and full of variety! The upgrade system keeps each run fresh, and the pixel art style looks fantastic. A must-play for action RPG fans!\" - Awesome Games Review Group FAST-PACED ACTION ROGUELITE COMBAT Vanquish Evil aims to take the most fun core gameplay from the ARPG/Roguelite/Survivors-like genres -- and crank up the action to one thousand percent! Where player skill and strategic upgrade choices strongly influence the runs. In addition to auto-abilities, the game is also a Twin-Stick Shooter where certain abilities have manual direction and activation. The game has a signature design, influenced by weapon evolutions, called Ascended Synergies. These abilities are created from combining 2 chosen upgrades after both have been chosen. These will usually have level requirements before they appear as upgrade choices. THE STORY A supernatural cosmic battle is being fought for the mortal realm, and much of the world is already in ruin. A final resistance of humanity are brave enough to fight. Many die, and are turned into undead armies who serve the formless evil cosmic presence known as Absolute Darkness (Vile). Ascended Light (Veil), a cosmic presence in favor of saving humanity, send Cosmic Guardians to fight alongside the mortal armies. The final elite army of heroes stand together to fight an apocalyptic war with one common goal... Vanquish Evil. GAME DEVELOPMEMENT GOALS Creating fun gameplay and high replay value is the goal. The game in it's current state is built to have all character progression in-game for that specific run, which gives the game a lot of replay value. The game now has meaningful meta progression. Trade Soul Crystals from vanquished enemies for Gold Coins, which are then used to buy Gear and Items. Gear enables spells and other abilities that the player can equip to then use within a run. You can choose what Gear you want to equip or un-equip in the SHOP menu. You can also buy Items, which are permanent meta progression in the form of buffs and passives. SPECIAL FEATURES Stylized hand-drawn pixel art, animation, and background art! Original Music Soundtrack, paying homage to classic retro RPG soundtracks!\n\nCASE 27 4232170 Emberion 2025 action\nSurvive endless waves of enemies in this action-packed roguelike. Master powerful spells, unlock heroes, and compete on Steam leaderboards. How long can you last?\n== ABOUT THIS GAME == Emberion is an intense action survival roguelike where you fight endless waves of enemies in the cursed Ashen Ruins. Master powerful spells, unlock legendary heroes, and see how long you can survive. Every run is different, every death makes you stronger. == KEY FEATURES == • AUTO-ATTACK COMBAT - Focus on positioning and strategy, not button mashing • MULTIPLE HEROES - Unlock and master the Battle Mage, Sorcerer, and Warlock • POWERFUL SPELLS - Lightning bolts, spectral skulls, flame slashes, and more • PROGRESSIVE DIFFICULTY - Each wave gets harder, bosses appear every 10 minutes • PERMANENT UPGRADES - Spend gold earned in runs to permanently strengthen your hero • STEAM LEADERBOARDS - Compete globally for the longest survival time • ENDLESS REPLAYABILITY - No two runs are the same == GAMEPLAY == Start each run weak, but grow stronger as you survive. Collect experience to level up and choose powerful upgrades. Earn gold to spend on permanent stat boosts back in the Sanctuary. Face increasingly difficult enemies and epic boss battles. How long can you survive? == HEROES == • BATTLE MAGE - Balanced warrior-mage hybrid • SORCERER - Master of arcane magic • WARLOCK - Dark magic specialist Each hero has unique abilities and playstyles. Unlock them all and find your favorite! == FUTURE UPDATES == Emberion is actively being developed with regular content updates planned: • NEW HEROES - More heroes coming in future patches • NEW MAPS - Additional areas to explore and survive in • ACHIEVEMENT SYSTEM - Unlock achievements as you progress • MORE CONTENT - New spells, enemies, bosses, and features == Social == Join our community to stay updated on the latest developments! == SYSTEM REQUIREMENTS == Requires Windows 10 or later. See store page for full system requirements.\n\nCASE 28 2955630 Catjinni Survivors 2026_H1 action\nFight with magical cats against demon hordes in this brutally charming, hand-drawn Survivors-like. Playable on your monitor, Steam Deck or fully in VR!Harness divine Catjinn powers, grow stronger by day, and survive the endless nights.\n\"If you like Vampire Survivors, then you are almost certain to like this one. It's an easy 8/10 from me.\" — Penguin, Reddit Community Admin \"The premise, the achievements, the power-ups, the music... everything is PURRRRRRFECT!\" — GamerBwoii, Steam Review \"Extremely 'broken' (OP) builds and a mystical aesthetic that stands out from the crowd... the satisfaction of clearing the screen is guaranteed!\" — Super Crítico, Steam Review \"Support this solo dev! The art direction is very clean, and the mix of 2D in a 3D world is so cool... The game is worth it!\" — wiwi, Steam Review Unleash the Power of Catjinni Make your wish. Slay the horde. Survive in style. Step into this solo-developed indie game that blends Roguelite action with RPG depth. In Catjinni Survivors, you’ll battle thousands of monsters, face off against mighty bosses, and unleash devastating magic to stop the dark Jinn Ifrit and his hordes of ghouls. Key Features Survive Day & Night: Face thousands of enemies and choose your battles wisely. At night, the ghouls become more dangerous. The game's start time syncs with your real-world time, affecting how your run plays out! 4 Unique Cat Classes: Francis: A blind, battle-hardened warrior monk. Fluppy: A brilliant professor of arcane magic. Sushi: A brave samurai and descendant of a legendary Ice Catalan. Flocki: A famous detective with a fiery temper. Explore 3 Legendary Biomes: Discover the mystical sands of Ancient Egypt, the sunken depths of Catlantis, and the forgotten Dungeons of the Catjinns. Tactical Traps & Progression: Use the environment to your advantage. Strengthen your team with visible gear, unlock skills, and master magical weapons on your way to claiming the ultimate crowns of the Catjinn Kings. The Numbers 15-20 Hours of core content. 60 Unique Weapons and abilities to master. 50 Challenging Enemy Types and over 18 Powerful Bosses . 5 Days & Nights of brutal survival challenges. Full VR & Steam Deck Support – play however you want! Roll the dice! Defeat Franky in the dice match and earn gold for upgrades. But beware: if you lose, it’s all gone! Become a Catjinn Master. Catjinni Survivors combines intense combat, hand-drawn worlds, and endless replayability. Whether you're into action, tactics, or adventure – in the world of the Catjinns, every wish is welcome. (OPTIONAL) VR-compatible Immerse yourself in the game like never before in this genre. Experience 3D sound and thrilling action in VR as you fight off hordes of enemies with your cosmic powers!\n\nCASE 29 3902210 Laby's Village 2026_H1 action\nWeary Wastelands, Malicious Forests, and Desolate Deserts...\r\nA 2D pixel action roguelike.\r\n58 unique weapons and 277 Demon Relics let you create your own build. Cut through hordes and challenge unique bosses.\r\nPlay as the goblin officer Laby on an absurd yet exhilarating conquest.\nLaby’s Village Laby’s Village is a fantasy-comedy 2D pixel action roguelike featuring unique dungeon-driven mechanics. You play as Laby, an officer of the Demon King’s army. In the chaotic era following the Demon King’s fall, you set out from a small village that safeguards the “Heart of the Demon King,” determined to conquer the ambitious warlords who refuse to submit. The game combines fast-paced combat, weapon progression, and relic-based builds. Purchase and upgrade a wide variety of weapons and Demon King relics to craft your own combat style. Each weapon has its own attack rhythm, projectile effects, and humorous personality, while relic enhancements can trigger unexpected and wildly satisfying chain reactions. ⚔️ Features 🗺 Unique Dungeon Mechanics Every dungeon features its own unique rules. Different environments, mechanics, and bosses create completely different combat pacing and challenges. 🎯 Weapon × Relic Build System Every run allows you to mix and match weapons and relics to create highly personalized builds. From a self-returning Bone Boomerang to the meteor-summoning Starfall Bringer, every weapon comes with its own quirks and temperament. 🧙 Quirky — and Dangerous — Villagers The Dark Elf Knight Commander “Blackheart,” the Great Rat Sage “Squeakmaster,” and the Dogfolk Traveling Merchant “Anna”… The villagers may seem ordinary at first glance—but they are anything but simple. Perhaps they will help Laby grow even stronger. 🔥 Ruthless Bosses Every dungeon has its own boss— Goblin Warlords, Elf Queens, Kobold Consuls... None of them are nice. They are all ambitious. And every one of them looks down on Laby. 🌈 Fantasy × Comedy This is Laby’s Village—a little run-down, yet full of absurd and lovable characters. A world filled with humor, weirdness, and adventure.\n\nCASE 30 4522710 Swarmbound 2026_H1 action\nUnleash pixel-art hell in a high-performance bullet heaven. Command 5 classes against massive 1,500-unit swarms with zero lag. Loot, level, and spec your way to Wave 30 with chain lightning and spiral hammers. Fully moddable, Steam Deck optimized, and built for pure retro destruction.\nWhat if an 8-bit console had unlimited power and an open-door policy? Swarmbound is a high-performance \"Bullet Heaven\" roguelite built on a simple premise: Retro aesthetics, modern scale, and total moddability . Face off against impossible swarms of up to 1,500 active enemies at a buttery-smooth 60fps, then change the entire game without ever hitting \"Quit.\" No bloated engines. No long load times. Just pure, unadulterated arcade chaos. Features The \"Impossible\" Swarm: Battle massive waves of enemies that would have melted original hardware. Our custom MonoGame engine handles 1,000+ mobs and hundreds of projectiles without breaking a sweat. Five Classic Classes: Master the frontline Paladin , the crowd-clearing Sorceress , the entity-commanding Necromancer , and more. Deep Tactical Survival: It’s not just about moving in circles. Balance class-specific abilities, manage area-denial spells, and hunt for shrines to turn the tide. Lo-Fi, High Fidelity: Immerse yourself in a world of chiptune SFX and MIDI compositions. Toggle the integrated CRT Post-Process Shader for that authentic scanline glow. Optimized for Handhelds: Built specifically for the Steam Deck , offering \"pick up and play\" sessions with near-instant loading and zero-compromise performance. The \"Open Engine\" Experience Swarmbound was built to be broken. We believe the community makes the best content, so we built the game as a modular platform. Full Mod Support: Every sprite, sound effect, item, and spell is mod-loadable. Hot-Swappable Content: Swap mods in and out live without restarting the game. The Druid Example: To prove the power of our modding tools, the entire Druid class —including its 8x8 sprites, unique spells, and SFX—was built and loaded using the same external mod system available to you. The \"Unlimited NES\" Philosophy Swarmbound is a love letter to the 8-bit era, stripped of its technical constraints. We’ve traded high-res textures for high-density simulation. Every 8x8 sprite and 1x1 pixel projectile is part of a massive, synchronized world designed to test your reflexes and your build. How long can you survive the swarm?\n\nCASE 31 4730540 TAILROUGE 2026_H1 action\nTAILROUGE is a fast-paced minimalist action roguelite where your tail is your weapon. Move with the mouse, boost through danger, circle enemies, slice through swarms, and collect powerful cards to evolve your build in short, intense arcade-style runs.\nTAILROUGE is a fast-paced minimalist action roguelite built around one simple idea: Your tail is your weapon. Move with the mouse, boost through danger, and use your trailing body to cut through waves of enemies. Every run is short, intense, and focused on clean movement, sharp turns, and satisfying chain reactions. Collect cards to evolve your abilities and create powerful builds. Some cards enhance your tail, some add new attacks, and others give you defensive tools when things get dangerous. Circle enemies. Slice them apart. Survive as long as you can. Features: - Fast arcade-style survival action - Tail-based combat built around movement and positioning - Card-based upgrades that change each run - Minimalist red-and-black visual style - Short, replayable runs - Simple mouse controls - No complicated menus, no long tutorials, just action TAILROUGE is designed for players who enjoy quick roguelite runs, clean mechanics, and the feeling of getting stronger one card at a time.\n\nCASE 32 4711500 Qarma Quad 2026_JulAug action\nControl four misfits working as a single Quad. Every interaction and choice is influenced by Qarma, the invisible force that pulls the strings. Complete difficult quests, earn Qoin, and stake your claim on the Qroma Galaxy!\nWelcome to the Qroma Galaxy, where Qarma is as real and measurable as gravity itself. A gig economy of freelance delivery crews carries the Qarmic burden across its watchful universe. The best crews balance their economy of Qarma and Qoin to live out the longest career. THE QARMA. Qarma is the invisible force that guides the actions of those who reside in the Qroma Galaxy. The universe responds to every action committed, both positive and negative. THE QUAD. Each quad unit is a randomly generated individual with a unique first and last name, qarmic standing, galactic race, and will die permanently. You can recruit and trade them for a price, or force strays to join your crew at gunpoint. Your quad is strongest when your team works together, but every Quest you take is trying to tear it apart. THE COMBAT. Your crew can be equipped with a diverse arsenal of equipment, tailored to suit your playstyle. Your crew’s loadout determines how you approach combat, and getting caught out of position can be deadly. Guns, Melee, and Shields are a rock-paper-scissors of offense vs defense. Kill with guns, shove with shields, and knock out with melee. THE QUESTS. Three factions reside in HQ, to which you can offer your services. The Crime faction promotes negative Qarma and offers more Qoin. The Cult faction is an advocate for positive Qarma, at the sacrifice of good Qoin. The Corpo faction deals with everything in between, and unpredictably. EARLY ACCESS The game is launching in Early Access, with the core gameplay loop in place. We’re looking forward to seeing how the game can expand and grow!\n\nCASE 33 4785950 Pirate Twist 2026_JulAug action\nA top-down naval roguelite where your ship is a mobile base. Build it module by module, sail through dangerous waters, fight pirates, and return to port richer — or don't come home at all. Every run is a gamble.\nBuild your ship. Brave the seas Come home richer — or don't come home at all. Pirate Twist is a top-down 2D naval roguelite where your ship is a mobile base. Cannons on the sides, hull plates in the center, repair station at the stern, storage up front — a 5×8 grid where every loadout creates a different playstyle. Sail out, fight pirates, collect resources, return to port to upgrade. Die, learn, run again. Three Layers of Progression Captain — earn XP from combat to spend on 9 skills across 3 branches (Combat / Navigator / Commander). A captain's XP and skill points are kept between runs — every voyage makes them permanently stronger, win or lose Ship — 17 modules, upgrades that survive defeat, a 5×8 grid you fill out over a run Crew — 6 crew archetypes with passive bonuses; your captain sets the roster (a signature crew plus auto-filled bonus slots), and supplies tick down between encounters Your Ship Is Yours Outfit your ship from 17 modules across a 5×8 grid. Light cannons fire fast and cheap. Heavy cannons punch holes. Mortars keep up a relentless rain of slow shells. Reinforced hull plates trade weight for survival. Repair stations let you patch damage mid-fight. A crow's nest hardens the ship the moment it's rigged. Rare ram prows and hull bracings — prized from sieges and bosses — turn your bow into a weapon and your hull into a wall. You fill the grid out as your run progresses — by the boss fight, your sloop has grown into something the pirates fear. Every Expedition Is a Choice The expedition map branches every step of the way. Calm seas that let you catch your breath. Dangerous pirate waters with elite ships and richer loot. Storms that batter your hull but reward the brave. A besieged port that needs its walls held through five waves. Unknown nodes that roll a surprise outcome on the way in. You pick your path through every expedition. Combat Is Naval Enemies don't stand still. Sloops circle-strafe and pepper your flanks. Brigantines and Galleons tank hits and barrage broadsides. Fire ships charge and explode. Every sea ends in a duel with a named boss captain — drawn from a pool of six — that asks the player to read patterns, not just out-DPS. No cursor, no twitch-aim — combat is pure positioning. Pick your live broadside (port or starboard), swing it onto the enemy, and your guns roar out that side the moment a target's in the arc — auto-fire by default, or switch to manual and time the perfect volley. Easy to learn, hard to master. Roguelite Loop A run is a three-sea campaign, played in a single sitting. Lose a fight and you rebuild at Port and sail back out — defeat costs resources, not the whole run. Captain XP / skill points and unlocked captain archetypes persist across runs (1-of-N pick at the start of every fresh run from your unlocked pool). Five captain archetypes — two unlocked to start, the rest earned by winning runs — each with a distinctive buff and signature crew. Replay is the loop — different captain, different route, different ship. The more you play, the harder it gets. At a Glance Modules — 17 across 6 types Enemy types — 5 + elite variants (Sloop, Brigantine, Galleon, Fire Ship, Boss) Encounters — pirate fleets, mystery waters (calm seas, storms, lucky treasure, elite duels), port sieges, boss finales Captain archetypes — 5 (Rourke, Reyes, Voss, Sable, Marisol) Captain skills — 9 across 3 branches Crew types — 6 (Gunner, Carpenter, Navigator, Cook, Quartermaster, Lookout) Resources — Gold, Wood, Iron, Supplies Session length — a single sitting (a full run is three seas)\n\nCASE 34 4922660 Frontier Survivors 2026_JulAug action\nOutlaw-survivor roguelike set on a doomed Western frontier. Pick a Gunslinger, build your posse, and bleed through 20 waves of bandits, bug-swarms, and worse. 15 characters. 12 biomes. 100+ items. Hand-painted pixel art. Built for runs you'll lose, then start again.\nOutlast the frontier. Bury the legend. Frontier Survivors is a Western-themed wave-survival roguelike where every minute is a stand against escalating bandits, bug-swarms, ghosts, gunslingers, and the boss that's been hunting your trail for three days. Pick from 15 Gunslingers, each with a unique active ability and starting kit. Use Showdown to drop bullet-time for 5 seconds. Use Tinkerer's Touch to upgrade your weapon mid-wave. Use Last Rites to revive once per run, but only when you're below 10 HP. The character you pick shapes the entire run. Survive 20 waves across 12 hand-painted frontier biomes: cracked deserts, copper mines, autumn cornfields, frozen pine forests, and the boneyard where most runs end. Each biome has its own enemy mix, its own destructibles, and its own atmosphere. Build your loadout from 100+ items across 6 set bonuses. The deeper you go, the more synergies break open. Recruit your Posse: 48 spirits of the fallen who follow you between runs and grant passive bonuses. The right posse turns a slow run into a steamroll. When you die, your WANTED poster marks the run: time survived, kills landed, biggest crit, gold earned. Save it, share it, or set it on fire and start again. FEATURES • 15 unique Gunslingers, each with active abilities • 12 hand-painted biomes with distinct enemy rosters • 100+ items across 6 synergy sets • 48 collectable Posse spirits with passive bonuses • 20-wave runs, roughly 15 minutes each, plus Endless mode • Daily and Weekly bounties • Full controller support • Persistent meta-progression: Stash currency, character unlocks, heirloom collection Built solo by MisterEoss. Made for people who lose a run, stare at the WANTED poster, and immediately hit Play Again.\n\nCASE 35 4952430 Sangreve 2026_JulAug action\nA gothic top-down survivor-roguelite. Attack, weave and dodge through undead hordes while picking level-up cards to build your character, then fight a boss at the end of each map. Nine characters, dodge-only defense, run-based roguelite progression.\r\n\nKNELL is a top-down survivor-roguelite with a gothic, undead theme. You pick one of nine characters and fight through waves of enemies across five maps, with a boss fight at the end of each stretch. Your weapon attacks on its own, so you spend the run moving, dodging, and choosing level-up cards to build your character. A single run takes about 20 to 30 minutes. Your only defensive move is a dodge roll, which gives a short window where you can't be hit. Features • Nine playable characters. Each has its own auto-attack, its own set of class cards, and its own ultimate ability. • Over 40 level-up cards, including chain lightning, orbiting orbs, flame rings, freezing totems, and blood explosions. You get a reroll and a banish each run to shape your draws. • Evolutions: max out a core ability and combine it with the right keystone card, and it upgrades into a stronger permanent version. You earn these by beating specific bosses. • Boss fights. A shrinking ring of fire, and each boss drops a relic that changes your later runs. • Five biomes: Overgrown Ruin, Rocky Wastes, Storm Realm, Hellscape, and Blood Desert. • Meta-progression: spend the souls you earn on permanent Armory upgrades, unlock new characters, and buy upgrades that make evolutions come sooner. Clearing the final map unlocks New Game+, and optional Dread difficulty levels make runs harder for bigger rewards. • Full controller support, rebindable controls, and an optional auto-aim. Characters • Mourne: melee, auto-swinging with a warping slash ultimate. • Harlow: a bow that fires piercing arrows in a line. • Warp: a mage with auto-bolts and a ring of elemental totems. • Crow: a grenade launcher for clearing packs. • Roan: a swordmaster with wide arcs and spinning rings of blades. • Vesper: an assassin who swaps between a sword and a spear, each with its own ultimate. • Sanguine: a heavy melee fighter who leeches health and rains firerocks. • Bram: a tank who turns into an invincible beast when his health runs low. • Malach: a summoner who fields a small army of demons. The art is from licensed pixel-art asset packs, and the game has a boss-battle soundtrack.\n"}

Script completed
Wall time 0.1 seconds
Output:

Warning: truncated output (original token count: 16568)
Total output lines: 1

{"chunk_id":"6a61f0","wall_time_seconds":0.0000026,"exit_code":0,"original_token_count":28319,"output":"Warning: truncated output (original token count: 28319)\nTotal output lines: 49\n\n\nCASE 36 3348510 Tower Arena Survive 2025 tower\nTower Arena Survive (TAS) is a dynamic mix of roguelike and tower defense, where your castle becomes an impregnable fortress. Use 80+ weapons and 88 upgrades to balance income, damage, and defense against up to 7 foes. Buy monsters to disrupt enemy waves and change the tide of battle!\nTower Arena Survive (TAS) Welcome to Tower Arena Survive – an electrifying fusion of roguelike strategy and tower defense where your wits decide the outcome of every battle. In TAS, you don’t control a lone hero – your mighty castle becomes the epicenter of your tactical prowess, evolving into an impregnable stronghold through a vast arsenal of over 80 weapons and 88 unique upgrades. Face off against up to 7 opponents in high-stakes matches where every decision counts and every second matters. Balance your gold income, damage output, and castle durability to withstand relentless enemy waves that test your strategy to the limit. And here’s a game-changing twist: purchase special monsters from a dedicated in-game store, and once deployed, these relentless creatures will appear in every wave for all your opponents, adding an extra layer of tactical depth and unpredictability. Each match is a fresh challenge, pushing you to devise innovative strategies to emerge victorious and cement your legacy as the ultimate defender. Choose Your Strategy Craft your unique playstyle by combining a diverse selection of weaponry and upgrades—whether you prefer a rapid offensive or a steadfast defense. Leverage your resources to reinforce your castle, tailoring its strengths to counter your opponents’ tactics. With every match demanding a fresh approach, your decisions on weapon and upgrade choices can turn the tide of battle. Command Your Fortress Your castle is more than a structure—it’s the core of your strategic empire. Invest your hard-earned gold to enhance damage, boost income, and fortify defenses. The special feature of sending waves of monsters, purchased from the exclusive store, forces opponents to constantly adapt to new threats, ensuring that every wave is a test of both your ingenuity and resolve. The Best of Tower Arena Survive Endless Replayability: Infinite tactical combinations ensure no two matches are ever the same. Innovative Mechanics: Use over 80 weapons and 88 upgrades while deploying special monster waves to keep your adversaries on their toes. Thrilling Multiplayer: Engage in epic battles against up to 7 players and prove your strategic supremacy. Fair Play: Success is determined solely by your skill and tactics—no pay-to-win advantages. Step into the world of Tower Arena Survive and let your strategic brilliance shine as you build an unbreakable fortress and outsmart every challenger on the battlefield!\n\nCASE 37 3610530 D.P.S: Weapons Testing Facility 2 2025 tower\nFind the best combinations of projectiles and components to create the ultimate weapon. Mutate and upgrade your components to create stronger weapons. Fight wave after wave of monsters to reach the top. \nBuild the weapon of your dreams Build a weapon from components to fight the incoming enemy waves. Use projectiles and modifiers to make a near infinite number of combinations. Upgrade your components and get new ones to improve your weapon. Do you want to build a gun that shoots cabbages that grow really big so they hit all the enemies? Or perhaps you want to program a gun where sawblades pierce all the enemies and huge amount of chain lightning against them? Everything is possible! Weapon is built left to right There are 2 different types of components, projectiles and modifiers each serving their unique purpose. Modifiers change and add properties for the projectile and projectiles are the ones that are actually fired from the gun. All the modifers to the left of a projectile will affect that projectile. Mix and match four disciplines Mix and match four different disciplines: Bounce, Teleport, Lightning and Spacetime. Create a lot of bouncing balls that grow with every bounce or make the projectiles teleport all over and shoot lasers from every portal. Shoot chain lightning at all enemies or control spacetime by slowing down enemies and hitting them with gravitational waves. Choose your battles Choose what you want to fight against, do you want to fight a huge swarm of enemies? Or perhaps you want to fight only hoppers or crabs. You can also choose based on the rewards; some waves will be more challenging, but you can get more rewards for beating those waves! Mutations improve your components Mutate components to make them even stronger! You can have up to 5 mutations on a single component. Many components have unique mutations, that help you create a stronger weapon. You can also merge three of the same components to improve its rank! More cannons more fun Program one large weapon and two smaller weapons to reach the furthest wave possible! Now with a total of 23 total slots for your components!\n\nCASE 38 3703380 Fantasy Monster Hunt 2025 tower\nUltra-massive hordes of monsters are invading! Select heroes, choose upgrades, collect relics, and strategically deploy skills to prevent the monsters from reaching the defensive line. Each match lasts 15 minutes.\nYou need to clear out massive waves of super ultra-mega-hordes of monsters. Each battle lasts about 15 minutes – perfect for quick sessions anytime! Warriors will automatically attack enemies. Each kill grants XP. When they level up, you'll get enhancement cards to strengthen your warriors! Warriors have synergies – for example, if you unlock the \"Energy Sense\" ability, magic waves will trigger a ring of smaller magic waves upon contacting electromagnetic fields. Relics drop during combat, and powerful ones will significantly boost your firepower. Monsters have specific weaknesses or traits. Exploiting these vulnerabilities allows more efficient elimination. Come and experience the battlefield in person!\n\nCASE 39 3916180 Transport Legion 2025 tower\nA unique roguelike card-based tower defense game. Strategically place cards, manage resource transport, balance production with combat, and craft distinctive decks to claim victory.\nLogistics Win Wars Where other tower defenses focus on troop synergies, Strategy Command makes supply lines your lifeline. Food, materials, tools, transport... Produce, ship, and consume resources across your buildings. Master this economy to train waves of castle guardians， who'll send those monsters packing! Hah! Procedural Maps, Dynamic Adaptation No fixed strategies here. As fog-of-war lifts, adapt to unpredictable terrain. Study roads and topography to pinpoint ambush zones. Complete quests, explore ruins, and evade disasters—special intel will elevate your tactics. Build Decks, Choose Battlegrounds One castle secured? March to the next! The Goddess grants legendary artifacts while merchants offer powerful cards. Combine them wisely—seek balance or embrace extremes. Outmaneuver diverse foes by exploiting weaknesses and amplifying strengths. More Clans,More Difficulties Pick your clan: Each with unique mechanics and cards, demanding fresh tactical approaches. Do you like challenge?From mental warm-up to tactical gymnastics—all playstyles welcome! Immerse yourself in this gradually unfolding strategic journey.Step to step, maybe you will become the master. <Transport Legion> Is: A turn-based game for deliberate planners (pause anytime, no rush). About adapting, not memorizing solutions A bit challenging, similar to a German-style board game. 1-2 hour sessions with no meta-progression grind. A game with conservative numerical values, small yet precise Not Included: No real-time reflexes or twitch gameplay. No rigid best build—every run demands innovation. No A game that can be cleared without thinking. No repetitive grinding or meta-progression. No endless number-chasing; just polished, meaningful choices. This game was developed by a solo developer. To make the game accessible to more players worldwide, the localization has been assisted by AI.\n\nCASE 40 4279880 Misty Land 2026_H1 tower\nDevelop and expand your town as you fend off eerie creatures from the mist. Choose from a variety of buildings and strategies, make tactical decisions, join defensive battles personally, protect your town and fulfill your ultimate goals.\nMisty Land is a strategy game that fuses tower defense, base building, action and roguelike elements. Wield your trusty weapons to venture deep into the wasteland valley, make choices based on distinct development strategies, build your exclusive town and defend it with all your might. Game Features: Selection & Purchase: Choose freely from a diverse range of unique buildings and strategies to craft your one-of-a-kind defense and economic systems. Construction : Place buildings at will, expand your town step by step, and forge an impenetrable defensive line with careful planning. Combat : Step onto the battlefield and face hordes of enemies head-on. Bullets, flames and blades are all deadly weapons against foes, and a variety of active skills deliver utterly distinct thrilling combat experiences. Upgrades : Every building and weapon comes with a wealth of upgrade options, offering you diverse strategic choices. Game Content : A variety of maps, difficulty levels and strategies paired with abundant weapons – plus dozens of distinctive buildings – create diverse gameplay and an immersive gaming experience.\n\nCASE 41 4523160 Slot Theory 2026_H1 tower\nA roguelike tower defense where every draft permanently mutates your build. Stack modifiers, trigger massive Surge events, and watch your defenses evolve into unstable, compounding systems that fight in ways you didn’t design.\nA roguelike tower defense where your build slowly becomes something you didn’t explicitly design. Each wave gives you one irreversible decision. That decision doesn’t just make your defense stronger, it changes how it behaves over time. You don’t build a loadout. You build something that emerges from your choices. Towers and modifiers interact, interfere, and reshape each other across the run. By the end of a run, your build has a distinct identity. Some builds become efficient. Some become chaotic. Some collapse into overwhelming synergies you didn’t plan for. Every run is a different machine formed from the same rules. Seven Operators give each run a distinct starting identity before your first pick. 20 waves. One evolving decision chain. Build it. Watch it take shape. Push it until it breaks.\n\nCASE 42 4666840 Glowbot Tower Defense 2026_H1 tower\nCompetitive endless sci-fi tower defense game with global leaderboards where you strategically place and upgrade Glowbots to stop waves of enemies across vibrant space maps.\nDefend. Survive. Compete. The Demon King's army is invading the galaxy, but simply surviving isn't enough. In Glowbot Tower Defense , you'll build powerful defensive networks of glowing robotic defenders, battle endless waves of monsters, and compete against players around the world for leaderboard supremacy. Every map features global Steam leaderboards that track your highest wave survived, turning every run into a battle for ranking, prestige, and rare achievements. Discover powerful strategies, optimize your defenses, and push farther than anyone else. Only the greatest commanders will reach the top. Climb the Global Leaderboards Every map in Glowbot Tower Defense features fully integrated online Steam leaderboards. Your highest wave survived is permanently recorded, allowing you to compete against players worldwide and prove your defensive mastery. As strategies evolve and players push deeper into enemy waves, rankings constantly shift, creating an ongoing challenge to remain among the best defenders in the galaxy. Whether you're competing with friends or challenging the world's top players, every wave matters. Ongoing Community Tournaments The battle never ends. Glowbot Tower Defense features regularly hosted leaderboard competitions and community tournaments where players can compete for exclusive recognition, Discord titles, and permanent placement in the game's Hall of Fame. Weekly and seasonal challenges encourage players to return, discover new strategies, and fight for the top positions on the global leaderboards. Whether you're aiming for a Top 10 finish or battling for the #1 spot, every tournament offers a new opportunity to prove your defensive mastery. The next champion could be you. Earn Prestigious Competitive Achievements Glowbot Tower Defense features a collection of rare competitive achievements designed for the most dedicated players. Competitive Achievements Include Reach the Top 10 on any map leaderboard Reach the Top 10 on all map leaderboards Become the #1 ranked player on any map Become the #1 ranked player across all maps These achievements are intentionally difficult and highly competitive. Reaching the top requires mastery of tower placement, resource management, upgrade strategies, and wave optimization. Only the greatest defenders will earn them. Build the Ultimate Defense Strategically place Glowbots along twisting space pathways to stop the Demon King's forces before they overrun your stars. Every tower serves a unique purpose, and success depends on building efficient defensive combinations that can survive increasingly difficult enemy assaults. Deploy Powerful Glowbots Command a variety of specialized robotic defenders, including: Energy-based attack towers Chain lightning towers Explosive area-damage towers Support units that enhance nearby defenses Experiment with different combinations to discover powerful synergies and create unstoppable defensive networks. Survive Endless Enemy Assaults The Demon King's army grows stronger with every wave. Face a variety of enemy types including: Fast-moving attackers Heavy armored units Specialized enemies with unique behaviors Massive boss monsters that test your defenses to their limits The farther you survive, the higher you climb. Explore Multiple Space-Themed Battlefields Defend your stars across vibrant neon-filled environments, each featuring unique layouts, strategic opportunities, and leaderboard rankings. Every map presents new challenges and requires different approaches to achieve the highest possible wave count. Upgrade and Optimize Earn diamonds through gameplay and strengthen your defenses. Upgrade towers, refine your strategies, and optimize your layouts as enemy waves become increasingly overwhelming. Small improvements can mean the difference between defeat and a new leaderboard record. Play Your Way Increase the game speed to accelerate the action and test your defenses under maximum pressure. When hundreds of enemies flood the battlefield, only the strongest strategies will survive. Can You Become the Galaxy's Greatest Defender? Build. Upgrade. Adapt. Compete against players around the world, climb the rankings, and claim your place among the greatest Glowbot commanders ever recorded. The Demon King's invasion never stops. How long can your defenses survive? 🌌 Join the official Glowbot Tower Defense Discord community: ⚡ Share strategies 🐞 Report bugs 🏆 Compete with other players 🎮 Follow development updates\n\nCASE 43 4785570 Ghosts Keep Out 2026_H1 tower\nDefend your base in this Roguelite survival shooter TD! Mix random skills + evolve gems to craft your Build. 5-min fast rounds, intense zombie & ghost hordes. Triple progression: Spirits, guns, heroes. Easy to learn, hard to put down. Low barrier, high ceiling. Every run is fresh!\nA casual strategy game that blends Tower Defense with Bullet Hell Shooter gameplay. As one of the few remaining human survivors, build your defenses, upgrade talents, combine hundreds of gems and Fantoms, hold the line against relentless ghost hordes — and stay alive. 🧠 Strategy × Tower Defense × Bullet Hell Every run is a brand new survival challenge. Enemies swarm like tidal waves. You must choose skills wisely, find your firing window amidst the bullet chaos, and never forget to defend your base core. ✨ Talent Builds · Freedom to Combine The game offers multiple talent paths: Gunslinging DPS, Fire, Ice, Lightning, Energy — each can stack and trigger chain reactions with one another. Craft your own Fire-Lightning Storm, Ice Explosion, Machine Gun Physical DPS build… different builds, completely different experiences! And remember — matching the right counter build against different enemies doubles your efficiency~ 💎 Gem System · Power Up Your Build Dozens of gems to mix and match. Socket them to unlock corresponding active skills or stat growth. Every run feels fresh and exciting. 🧬 Progression Systems Fantom – A spirit unit that fights alongside you, providing passive bonuses and an active skill. Gun System – Different weapon types offer unique bullet patterns and handling. Multiple Characters – Each with their own unique talent skill, affecting your early-game strategy and mid-game transitions. ⚡ Easy to Pick Up, Quickly Addicting Each round lasts only 5 minutes — fast-paced, high-feedback gameplay. No need to worry about failure — just restart quickly and adjust your build. Perfect for short play sessions, but also for that \"just one more run\" immersive experience. 👻 You'll Love This If You… Want to try a fresh Tower Defense + Bullet Hell hybrid Love theorycrafting builds, studying numbers, and experimenting with combinations Want a game that's relaxing but not brain-dead, addictive but not stressful Enjoy the rhythm of \"one more run and I'll win\" 🔥 Join Now Hold the base, combine talents, socket gems, summon Fantoms. In this apocalypse, the only truth is survival. Советская армия,Немецкая армия,Технологии,Исторические бои,Heroes of Might and Magic V,Герои Меча и Магии 5,Герои,Армии,Прокачка героев,Beholder,Бихолдер,Приключение,Тоталитаризм,Управление домом,Шпионаж,Моральные дилеммы,Crossout,Кроссаут,Автомобильные бои,Крафт автомобилей,Оружие на автомобилях,Создание автомобилей,Бои на больших картах,Loop Hero,Луп Геро,Карточная стратегия,Рогалик,Циклические бои,Простота управления,Глубокий геймплей,Cities: Skylines,Города: Небоскребы,Управление городом,Транспорт,Жители,Экономика,Экология,Красивые города,Russian Fishing 4,Русское Рыбление 4,Симулятор рыбалки,Рыбалка,Разные виды рыбы,Рыбаки,Карты для рыбалки,Снасти для рыбалки,Коллекционирование рыб,Лёгкий геймплей,Euro Truck Simulator 2,Евро Трак Симулятор 2,Симулятор вождения,Грузовики,Европа,Вождение грузовика,Транспорт,Карты Европы,Перевозки,Farming Simulator 22,Фармерский Симулятор 22,Симулятор фермы,Ферма,Аграрный симулятор,Тракторы,Растения,Животные,Симулятор Русской Деревни,Симулятор жизни,Русская деревня,Деревня,Огород,Грибы,Ягоды,Дрова,Охота,Самогон,Жизнь в деревне,World of Warships,Мир Кораблей,Симулятор кораблей,В…6568 tokens truncated…，好評如潮，ポジティブ，ネガティブ，開発，制作，インディー開発，デザイン，プログラミング，グラフィック，音楽，サウンド，キャラクター，キャラデザ，ストーリー，ストーリー性，感動，泣ける，面白い，楽しい，刺激的，緊張感，リラックス，療癒，育成，育成シミュ，ペット育成，キャラ育成，農園，農業シミュ，街づくり，都市開発，建造，ビルディング，冒険，探検，謎解き，推理，戦争，歴史，時代劇，魔法，剣と魔法，魔物，ドラゴン，エルフ，ドワーフ，人間，ゾンビ，ヴァンパイア，悪魔，天使，神，女神，英雄，勇者，魔王，悪役，正義，悪，善悪の選択，選択肢，多結末，マルチエンディング，隠し要素，隠しキャラ，隠しステージ，アチーブメント，トロフィー，課金，ガチャ，アイテム，武器，防具，アクセサリー，魔法，スキル，アビリティ，レベルアップ，経験値，ゴールド，通貨，クエスト，メインクエスト，サイドクエスト，ダンジョン，ボス戦，ボス，最終ボス，難易度，イージー，ノーマル，ハード，ナイトメア，難易度調整，オプション，コントローラー対応，キーボードマウス，グラフィック設定，解像度，フレームレート，60fps,4K,HD, フル HD,VR, バグ，不具合，修正パッチ，アップデート，アップグレード，ゲームパッド，周辺機器，椅子，ヘッドセット，イヤホン，マイク，スティック，ジョイスティック，ペダル，ソフト，ダウンロード，インストール，起動，セーブ，ロード，オートセーブ，クイックセーブ，セーブデータ，バックアップ，クラウドセーブ，オンライン，オフライン，LAN 対戦，ランキング，順位，トップ，ベストセラー，売れ筋，人気ランキング，ダウンロードランキング，レビューランキング，新作ランキング，無料ランキング，割引ランキング，ジャンルランキング，アクションランキング，RPG ランキング，シミュレーションランキング，ストラテジーランキング，アドベンチャーランキング，ホラーランキング，恋爱ゲームランキング，音楽ゲームランキング，格闘ゲームランキング，レースゲームランキング，シューティングランキング，MMORPG ランキング，インディーゲームランキング，日本語ゲームランキング，日本製ゲームランキング,Escape from Tarkov,Saída de Tarkov,jogo russo,tiro tático,supervivência,tiro em primeira pessoa,jogo online,multijogador,classificação 18+,tiro difícil,Tarkov,combates corpo a corpo,coleta de loot,upgrade de personagem,supervivência online,desenvolvedor russo,batalha pela sobrevivência,tática,disparos,armas,munição,mapa de Tarkov,jogo com legendas em português,jogo compatível com português,tiro realista,modos de jogo competitivos,jogo de tiro para pc,steam,Atomic Heart,Coração Atômico,ação,ficção científica,mundo aberto,antitopia,futurismo,robôs,gráficos bonitos,tiro com história,jogo single player,cooperativo,mecânicas de combate,magia e armas,energia atômica,pós-apocalipse,história da Rússia,jogo clássico russo,dublagem em português (brasil),jogo de ficção científica,pc,Call of Duty: Modern Warfare II,Call of Duty: Guerra Moderna II,jogo de tiro,franquia clássica,guerra moderna,ações militares,PVP,batalhas táticas,mapa mundial,modos de jogo,história single player,skins,armas personalizadas,Apex Legends,battle royale,legendas,combates dinâmicos,mapas de batalha,competições,ranked,cosméticos,jogo free to play,f2p,Ill Will,Il Will,tiro old school,tiro boomer,classificação 100% no Steam,tiro brutal,música,design de níveis,dificuldade alta,versão demo,jogo pago,combates com monstros,jogo indie russo,compatível com português,Hogwarts Legacy,Herança de Hogwarts,jogo de RPG,tiro em terceira pessoa,fantasia,magia,Hogwarts,aprendizado de feitiços,excursões por Hogwarts,coleta de itens,criaturas fantásticas,história envolvente,personagens customizáveis,Baldur's Gate 3,CRPG,Dungeons & Dragons,escolhas e consequências,classes de personagens,combates em tempo real,combates por turnos,história épica,Elden Ring,difícil,fantasia épica,dragões,combates com chefes,exploração,armas e armaduras únicas,danças com espada,desenvolvedor FromSoftware,ATOM RPG,RPG difícil,visão de cima,coleta de recursos,crafteo,dialogos com NPC,mundo pós-apocalíptico,tema russo,idioma português,jogo indie,Black Book,Livro Negro,mitologia eslava,bruxa,combates com cartas,upgrade,coleta de cartas,missões,leyendas eslavas,arte bonita,RPG de cartas,compatível com português (brasil),Civilization VI,estratégia por turnos,construção de cidade,gerenciamento de recursos,culturas e civilizações,exploração,guerras,diplomacia,Age of Empires IV,estratégia em tempo real,estratégia histórica,exércitos,civilizações históricas,eventos históricos,construção de base,técnicas militares,StarCraft II,estratégia espacial,três raças,Zerg,Protoss,Terranos,Crossout,combates com veículos,crafteo de veículos,armas em veículos,criação de veículos,combates em mapas grandes,PVP,PVE,Beholder,aventura,totalitarismo,gerenciamento de casa,espionagem,dilemas morais,estilo bonito,Euro Truck Simulator 2,simulador de caminhão,simulador de direção,caminhões,Europa,direção de caminhão,transporte,mapas da Europa,entregas,personalização de caminhão,Farming Simulator 22,simulador de fazenda,agricultura,tratores,plantas,animais,colheitas,gerenciamento de fazenda,equipamentos agrícolas,Russian Fishing 4,simulador de pesca,pesca,tipos diferentes de peixes,pescadores,mapas para pescar,ferramentas de pesca,coleta de peixes,gameplay leve,jogo gratuito,f2p,Cities: Skylines,simulador de cidade,gerenciamento de cidade,transporte,moradores,economia,ecologia,cidades bonitas,construção de infraestrutura,Simulador de Aldeia Russa,simulador de vida,aldeia,horta,cogumelos,frutas,lenha,caça,licor caseiro,vida no campo,Tetris,jogo clássico,quebra-cabeça,jogo lendário,Aleksei Pajitnov,URSS,entretenimento,quebra-cabeça clássico,Minecraft,jogo de construção,blocos,modo criativo,modo sobrevivência,construções,mods,Stardew Valley,jogo de fazenda,casual,simulação de vida,amigos,relacionamentos,festivais,Among Us,jogo de espionagem,detetive,impostor,tarefas,Pummel Party,jogo de festa,minijogos,competição,jogar com amigos,Escape from Tarkov,Escapar de Tarkov,juego ruso,disparo táctico,supervivencia,disparo en primera persona,juego online,multijugador,clasificación 18+,disparo difícil,Tarkov,combates cuerpo a cuerpo,colección de botín,mejora de personaje,supervivencia online,desarrollador ruso,batalla por la supervivencia,táctica,disparos,armas,munición,mapa de Tarkov,juego con subtítulos en español,juego compatible con español,disparo realista,modos de juego competitivos,juego de disparos para pc,steam,Atomic Heart,Corazón Atómico,acción,ciencia ficción,mundo abierto,antitopía,futurismo,robots,gráficos bonitos,disparo con historia,juego single player,cooperativo,mecánicas de combate,magia y armas,energía atómica,postapocalipsis,historia de Rusia,juego clásico ruso,doblaje en español (latinoamérica),juego de ciencia ficción,pc,Call of Duty: Modern Warfare II,Call of Duty: Guerra Moderna II,juego de disparos,franquicia clásica,guerra moderna,acciones militares,PVP,batallas tácticas,mapa mundial,modos de juego,historia single player,skins,armas personalizadas,Apex Legends,battle royale,leyendas,combates dinámicos,mapas de batalla,competencias,ranked,cosméticos,juego free to play,f2p,Ill Will,Il Will,disparo old school,disparo boomer,clasificación 100% en Steam,disparo brutal,música,diseño de niveles,dificultad alta,versión demo,juego pagado,combates con monstruos,juego indie ruso,compatible con español,Fortnite,construcción,combates en tiempo real,eventos en línea,modos de juego creativos,mapas personalizados,juego popular en latinoamérica,Hogwarts Legacy,Legado de Hogwarts,juego de RPG,disparo en tercera persona, fantasía,magia,Hogwarts,aprendizaje de hechizos,excursiones por Hogwarts,colección de objetos,criaturas fantásticas,historia envolvente,personajes personalizables,Baldur's Gate 3,CRPG,Dungeons & Dragons,elecciones y consecuencias,clases de personajes,combates en tiempo real,combates por turnos,historia épica,Elden Ring,difícil, fantasía épica,dragones,combates con jefes,exploración,armas y armaduras únicas,danzas con espada,desarrollador FromSoftware,ATOM RPG,RPG difícil,visión cenital,colección de recursos,crafteo,dialogos con NPC,mundo postapocalíptico,tema ruso,idioma español,juego indie,Black Book,Libro Negro,mitología eslava,bruja,combates con cartas,mejora,colección de cartas,misiones,leyendas eslavas,arte bonito,RPG de cartas,compatible con español (latinoamérica),The Life and Suffering of Sir Brante,La Vida y el Sufrimiento del Señor Brante,RPG textual, fantasía histórica,estilo bonito,estilo literario,RPG narrativo,juego de texto,historia profunda,Civilization VI,estrategia por turnos,construcción de ciudad,gestión de recursos,culturas y civilizaciones,exploración,guerras,diplomacia,Age of Empires IV,estrategia en tiempo real,estrategia histórica,ejércitos,civilizaciones históricas,eventos históricos,construcción de base,técnicas militares,StarCraft II,estrategia espacial,tres razas,Zerg,Protoss,Terranos,Crossout,combates con vehículos,crafteo de vehículos,armas en vehículos,creación de vehículos,combates en mapas grandes,PVP,PVE,Beholder,aventura,totalitarismo,gestión de casa,espionaje,dilemas morales,estilo bonito,Pathfinder: Kingmaker,Pathfinder: Rey Mago,CRPG,mundo abierto,criaturas fantásticas,ubicaciones hermosas,doblaje en español (latinoamérica),juego de rol,Euro Truck Simulator 2,simulador de camión,simulador de conducción,camiones,Europa,conducción de camión,transporte,mapas de Europa,entregas,personalización de camión,Farming Simulator 22,simulador de granja,agricultura,tractores,plantas,animales,cosechas,gestión de granja,equipos agrícolas,Russian Fishing 4,simulador de pesca,pesca,tipos diferentes de peces,pescadores,mapas para pescar,herramientas de pesca,colección de peces,gameplay ligero,juego gratuito,f2p,Cities: Skylines,simulador de ciudad,gestión de ciudad,transporte,habitantes,economía,ecología,ciudades bonitas,construcción de infraestructura,Simulador de Aldea Rusa,simulador de vida,aldea,huerto,hongos,frutas,leña,caza,licor casero,vida en el campo,Tetris,juego clásico,rompecabezas,juego legendario,Alekséi Pázhitnov,URSS,entretenimiento,rompecabezas clásico,Minecraft,juego de construcción,bloques,modo creativo,modo supervivencia,construcciones,mods,Stardew Valley,juego de granja,casual,simulación de vida,amigos,relaciones,fiestas,Among Us,juego de espionaje,detective,impostor,tareas,Pummel Party,juego de fiesta,minijuegos,competición,jugar con amigos,Fall Guys,Chicos Caídos,juego popular en streams de latinoamérica\n\nCASE 44 3998900 Claws Of Power 2026_JulAug tower\nClaws of Power is a cute, real-time card battle game. With a variety of unique and adorable cat cards, each one could be your key to victory. Strategize by building different decks and using items to overcome challenges and claim the throne as the ultimate cat leader.\nClaws of Power is a real-time combat casual card game with an adorable art style. Due to certain circumstances, you must become the boss of this turf. Collect Cats Throughout the game, you will collect various feline characters, each possessing unique skills. Once they become part of your paw-squad, spend dried fish to summon them into battle. Upgrade Your Buildings Enhance your base to increase its health, and upgrade your Dried Fish Factory to boost production speed—all of which will greatly aid you in combat. Purchase Items Acquire special skills that can be used during battles. Equipment matching Try different equipment combinations to maximize your cat's abilities. Lord Challenge If you feel powerful enough, you can challenge powerful lords and receive generous rewards! Socket Gems Conquer one stronghold after another and dominate this land as the ultimate top cat!\n\nCASE 45 4402950 HEXSTORM: Tears of Arcadia 2026_JulAug tower\nHEXSTORM: Tears of Arcadia is a tactical tower defense with roguelike structure. Build your flying fortress, upgrade your ship, and combine powerful boosts. Every battle is unique. Uncover the mystery of the fallen captain.\nBuild a flying fortress and defend it from every direction. HEXSTORM: Tears of Arcadia is a tactical tower defense with a roguelite structure where you command the flying fortress Arcadia and build a unified combat network of turrets, support modules, and resource collectors around it. This is not a tower defense about fixed paths and passive waiting. The battlefield evolves with you. Enemies attack from all directions, your economy depends on territory control, and one wrong structure can leave an entire sector of your defense without power. At the center of everything is Arcadia . It is your ship, your base, and the heart of the entire system. Expand the network, harvest resources, strengthen your weapons, and assemble combat combinations that change the course of every mission. Build a Living Defense System Instead of placing isolated towers, you construct an interconnected infrastructure. Collectors expand your territory, power the network, and extract resources. Turrets cover weak directions. Support modules reinforce your defenses and unlock new tactical options. Every decision has a cost. Invest resources into economic expansion or urgently seal a breach in your defense. Focus on income or commit everything to firepower. You cannot simply place towers and watch. Survival depends on maintaining the balance between growth and defense. Two Mission Types, Two Combat Rhythms In ground missions , Arcadia lands on the surface and you deploy your defenses around the ship, holding back waves of enemies on a static battlefield filled with crystal deposits. In air missions , the ship continues its flight. The world scrolls beneath you, resources drift past, hostile missile batteries operate in the background, and rare depot containers give you the chance to dramatically reshape your build in the middle of battle. Both modes demand different pacing and decision-making, but they share one rule. Arcadia must survive. Every Mission Builds a New Strategy During aerial operations you discover depot bonuses and choose between several upgrades. Some provide immediate effects, while others stack and reshape your playstyle. You can build a defense focused on lightning strikes, shields, and repairs. You can accelerate turret fire rate and turn the perimeter into a relentless barrage. Or you can focus on aggressive drone assaults and precision elimination of high-priority threats. Even familiar missions feel different when Arcadia enters battle with a new combination of upgrades. Enemies Demand Priorities Every enemy type introduces a specific threat. Some rush the base and detonate on impact. Others circle the battlefield and fire from range. Heavy armored units break through the front line, saboteurs disable critical structures, and missile batteries force you to constantly monitor the safety of your perimeter. Victory does not belong to the player who builds the most turrets. It belongs to the one who reads the battlefield and reacts to the right threat at the right moment. Upgrade Arcadia Between Missions After each battle you earn meta progression that strengthens Arcadia for future operations. Upgrade the hull, generators, starting resources, turret firepower, and drone combat systems. Over time Arcadia transforms from a damaged survivor into a fully operational flying fortress capable of surviving increasingly dangerous missions. Campaign and Challenge Mode Experience a story campaign that gradually unlocks new mechanics, structures, and defense systems, as well as a challenge mode with endless waves for players who want to push their builds to the limit. HEXSTORM: Tears of Arcadia is a game about pressure, rhythm, and constant tactical decisions. A defense system built in motion. A ship that grows stronger after every battle. And a world whose story has only just begun.\n\nCASE 46 4799450 Kingdoms vs Zombies 2026_JulAug tower\nKingdoms vs Zombies is a blend of strategy, tower defense, and roguelike gameplay, where you must defend your kingdom against endless hordes of the undead.\nWhen night falls, the dead rise to hunt. With each passing day, their numbers grow, and their attacks become more relentless. Only careful resource management, a well-planned defense, and a powerful army will help you survive the looming threat! Build fortifications, construct towers, strengthen your walls, and command your troops as you fend off wave after wave of enemies. Every run unfolds differently: choose random upgrades, create unique build combinations, and adapt to ever-changing battlefield conditions. Develop your settlement, unlock new opportunities for your army, and transform a small village into a mighty fortress. Experiment with different strategies, combine various unit types, and discover the best ways to stop the endless zombie invasion. Will you withstand the undead onslaught and build an impregnable kingdom, or will your walls crumble beneath countless enemies? Features: A unique blend of strategy, tower defense, and roguelike mechanics Massive army-versus-army battles Random upgrades and unique builds in every run Build and expand your own kingdom Diverse units, towers, and defensive optionsWaves of increasingly dangerous enemies High replayability through random events and upgrades Every decision matters. One smart choice can save the kingdom. One mistake, and the zombies will reach your gates. Only you can defeat the endless hordes and emerge victorious!\n\nCASE 47 4918020 Orbital Tower Defence 2026_JulAug tower\nBuild towers on rotating orbital rings to defend your planet. Discover powerful upgrades, battle elite enemies and bosses, and master unique planets with distinct challenges.\r\n\nOrbital Tower Defence Defend your planet from every direction. Build powerful orbital defenses on rotating rings surrounding your planet. As enemy fleets close in from all angles, every placement matters—and every rotation changes the battlefield. Create devastating combinations by unlocking dozens of towers, powerful upgrades, and branching upgrade paths. Adapt your strategy to survive relentless enemy waves, elite variants, and screen-filling boss encounters. No two planets are the same. Each world features unique orbital layouts, special abilities, and new tactical challenges that force you to rethink your defenses. Features Unique Orbital Gameplay – Build towers on rotating rings that constantly reposition your defenses. Many Towers & Upgrades – Discover a wide variety of towers with branching upgrade paths. Elite Enemies & Epic Bosses – Face dangerous elite units and massive bosses that demand new strategies. Elite Upgrades – Earn powerful upgrades to push your builds beyond their limits. Defend Multiple Planets – Every planet offers distinct ring patterns, abilities, and difficulty levels. Endless Build Variety – Experiment with different tower combinations and synergies on every run. Clean Geometric Visuals – A minimalist style where everything is generated from simple shapes, keeping the action clear and readable. Can you build the ultimate orbital defense and protect humanity's last worlds?\n"}

**Checking example specificity**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"364bca","wall_time_seconds":0.00000276,"exit_code":0,"original_token_count":5252,"output":"\nCASE 48 1004640 FINAL FANTASY TACTICS - The Ivalice Chronicles 2025 other\nFeaturing fully voiced dialogue, beautifully enhanced graphics, and countless gameplay improvements, the genre-defining tactical RPG makes its long-awaited return.\nIvalice—a kingdom blessed by the light of the gods and ruled by the Two-headed Lion. A year after her defeat to Ordallia in the Fifty Years' War, the king succumbed to malady, leaving a mere boy of two to ascend the throne. Whoever was named regent would rule in truth, and thus emerged two rival candidates: dukes Goltanna and Larg, whose deeds during the war had earned them great renown. Goltanna the Black Lion and Larg the White Lion were destined to clash, and the resulting conflict would come to be known as the War of the Lions. Against this backdrop appear Ramza, third son of House Beoulve, one of Ivalice's leading military families, and his childhood companion Delita, a commoner raised amongst nobles. As they endeavor to chart a course through this tumultuous age, a mighty current draws them ever closer to the abyss. Lead your party in exhilarating encounters where strategic positioning and forward planning determine the outcome. Use the varied terrain of 3D battlefields to your advantage, skillfully manipulate the turn order, and victory will be yours for the taking. Combine a wide variety of jobs and abilities to craft your very own strategies. As your party members grow, they will gain access to more jobs, each with its own unique abilities. Once learned, abilities can be used even after changing to another job. With over 20 jobs to master—including white mage, black mage, dragoon, and many others from the FINAL FANTASY series—and hundreds of abilities to learn, you have the freedom to experiment and discover the combinations that best suit your party. Chocobos and other creatures can also be tamed and led into battle. With improved graphics, fully voiced dialogue, and refined gameplay, the enhanced version provides the perfect way to immerse yourself in the world of Ivalice. Along with the fine-tuning of several features and battle mechanics, the inclusion of the more accessible \"Squire\" difficulty setting makes the enhanced version an excellent starting point for first-time players. The classic version unites the graphics and gameplay of the original 1997 version with the renowned War of the Lions translation for a unique way to experience this beloved title. The Enhanced version of the game supports a wide variety of controller options. It also supports keyboard and mouse controls. The Classic version of the game only supports XInput controllers. Keyboard controls are also supported, but mouse controls are not supported.\n\nCASE 49 3540430 Dungeon Weiqi 2025 other\nDungeon Weiqi is a card roguelike strategy game. In this game, the player assume the role of a kingdom's lord. Through your journey, you'll collect cards, construct decks, progressively empower your capabilities, and confront escalating challenges.\"\nDungeon Weiqi is a pixel-art roguelike card game primarily inspired by the ancient game of Weiqi. Weiqi Style Card Combat Deploy cards using Crystals that regenerate automatically each turn. Once placed on the chessboard, cards autonomously attack adjacent enemies at round's end. Try to occupy advantageous positions and surround the enemy to create more advantages! Your Lord enters battle as a card. Will you use the lord's powerful abilities to charge forward, or protect them in the rear and seek victory? The choice is yours! Diverse Factions & Card Types Players can choose from multiple factions. You can collect unique cards, and form the most suitable deck to fight! There are three types of cards for both the enemy and us: Warrior Cards | Skill Cards | Structure Cards. Master their synergies to overturn desperate situations! Enemies also possess unique cards and combat styles. Be careful to deal with them and familiarize yourself with their combat styles and weaknesses. Then victory is yours! Explore the Map Freely Avoiding strong enemies or engaging in direct combat? The decision is entirely yours! You can freely plan your path, choose the order of battle with enemies, or even completely ignore them (if you are willing to give up the treasure they are guarding). Be careful with your supplies. Every step further will cost your food. If the food runs out, you will be at a huge disadvantage and even lose the entire game! Plan your path carefully to avoid the worst-case situation.\n\nCASE 50 3604960 Legacy of Shadows 影の遺産 2025 other\nA survival horror exploration game. The debt-ridden protagonist inherits his late uncle’s mansion to find valuables and sell them. But something sinister lurks inside. Can you gather items, evade ghostly threats, and make it out alive? Tense gameplay where exploration meets survival.\nLegacy of Shadows A decaying mansion deep in the mountains. Leftover relics. And a lingering, ominous presence— You inherit your late uncle’s house, only to discover it comes with a crushing debt. What was supposed to be a visit to sell off the remaining valuables and repay what you owe... turns into something far more unsettling. Explore, collect, and survive. Can you navigate the mansion, gather valuable items, and make it out safely? As you delve deeper, the house’s buried past begins to reveal itself... 💀 Tense Exploration – A survival experience where every item you collect could be your last. 👻 Lurking Terror – Evade the shadows that wander the halls of the mansion. 💰 Debt Repayment as Goal – Sell what you find and aim for financial freedom. 🔍 Uncovered Truths – The items you collect start to whisper secrets of a long-held grudge. 🧠 Growth & Choice – Learn new skills through exploration and tailor your growth to your playstyle. 🏆 Competition & Challenge – Compete with others in clear time and score for replayability. Can you repay the debt and escape the mansion? Or will you be swallowed by the shadows—?\n\nCASE 51 3927630 魔姬征服者  Twelve Succubus 2025 other\nSemi-Turn-Based RPG! Ruiya faces doom. You, interdimensional hero, are its last hope! Master skills & gear, battle charismatic witches. Defeat → Recruit → Unlock galleries.  endless monster battles with abundant loot, plus unlimited revives and   to carve your legend!\nThe Ruiya Continent has been invaded by hordes of succubus! Otherworldly heroes are summoned to confront this threat. Light on narrative, your mission is clear 🗡️ Conquer Succubus Boss Defeat witch bosses in each stage to subdue and recruit them! Recruited succubus: Fight alongside you Unlock exclusive galleries! ⚡ Unrestricted Skill System Mix and match countless skills freely with no limitations. Forge your ultimate squad through limitless customization! ✨ Gear Enhancement Empower equipment with game-changing special effects through upgrades. 👁️ Visible Encounters No annoying random battles! Spot enemies on the field and engage strategically. 🔄 Stress-Free Progression Auto-replenish HP/MP before every battle Zero penalties upon defeat Auto-battle mode\n\nCASE 52 2529790 GRIME II 2026_H1 other\nBecome the stealer of forms in GRIME II, a surreal action-adventure metroidvania. Launch tendrils made of hands to absorb foes and summon molds in their shape, as you explore a bizarre lived-in-world obsessed with art.\nBecome a stealer of forms in GRIME II, the sequel to 2021's acclaimed action-adventure metroidvania. You are a Formless - an art mimic, absorbing creatures and summoning molds in their shape. Venture into a new and mysterious land in the GRIME universe, where danger and wonder await around every corner. Use your surroundings in tandem with your summoned molds to overcome deadly enemies and epic bosses, all the while exploring a deep world full of diverse cultures and characters. Create your own playstyle by choosing from a variety of abilities, weapons, and mold summons. Learn the makings of your enemies as you summon their form to both aid you in combat, as well as in exploring the various secrets paths of the world. Molds - Absorb defeated enemies, shaping them into molds which can then be used for special attacks. Use molds to throw enemies, fire projectiles, stun, and even turn into summons. Environmental Combat - Wherever you go, the environment provides ample opportunities to damage enemies nearby or from a distance. Be wary, for they can do the same to you! Parry and Grasp - Send hand-tendrils to either parry enemy attacks or grasp and absorb them, drawing you closer to learning their shape. A Lived-in World - Discover incredible locations made of painted nails and giant vases, each with its own unique civilization, culture, and deep characters to converse with. Customise your playstyle with over 30 weapons, 60 armours and 12 abilities.\n\nCASE 53 4402660 METRO QUESTER | OSAKA 2026_H1 other\nMETRO QUESTER | OSAKA is a dungeon exploration RPG based in a post-apocalyptic futuristic world created by the manga artist Kazushi Hagiwara, with a deep game system designed by Hironori Kato that offers the excitement and surprises reminiscent of 1980s computer games through hack and slash.\nMETRO QUESTER | OSAKA is a dungeon exploration RPG based in a post-apocalyptic futuristic world created by the manga artist Kazushi Hagiwara, with a deep game system designed by Hironori Kato that offers the excitement and surprises reminiscent of 1980s computer games through hack and slash. Moving the location to the city of water, Osaka, it features 32 characters and 8 classes, enjoyable combo skills, numerous customizable weapons, a bestiary, and features like NEW GAME+ for replayability, offering plenty of depth for dedicated gamers.\n\nCASE 54 4631030 TarotBullFight 2026_H1 other\nA bullfighting-style card game based on classic bull poker rules, featuring special abilities from select tarot cards or tarot-themed decks. Players can engage in turn-based battles against both robot opponents and other players to win matches.\nGame Description Step into a strategic card game centered around tarot! In TarotBullFight , you take control of the table, building your custom deck using all 22 Major Arcana and 56 Minor Arcana. Engage in bullfighting-style card battles against opponents from around the world, or test your wits and luck against AI robots. Victory isn’t just about luck here. Every move you make, every tarot ability you play at the right moment, will shape the outcome of the game. Key Features ✨ Tarot Deck Building Selected tarot cards come with unique effects and mechanics. Playing the right card at the right time can turn the tide of battle in unexpected ways. ⚔️ Bull-Style Card Duels Classic bull poker rules are deeply integrated with card abilities. You’ll need to calculate hands quickly, read your opponents, and win rounds with strategy—not just luck. Every match is full of twists and surprises; the outcome is never decided until the final play. 🌐 Online Multiplayer Play against friends or strangers from all over the world, and prove your skill and tactical mastery on the table. How to Play Build Your Deck: Leverage your strategic thinking to pick the right tarot cards, creating a deck balanced for attack, defense, and disruption. Join Bull Matches: In turn-based duels, play your cards to form the strongest possible hand combinations and defeat your opponents.\n\nCASE 55 4775190 BlassHole 2026_H1 other\nA space action, looting and ship progression game first made over 25 years ago while I was studying programming, now remade: explore systems, fight, collect loot and upgrade your ship module by module.\nBlassHole is a space action, looting and survival game where you start with a modest ship and carve your way through the galaxy by raiding convoys, defeating threats and collecting whatever drifts out of the wreckage. This is a game I made more than 25 years ago while I was studying programming. Now, finally, I have been able to remake it with everything I always imagined for it. Explore star systems, upgrade your ship with modules, buy and sell with merchants, gather credits and parts, and decide when you are ready to jump into more dangerous territory. Every fight can leave you with better weapons, engines, shields, hull plating or power cores. Survive long enough and your small ship can become a brutal machine. But the galaxy does not wait: aliens, raiders, hostile stations, minibosses and final bosses guard the most valuable systems. Features OpenGL 3D space combat with cannons, missiles, auto-turrets, shields, cloak and afterburner Module system with rarities, tiers and ship power management Skill progression for hull, shields, weapons, engines, fuel and navigation Galaxy systems with rising difficulty, merchants, special bosses and NG+ content Loot credits, parts and modules to sell, equip or fold into your build Ship codex and local achievement tracking prepared for future Steam integration Gameplay Loop Explore a system Find targets, merchants and threats Fight and collect loot Sell, buy and equip new modules Upgrade your ship and skills Jump to a more dangerous system Note BlassHole is a compact space action experience built for quick runs, steady progression and getting straight back into combat.\n\nCASE 56 4469500 Ninety Nine Dungeons 2026_JulAug other\nJapanese-style action RPG dungeon crawler adventure. Explore a top-down map enhanced with full verticality and multi-level dimensions. Discover rare equipment, customize your own unique build and playstyle, and conquer the ever-changing depths!\nChoose your starting class and embark on an epic adventure through dungeons brimming with magic and dragons. Defeat powerful bosses across the map, uncover hidden treasures, and combine them to forge your own unique build and playstyle. Whether you obliterate enemies in seconds or dominate the battlefield, plunge straight to the depths of the 99-layer procedurally generated dungeon! Every floor teems with deadly traps, formidable guardians, and unpredictable random events. Harness raw power, masterful techniques, or clever strategies—equip and optimize your one-of-a-kind build to conquer them all. Push your limits to the extreme and etch your own legendary tale of heroism. The ultimate challenge awaits—dive in now!\n\nCASE 57 4840190 VOIDRAVE 2026_JulAug other\nA bullet hell shmup with a secret. 100 waves, 20 bosses, and the slow realization that the sector you're purging isn't a sector. You are not crossing the void. You are inside it. Now with local co-op: two pilots, one screen.\nYou are not crossing the void. You are inside it. A routine purge run past the edge of the charts. The wrecks out here form spirals. The ice over the First Fleet tombs is warm. The deeper you fly, the less this place looks like space. \"Command says routine. Command is far away.\" 20 BOSSES Every five waves, something big. The Watcher. The Twin Pulse. The Coiled God. Each one has its own attack patterns, its own phases, its own place in the story. Because they aren't invaders. They're organs. And at the bottom waits the Void Heart, the heart of the thing you've been flying inside all along. 100 WAVES, ONE STORY Classic arcade bullet hell: swarms, elites, mini-bosses, wave mutators that change the rules mid-run. Between waves you get sector logs and transmissions from the First Fleet, the ones who flew here before you and stopped answering. \"They are our own voices. Played back wrong.\" Everything you kill goes into the library. Dozens of bestiary and codex entries, every enemy and boss documented. Half the story is hidden in there. MORE THAN ONE RUN BOSS RUSH - all 20 bosses back to back, no waves in between. DAILY RUN - one attempt per day, same seed for everyone, global leaderboard. You get one shot. WORLD BOSS - one giant boss, one health pool shared by every player in the world. Your damage counts. It has enrage phases. Good luck. 4 SHIPS Four ships, four weapon loadouts, four ways to die slower. Between waves, the upgrade shop opens: drones, weapons, shields, crits, and the limitless VOID OVERCHARGE. Every 10 waves you pick 1 of 3 mutations that bend the rules: vampiric rounds, ricochet, chain lightning, gravity wells. No two runs build the same. Find a synergy that breaks the game and ride it to wave 100. LOCAL CO-OP Bring a second pilot. The whole campaign, one screen, one controller and one keyboard. A tether runs between your ships and cuts enemy bullets while you stay close, charging a shared meter. Fill it, both press special within half a second, and the screen clears. The phase shield covers both of you, bombs and overdrive come from one pool, your ship special stays your own. Go down and you drift as a wreck until your partner flies over and brings you back. BUILT FOR ARCADE Graze scoring, combo multipliers, and a leaderboard that remembers everything. Three difficulty settings: from first-timer friendly to absolute punishment. Full controller support and Steam Deck compatible. Hand-crafted pixel art, original soundtrack, and zero microtransactions. Built from scratch in C++ and OpenGL. No engine. No middleware. Just the void. \"It knows your callsign. It knows your old one.\"\n\nCASE 58 4894180 Thrice-a-dice 2026_JulAug other\nTrice-a-Dice is a roguelike Dice-Builder. Roll dice to earn money. Raplace dice faces for better ones and combine into the best. Buy Cats to create synergies with faces and upgrades. Deal with the Devil, gamble in slot machine and buy Angel's help. All of this just to pay to play further.\r\n\nTrice-a-Dice is a roguelike Dice-Builder about building powerful synergies just to pay for one more round. Roll, build, and manage. Customize your dice, collect powerful Cats, and manage escalating costs. The goal is simple: generate enough income to stay ahead of the payments. Each run, you manage a collection of dice that determine your payouts. By swapping faces, buying Cats, and stacking synergies, you turn a basic roll into a high-yield economic engine. Customize Your Dice Your dice are modular. Instead of being locked to standard numbers, you can swap, replace individual faces or combine them to suit your strategy. Buy cats Cats would give money or even multiply them with certain conditions. They can make some your dice stronger or just copy one another. Meta-Progression Unlock characters with different starting set up. Each run can be impacted by random Blessings, Curses, and Slot Machine prizes. Increase difficulty by choosing from various rent curves and starting conditions. Decay Mode: Test your build in a never-ending loop and increasingly more difficult endgame. Unlock new faces, cats, and shop upgrades between runs, and track your best stats globally. Try them all 104 cats 35 base dice faces 19 special and combo-faces 45 round upgrades 6 characters 6 additional difficulties 16 Devil's curses and 9 Angel's blessings\n\nCASE 59 4973340 Neko Rogue 2026_JulAug other\nA cyberpunk roguelike deckbuilder set in a dystopian neon city. Build a deck, sabotage enemy Intent Cards, steal block and cut damage, then loot defeated foes’ cards for the next fight. Outplay unique enemy decks, climb hostile districts, and conquer the underworld—one hand at a time.\nNeko Rogue is a roguelike deckbuilder where every fight is decided by your hand, and every wrong play could end in disaster! This isn’t a simple reskin of “play cards, see a damage number, end turn.” In Neko Rogue, the enemy plays by the same rules as you do: CARDS. EVERY ENEMY HAS A DECK • Each foe brings a unique card set—no generic attack-block cookie-cutter packs. • Multi-enemy street fights and bosses each telegraph threat through real card play. • Rare finishers and specialized skills make every encounter feel like a different. INTENT CARDS, NOT JUST INTENT ICONS • Each round, enemies reveal an Intent Card on the field—not just a symbol. • You’ll read the board like a real duel: what’s coming, what’s blocked, what’s about to hurt. • Fight packs and bosses with style—and plan around plays you can actually see. PLAY THE ENEMY’S TURN BEFORE IT HAPPENS • Build a deck that hits hard—but also includes skills that target enemy Intent Cards. • Steal block points, reduce incoming damage, or negate an intent completely. • Outplay the board, not just the HP bar. LOOT THE ENEMY’S DECK • After victory, choose cards and skills from the enemy’s deck to add to your own. • Turn a street punk’s tricks—or a boss’s signature moves—into weapons for the next fight. • Grow stronger by absorbing the city’s underworld, one conquered deck at a time. A CITY THAT SOUNDS ALIVE • Adaptive music shifts with zones, standard fights, and boss battles. • Each map district and encounter type has its own sonic identity to pull you deeper into the run. One more hand. One more intent stolen. One more enemy card claimed. Build your deck. Break theirs. Take what’s yours.\n"}

**Reviewing entries for concreteness**

Script completed
Wall time 0.4 seconds
Output:

{}
{"chunk_id":"df9bed","wall_time_seconds":0.231878147,"exit_code":0,"original_token_count":62,"output":"label   concrete  generic  unclear\nfamily                            \naction         6        5        1\nauto           5        6        1\ndeck           6        4        2\nother          5        6        1\ntower          7        4        1\n"}

Script completed
Wall time 0.5 seconds
Output:

{"chunk_id":"1c9ac8","wall_time_seconds":0.379380669,"exit_code":0,"original_token_count":184,"output":"['appid', 'tagid', 'weight', 'rank', 'tag']\n['Action Roguelike', 'Automation', 'Base Building', 'Baseball', 'Based On A Novel', 'Board Game', 'Card Battler', 'Card Game', 'City Builder', 'Class-Based', 'Deckbuilding', 'Grid-Based Movement', 'Hex Grid', 'Inventory Management', 'Music-Based Procedural Generation', 'Party', 'Party Game', 'Party-Based RPG', 'Puzzle', 'Puzzle Platformer', 'Real Time Tactics', 'Resource Management', 'Roguelike', 'Roguelike Deckbuilder', 'Roguelite', 'Skateboarding', 'Snowboarding', 'Sports', 'Tactical', 'Tactical RPG', 'Team-Based', 'Text-Based', 'Trading Card Game', 'Traditional Roguelike', 'Transportation', 'Turn-Based', 'Turn-Based Combat', 'Turn-Based Strategy', 'Turn-Based Tactics', 'eSports']\n"}

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"1db2ec","wall_time_seconds":1.000681295,"session_id":6715,"original_token_count":0,"output":""}

Script completed
Wall time 3.8 seconds
Output:

{"chunk_id":"2f35e3","wall_time_seconds":3.83381957,"exit_code":1,"original_token_count":424,"output":"Traceback (most recent call last):\n  File \"/home/adn/a327ex/steam-research/continuation-2026-09-06/analyze.py\", line 161, in <module>\n    pilot=labels.merge(d[['appid','reviews','positive_pct','usd_list_price','text_words','quantity_max','store_url','developer','publisher']],on='appid',validate='one_to_one')\n  File \"/home/adn/a327ex/steam-research/.venv/lib/python3.14/site-packages/pandas/core/frame.py\", line 12902, in merge\n    return merge(\n        self,\n    ...<10 lines>...\n        validate=validate,\n    )\n  File \"/home/adn/a327ex/steam-research/.venv/lib/python3.14/site-packages/pandas/core/reshape/merge.py\", line 385, in merge\n    op = _MergeOperation(\n        left_df,\n    ...<10 lines>...\n        validate=validate,\n    )\n  File \"/home/adn/a327ex/steam-research/.venv/lib/python3.14/site-packages/pandas/core/reshape/merge.py\", line 1018, in __init__\n    ) = self._get_merge_keys()\n        ~~~~~~~~~~~~~~~~~~~~^^\n  File \"/home/adn/a327ex/steam-research/.venv/lib/python3.14/site-packages/pandas/core/reshape/merge.py\", line 1620, in _get_merge_keys\n    right_keys.append(right._get_label_or_level_values(rk))\n                      ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^\n  File \"/home/adn/a327ex/steam-research/.venv/lib/python3.14/site-packages/pandas/core/generic.py\", line 1757, in _get_label_or_level_values\n    self._check_label_or_level_ambiguity(key, axis=axis)\n    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^\n  File \"/home/adn/a327ex/steam-research/.venv/lib/python3.14/site-packages/pandas/core/generic.py\", line 1717, in _check_label_or_level_ambiguity\n    raise ValueError(msg)\nValueError: 'appid' is both an index level and a column label, which is ambiguous.\n"}

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"b91f6a","wall_time_seconds":1.001338155,"session_id":36528,"original_token_count":0,"output":""}

**Inspecting summary outputs**

Script completed
Wall time 0.0 seconds
Output:

{"chunk_id":"54fc02","wall_time_seconds":0.000001742,"exit_code":0,"original_token_count":2986,"output":"GROWTH\n  year       family  eligible  strict  inclusive  build_share  strict_median_words  all_median_words  english_support_share  unique_developers  median_price\n 2025          all      9140     843        985        0.092                289.0             225.0                  0.972                831          8.99\n 2025         deck       128     102        106        0.797                252.0             245.0                  0.980                102          9.99\n 2025       action       718     274        308        0.382                275.0             236.0                  0.985                272          6.99\n 2025         auto       157      73         84        0.465                274.0             241.0                  0.959                 73          8.99\n 2025        tower       314      64         74        0.204                280.0             226.5                  1.000                 64          9.99\n 2025 none_of_four      7931     386        477        0.049                309.0             224.0                  0.961                383          9.99\n 2026          all     12409    1717       1920        0.138                309.0             249.0                  0.983               1677          6.99\n 2026         deck       232     199        200        0.858                295.0             287.0                  1.000                199          9.99\n 2026       action      1068     565        594        0.529                289.0             260.0                  0.991                562          5.99\n 2026         auto       269     136        160        0.506                298.0             265.0                  1.000                135          6.99\n 2026        tower       439     132        151        0.301                273.5             252.0                  0.985                130          6.99\n 2026 none_of_four     10587     790        930        0.075                327.5             247.0                  0.970                771          7.99\nDECOMPOSITION\n         signature  n2025   n2026  b2025  b2026  share2025  share2026  delta  catalog_volume_composition  within_signature_build_share\n           action  660.0   972.0  243.0  511.0       0.37       0.53  268.0                      139.45                        128.55\n      action+auto    9.0    26.0    5.0   15.0       0.56       0.58   10.0                        9.63                          0.37\naction+auto+tower    5.0     5.0    2.0    2.0       0.40       0.40    0.0                        0.00                          0.00\n     action+tower   33.0    53.0   16.0   26.0       0.48       0.49   10.0                        9.75                          0.25\n             auto  105.0   160.0   48.0   76.0       0.46       0.48   28.0                       25.63                          2.37\n       auto+tower   26.0    49.0   10.0   18.0       0.38       0.37    8.0                        8.65                         -0.65\n             deck  100.0   188.0   83.0  160.0       0.83       0.85   77.0                       73.97                          3.03\n      deck+action   10.0    11.0    7.0   10.0       0.70       0.91    3.0                        0.80                          2.20\n deck+action+auto    1.0     1.0    1.0    1.0       1.00       1.00    0.0                        0.00                          0.00\n        deck+auto   10.0    25.0    6.0   22.0       0.60       0.88   16.0                       11.10                          4.90\n  deck+auto+tower    1.0     3.0    1.0    2.0       1.00       0.67    1.0                        1.67                         -0.67\n       deck+tower    6.0     4.0    4.0    4.0       0.67       1.00    0.0                       -1.67                          1.67\n     none_of_four 7931.0 10587.0  386.0  790.0       0.05       0.07  404.0                      163.73                        240.27\n            tower  243.0   325.0   31.0   80.0       0.13       0.25   49.0                       15.32                         33.68\nMAIN RECEPTION\n      period                 group    n  under10  ge100  liked100  ge1000  median_reviews  satisfied_among100  volume_ratio  joint_ratio  joint_text_ratio\n       2025            all_builds 1365      611    296       181      99            13.0               0.611         1.000        1.000             1.000\n       2025           deck_strict  169       38     62        41      14            40.0               0.661         1.379        1.502             1.496\n       2025          other_builds 1196      573    234       140      85            11.0               0.598         0.932        0.911             0.912\n       2025     deck_all_eligible  203       47     72        48      16            36.0               0.667           NaN          NaN               NaN\n       2025 deck_missed_by_strict   34        9     10         7       2            28.5               0.700           NaN          NaN               NaN\n    2026_H1            all_builds 1085      613    171       101      50             5.0               0.591         1.000        1.000             1.000\n    2026_H1           deck_strict  132       56     33        22       8            15.0               0.667         1.189        1.389             1.396\n    2026_H1          other_builds  953      557    138        79      42             5.0               0.572         0.963        0.928             0.927\n    2026_H1     deck_all_eligible  151       61     38        26       8            16.0               0.684           NaN          NaN               NaN\n    2026_H1 deck_missed_by_strict   19        5      5         4       0            19.0               0.800           NaN          NaN               NaN\n2026_JulAug            all_builds  632      418     75        45      19             3.0               0.600         1.000        1.000             1.000\n2026_JulAug           deck_strict   67       32     11        10       0            10.0               0.909         0.970        1.486             1.389\n2026_JulAug          other_builds  565      386     64        35      19             2.0               0.547         1.005        0.915             0.926\n2026_JulAug     deck_all_eligible   81       43     12        11       0             7.0               0.917           NaN          NaN               NaN\n2026_JulAug deck_missed_by_strict   14       11      1         1       0             5.5               1.000           NaN          NaN               NaN\nHYBRIDS\n           hybrid  present   n  ge100  liked100  liked100_rate  satisfied_among100  within_deck_joint_ratio\n   party_tactics     True 131     41        29          0.221               0.707                    1.104\n   party_tactics    False 237     65        44          0.186               0.677                    0.941\neconomy_building     True  35     17        10          0.286               0.588                    1.426\neconomy_building    False 333     89        63          0.189               0.708                    0.955\n       inventory     True  27     13        10          0.370               0.769                    1.748\n       inventory    False 341     93        63          0.185               0.677                    0.936\n     auto_battle     True  42     20        13          0.310               0.650                    1.668\n     auto_battle    False 326     86        60          0.184               0.698                    0.920\n   tower_defense     True  12      4         2          0.167               0.500                    0.784\n   tower_defense    False 356    102        71          0.199               0.696                    1.008\naction_roguelike     True  23      6         4          0.174               0.667                    0.756\naction_roguelike    False 345    100        69          0.200               0.690                    1.019\n             any     True 208     70        48          0.231               0.686                    1.141\n             any    False 160     36        25          0.156               0.694                    0.809\nPILOT\n                 axis       value  n  under10  ge10  ge50  ge100  ge200  ge1000  liked100  rate100  liked100_rate  satisfied_among100  median_reviews  median_positive_among100  median_price  developers  publishers  reviews_sum  top1_review_share  median_words  claims100\n               label    concrete 29       16    13     9      7      5       2         7    0.241          0.241               1.000             5.0                      88.0         8.990          29          29         8912              0.632         379.0          5\n               label     generic 25       15    10     5      4      1       0         2    0.160          0.080               0.500             5.0                      79.0         6.990          25          25          919              0.256         231.0          3\n               label     unclear  6        3     3     2      1      1       0         1    0.167          0.167               1.000            14.0                      96.0         5.845           6           6          373              0.740         334.0          1\nexplicit_interaction       False 42       24    18    11      9      5       1         7    0.214          0.167               0.778             6.5                      86.0         7.990          42          42         3978              0.468         295.5          6\nexplicit_interaction        True 18       10     8     5      3      2       1         3    0.167          0.167               1.000             4.0                      89.0         8.990          18          18         6226              0.905         392.0          3\n              family      action 12       11     1     0      0      0       0         0    0.000          0.000                 NaN             2.0                       NaN         8.490          12          12           52              0.481         388.0          1\n              family        auto 12        6     6     4      2      2       0         2    0.167          0.167               1.000            10.0                      91.0         7.990          12          12          653              0.366         375.5          3\n              family        deck 12        2    10     6      6      3       0         5    0.500          0.417               0.833            81.0                      87.5        10.490          12          11         1442              0.291         309.0          5\n              family       other 12       10     2     2      2      2       2         2    0.167          0.167               1.000             1.0                      87.5         7.490          12          12         7508              0.750         272.0          0\n              family       tower 12        5     7     4      2      0       0         1    0.167          0.083               0.500            33.5                      78.5         6.845          12          12          549              0.295         260.0          0\n              period        2025 20        9    11     8      6      4       1         5    0.300          0.250               0.833            23.5                      88.0         8.990          20          20         7059              0.798         343.0          2\n              period     2026_H1 20       12     8     4      3      2       1         2    0.150          0.100               0.667             7.5                      86.0         7.845          20          20         2463              0.756         295.0          1\n              period 2026_JulAug 20       13     7     4      3      1       0         3    0.150          0.150               1.000             3.0                      88.0         7.490          20          20          682              0.350         316.0          6\n"}

**Auditing 27 hybrid entries**

Script completed
Wall time 0.7 seconds
Output:

{"chunk_id":"70c14d","wall_time_seconds":0.489978731,"exit_code":0,"original_token_count":3631,"output":"\n 2468550 Vivid World 2025 \n A follow-up to the hit game Vivid Knight, Vivid World is a completely new party-building, roguelike strategy game!\r\nOur story takes place in the Underworld, a land ruled by demons.\r\nUtilize and master our protagonist, Lemuria's new Jewel Magic to conquer strange dungeons.\n\n\n 2500460 Inventorix 2025 \n INVENTORIX is an inventory-management money-making roguelike deckbuilder. Creatively pack and synergize items and alien perks to create astronomical wealth and pay a mysterious space debt!\nUse your INVENTORIX device (crafted by space gnomes) to extract maximum value from interstellar items and recruit aliens to boost your earnings. Pay off a mysterious debt or push your luck  | Loan Sharks Summary Key Features Galactic profits : Earn cosmic cash, ka-ching, ka-ching, ka-ching! Inventory Management : Pack items and aliens to uncover powerful combos. 95+ Items : Pick them wisely. 150+ Alien Perks : Which ones best support your build? 15+ Item Classes : Go | class or mix them up for cosmic synergies. Roguelike: Fast, replayable runs packed with variety. 10 INVENTORIX Devices : Unlock them and explore unique strategies with each. 20 Difficulty Levels : Conquer escalating challenges. Boss Encounter : Dubious Space Loan Sharks to defeat\n\n 2686020 Faith in Despair 2025 \n Forge your own cards with triggers, actions and modifiers. Shape your deck and gamble against Death itself. Make one mistake and you will descend into despair.\n\n\n 2916670 Neon Village 2025 \n This game is a fusion of a \"roguelike deck-builder\" and \"match-3 puzzle\". Collect unique residents to create explosive synergies and develop a village glittering with neon!\n\n\n 3177890 Rana Card 2025 \n Rana Card is a roguelike deck-builder where you play as a farm manager chasing TimeTech's KPI targets. Draw cards in high-stakes moments, survive Time Explosions, and unlock powerful cards. Build your deck, chart your own path, and face branching events and cataclysmic years on the road to success.\n\n\n 3215240 Probability Dungeon 2025 \n A turn-based roguelike where skills have a probability of execution. ¡You will have to keep improving your list of skills and decide in which order you want them to be executed! Set your strategy, create synergies and try to overcome the different threats.\n\n\n 3296910 Stash: A Card Looter 2025 \n Embark on a unique drafting adventure with an endless array of randomized cards inspired by classical ARPG-Loot.\r\nTransform your one-of-a-kind loot into powerful battle tactics! Take on enemy squads drafted by real players in a post-human world teeming with beasty conflicts.\n\n\n 3325030 Glass Cannon 2025 \n Challenging and chill roguelike where you plan your shots carefully and creatively eliminate enemies using different weapon combinations. Buy weapon mods to create powerful synergies and clear all 24 rounds.\n\n\n 3410180 Overlooting 2025 \n Adapt your strategy to a skill tree that changes every run, combine pieces of equipment to form powerful synergies and create a game-changing build capable of defeating the bosses that will try to stop you. In this Inventory Management Roguelike two games never feel the same!\n\n\n 3470920 Territors 2025 \n Enter the world of particles in this fast-paced card game, where the Proton and Neutron factions clash in intense matchups! Enhance your factions with new powerful particles, set and upgrade your deck, and prepare to face the ultimate challenge.\ntages based on their abilities, unique card powers and bonuses, enemy special traits, and a dynamic inventory size that adjusts based on your performance, enabling players to strategize and customize their decks for maximum impact. The game itself offers great replayability than\n\n 3573070 Toy Smash Kaboom! 2025 \n Toy mayhem unleashed! Arrange, combine, and craft the wildest toy army inside your desk, then crash into chaotic cartoon brawls across the tabletop!\nkes and magical blasts. Through each tabletop battle, fragments of your long-lost childhood return… Arrange, Combine, and Evolve Master the art of toy arrangement on your tabletop! Every toy slot is a strategic decision. Plan your layout, combine rare units, and arrange them w | and exclusive class abilities. Doll Chef: A kitchen defense expert who controls space and fortifies positions, perfect for defensive-minded players. Dr. Hamster: Commander of Hamster Legion that leads a swarm of tiny units to overwhelm enemies with relentless numbers. Puppet Mus\n\n 3665300 Rogue Ninjas 2025 \n In Rogue Ninjas, survival demands cunning and precision. Manage a squad of elite ninjas through twisting dungeons, strengthen your inventory with rare enhancers, and battle ruthless bosses in this dice-driven roguelike adventure.\ngeon crawls, tactical battles, and boss fights. Master the art of strategy by managing your squad’s inventory, upgrading with powerful enhancers, and using dice rolls to shift the odds in your favor. Every move matters as you face relentless enemies and forge your path to victor | -Driven Combat: Roll dice at the start of each turn and use the results to activate cards from your inventory. Match odd, even, or specific dice values to unleash powerful abilities and overcome enemies with smart dice management! Squad Management: Control a squad of unique ninj | as, each with distinct abilities and roles. Strategically swap their positions on the battlefield to maximize their strengths, protect vulnerable allies, and adapt to the flow of combat. Master positioning to turn the tide of battle! Inventory Load | out System: Strategically assign powerful cards from your inventory to each ninja. With limited slots and unique card effects, careful planning is crucial to maximize your squad’s potential in battle. Whether you're crawling through twisted dungeons, battling overwhelmin | g bosses, or managing your inventory to survive another run, Rogue Ninjas offers a fresh and replayable adventure in every playthrough.\n\n 4026180 REDTAPE 2025 \n Climb the corporate ladder in this roguelite deckbuilder. Recruit Employees, stack rule-bending abilities and bonuses, and purchase Items to boost your plays. Score massive point combos to game the system harder than the C-suite games you.\n\n\n 2652600 SealChain: Call of Blood 2026_H1 \n SealChain is a roguelike game where items can link together to form chain reactions. Items can also be arranged into specific formations to create seals, granting special effects.\nlChain is a roguelike game where items can link together to form chain reactions. Items can also be arranged into specific formations to create seals, granting special effects. The world is collapsing. Chaotic evil shrouds the land like thick purple mist. In this catastrophe, c |  other to enhance attributes Items can resonate with each other for additional effects Items can be arranged into specific formations to create seals, granting special effects Multiple playable characters, each with unique builds, skills, armor, and weapons Event system: obtain\n\n 3284290 Moonsigil Atlas 2026_H1 \n The deckbuilding roguelike with no energy – play cards using physical space! Slot cards together, manipulate the shape of your cards, and create overpowered combos to break the game!\n no limit to how many cards you can play, as long as they fit! Every card is a sigil you place on a grid, each with its unique shape, keywords, and effects - turning each combat into a tactical optimization challenge. Create powerful synergies through rune positioning, adja | e dozens of astral entities and three Titans that reshape the board itself. Hostile zones, shifting grids, and unique win conditions turn every encounter into a test of adaptation and mastery. A powerful deck is only half the battle…\n\n 3327710 Shadows of Wyrmwood 2026_H1 \n Shadows of Wyrmwood is a roguelike adventure RPG. Use spells, swords, and strategy to defeat every corrupted foe. Travel across procedurally generated maps that grow harder with every encounter. With each victory, obtain new armor, weapons, and spells. Do you have what it takes to save Wyrmwood?\n\n\n 3355940 Card Coder 2026_H1 \n Card Coder is a card-building roguelike: A novel mix of tactical card battles, inventory management and roguelike deckbuilding, where YOU create custom cards during play. Loot, shop & collect modular ability components to make truly unique cards out of 10 billion possibilities.\nCard Coder merges tactical card battles, inventory management and roguelike deck-building with a novel twist: Create your own custom cards during each run by constructing cybernetically enhanced warriors from all over th | e galaxy. Defeat challenging enemies to loot, shop and collect many different ability components. Slot them into your cards to define their behavior according to your playstyle: Discover, research and unlock a large variety of components to construct over 10 billion diffe | cked. But keep your Commander alive at all costs! Features Card Coder merges tactical card battles, inventory management and roguelike deck-building with a novel twist: Create your own custom cards during each run. B eyond deckbuilding : Create your own cards as you play by comb | sibilities. Build your cards to fit your playstyle -- not the other way round! Tactical combat with positioning : Manipulate enemy cards' targeting aggro by placing your cards strategically on the board. Take advantage of superior positioning. Meta Progression Unlock new ability c\n\n 3481020 Apokerlypse 2026_H1 \n A card-strategy roguelike inspired by a variety of poker games. You play as a highly skilled card shark in the Tavern at the End of the World, combining all kinds of demon chips to rewrite the rules of play, compete with players from across the globe, and ultimately rise to the top.\n\n\n 3802470 Turnbound 2026_H1 \n Turnbound is an auto battler, where you manage a hero's inventory to escape a haunted boardgame. Each run is different from the last, with unique upgrades, combos and synergies. Async PvP lets you battle the ghosts of other players at your own pace in this unique new inventory battler.\r\n\r\n\nhe builds left behind by other real players in asynchronous PvP battles. It’s a game of competitive inventory management, where every tile you place in your grid has a strategic cause-and-effect loop that impacts the tiles around it. It is a game that rewards foresight, creativi | the helpful playback tools to better understand how your choices played out so you can improve your grid. If you keep playing your tiles right, you might just keep your soul intact - and make it to the top of the highest competitive chambers. Once you have a winning build o | d tile customization Code & build sharing Still not sure? Some of our players have described it as “Backpack Battles crossed with Hearthstone Battlegrounds,” and said that it's like “if The Bazaar met Backpack Battles” or “if Inventory Tetris and Hearthstone Battlegrounds had a\n\n 3941130 Queue Queen 2026_H1 \n A card-stacking roguelike where you quickly stack cards and trigger items to get points. Discover thousands of different items and make synergistic builds by customizing your deck and inventory to reach higher scores and break the game.\node after a victory to see how far your build can go. New Game +, up to NG+5. Each NG+ expands your inventory and stack size by 1 and also increases the required points for each round.\n\n 4047130 Cardwheel 2026_H1 \n Endless runs, endless combos — every spin is a new roguelite adventure.\ne rules (e.g., more chips for red balls, multipliers for pairs), while card packs change how balls, slots, or colors behave. 👹 Boss Challenges: Every three rounds, face a unique twist. Some colors stop scoring, modifiers shut down, or jokers get disabled. Adapt fast; there a\n\n 4180100 Emily's Desire 2026_H1 \n Help Emily fulfill her desire of conquering dungeon mazes. Embark on RPG Adventures in this Roguelite Auto Battler that rewards exploration and creativity. Gain collectible Victory Cards of Emily whenever you win!\n\n\n 4206270 Wireworks 2026_H1 \n Design your defenses, build your module board, and manage your wires. Combine signals and create synergies to empower your weapons and survive in this base-defense roguelike auto-battler.\n\n\n 3380700 Papercraft Tactics 2026_JulAug \n Arrange a living backpack of swords, potions, and trinkets to auto‑battle foes in a paper‑theater roguelike. Link items to trigger chain reactions, play tactical cards, and outsmart bosses run after run.\nPapercraft Tactics is a run‑based, grid‑inventory roguelike with semi‑auto, turn‑based combat. You don’t just equip items—you wire them. Put swords next to potions, feed power from wands into shields, and rotate gear so |  inputs meet outputs. When a turn starts, your backpack comes alive: links fire, effects cascade, and enemies learn to fear tidy layouts. Between rounds, draft new items, shop, and re‑arrange for ever‑bigger chains. During fi |  Draft new items with distinct inputs/outputs and tags (elemental, heavy, fragile, scaling, mimic). Arrange items on a grid ; rotate to align green inputs and yellow outputs ; place for adjacency bonuses. Battle (auto‑resolve, turn‑based): At turn start, linked items trigger i | ue). Synergies are predictable and readable—builds reward planning, not twitch. Key Features Living Backpack: Your grid inventory is the build. Rotate, link, and position to forge synergies. Chain‑Reaction Combat: On OnPlayerTurnStart , items trigger and fill the action queue—c\n\n 4092940 Lucky Shot 2026_JulAug \n A shooting dopamine machine! Lucky Shot is a roguelike mag-builder FPS where you master different weapons, build wild combos, and break the game to reach massive high scores.\nbuilder FPS where precise shooting meets wild buildcrafting. Master different weapons, stack Holos, arrange special bullets, and combine charms, energy drinks, and target upgrades into absurd, run-breaking combos. Land clean shots, grow huge multipliers, and push the score unt\n\n 4474960 Ratatouille Builder 2026_JulAug \n In this cooking roguelike, inspired by many deckbuilder games, set up your garden to grow fresh vegetables and serve delicious dishes every day in your restaurant! Discover new recipes, exploit synergies, and learn your customers' favorite dishes!\n\n\n 4537600 Dice Gun Commando 2026_JulAug \n Guns roll dice. Dice trigger mods. Mods cause MAYHEM! Destroy waves of alien horrors in this bullet heaven roguelite by building an arsenal of dice-driven, synergistic weaponry.\n\n"}
{"chunk_id":"70a3e3","wall_time_seconds":0.000002656,"exit_code":0,"original_token_count":389,"output":"{\n  \"growth\": {\n    \"delta\": 874.0,\n    \"catalog_volume_composition\": 458.03311281635337,\n    \"within_signature_build_share\": 415.9668871836466\n  },\n  \"unique_game_count\": 184664,\n  \"pilot_labels_sha256\": \"de15007f4ce3c6f2ac1bb6fad97ab3d67bbb7d6c1a1bba9e3b494635496f0042\",\n  \"network_paused\": true\n}year,family,fractional_games\n2025,deck,92.16666666666666\n2025,action,258.0\n2025,auto,59.83333333333333\n2025,tower,47.0\n2025,none_of_four,386.0\n2026,deck,179.0\n2026,action,537.5\n2026,auto,105.16666666666666\n2026,tower,105.33333333333333\n2026,none_of_four,790.0\nyear,object,n,share,fractional_games\n2025,card_deck,209,0.2479240806642942,178.66666666666669\n2025,party_composition,120,0.1423487544483986,84.5\n2025,equipment_skills,448,0.531435349940688,392.16666666666663\n2025,machine_assembly,23,0.027283511269276393,12.666666666666666\n2025,unassigned,175,0.20759193357058126,175.0\n2026,card_deck,454,0.2644146767617938,390.66666666666663\n2026,party_composition,231,0.13453698311007573,158.0\n2026,equipment_skills,892,0.5195107746068724,774.3333333333333\n2026,machine_assembly,41,0.023878858474082703,20.0\n2026,unassigned,374,0.21782178217821782,374.0\nperiod,group,n,positive80,expected,ratio\n2025,deck,62,41,38.020404403025616,1.0783683299469922\n2025,other_builds,234,140,142.97959559697438,0.9791606936323057\n2026_H1,deck,33,22,19.36269841269841,1.1362052711398944\n2026_H1,other_builds,138,79,81.6373015873016,0.9676948951518037\n2026_JulAug,deck,11,10,6.1215320910973094,1.633577975445598\n2026_JulAug,other_builds,64,35,38.87846790890269,0.900241235894623\n"}

**Auditing deck premise**

Script completed
Wall time 0.4 seconds
Output:

{"chunk_id":"0eb0cf","wall_time_seconds":0.196786869,"exit_code":0,"original_token_count":6015,"output":"\n 2468550 Vivid World \n Vivid World is a completely new party-building roguelike auto-battler, where you assemble a team of characters that have transformed into jewels to engage in combat! Transverse a new world with quirky characters who've all been redesigned from Vivid Knight! Create your own unique and powerful party to clear every dungeon in your way! More in-depth party building than before! Do you know the proverb: \"Three heads are better than one\"? Gather several of the same jewels, enhance them, and see your buffs and synergy explode with power! And now, introducing \"Orbs.\" You can equip Orbs to each character in your party, which means your ability to create and change your party has now expanded infinitely! Strategic battles. In this game, the battles advance automatically while the player utilizes \"Gems\" to support their party during battle phases. With our new and improved battle system, the player can combine several Gems together or unleash the power of individual Gems! Enjoy strategic battles with greater freedom than ever before! Resource management that will test your thinking. The characters who battle for you are, after all, only jewels. That means you can \"buy and sell\" them. You can buy different types of Jewels or strengthen a single Jewel by merging several of the same Jewel with it. The choice is yours. Feel free to play around with your resources however you see fit. But be careful. If you're too greedy, you might find yourself being sucked into a dungeon here in the Underworld that you can't get out of. A world shrouded in mystery. The Underworld's deep, dark secrets. Those secrets will be revealed as you conquer dungeons and interact with the denizens of the Underworld. Just who are these characters who have transformed into Jewels? By strengthening your bonds with them and unleashing their hidden powers, you might find that out as well.\n\n 2686020 Faith in Despair \n You are Faith. The last of your kind. The epilogue to an age-old tale of demise. You gamble for your life against Death itself. Day after day after day. Yet you persevere. You dare not abandon hope. You struggle against your inevitable fall into despair. To live – and play – another day. Faith in Despair is a grimdark roguelike deckbuilder that allows you to create your own cards using triggers, actions and modifiers to concoct devastating synergies. Every run holds new dangers and every action will inevitably push you closer towards despair . FIND YOUR PATH TO PREVAIL Forge your cards: Collect triggers, actions and modifiers and define the abilities of your creatures and spells. Shape your deck: Formulate your strategy and devise devastating synergies to level the legions of Death. Give in to despair: Spells you cast and creatures you raise push you deeper into despair and put curses on you. Explore your own path: Follow Faith's journal, reveal her story and unlock new starter decks to challenge Death. FEATURED ARTISTS & MUSICIANS Cover art by Magic: The Gathering artist Marcela Bolívar In-game card illustrations by Vergvoktre , Marvin Blattert , Vil , Gwenypatu , Lyfeu Music by Muddasheep feat. Anete Liepina (opera singing) and Lou Dunand-Vincent (cello) Sound Design by Muddasheep and KAON\n\n 3215240 Probability Dungeon \n A turn-based roguelike where skills have a probability of getting executed. Your characters have an ordered list of skills. Every turn, you will need to decide the skill order to adapt to the enemy intents. Through the game, you will need to: Add, replace or improve skills. Equip accessories to characters. Adapt and establish different strategies. For this, you will have to look for shops, try the different events that you will find throughout the map or gain fight rewards! 4 Characters to combine, with different game styles and skill pools +150 Skills +80 Accessories\n\n 3410180 Overlooting \n An unlikely group has been formed with the mission of defeating a new dark force corrupting the world. Maximilian, a beloved hero coming from a noble family, Liz, a legendary thief wanted in multiple countries and The Shadow, an enigmatic being that everyone thought was the origin of the corruption. Looting equipment will be your main task. Each piece is part of a set and equipping enough pieces of a set will unlock unique bonuses. You will be able to acquire special skills from a tree that changes every game. Skills can offer you general utility, an increase in power or make one specific build go beyond its limits. At the end of each region a boss will try to stop you from completing your quest. This powerful enemies have special abilities and will test your strength, so make the right choices at every step of your journey to make sure no one will stand in your way!\n\n 3573070 Toy Smash Kaboom! \n Join Our Official Discord ! (You can access Discord link from the panel on the right.) In the dusty corner of memory, the toy room stirs once more. Awaken forgotten childhood toys and fuse them within the small table to craft your own kaboom legion — from ground collisions to aerial strikes and magical blasts. Through each tabletop battle, fragments of your long-lost childhood return… Arrange, Combine, and Evolve Master the art of toy arrangement on your tabletop! Every toy slot is a strategic decision. Plan your layout, combine rare units, and arrange them wisely to unlock your build’s full potential! Build Your Playstyle Craft your own tactical system from over 400 strategic toys! 100+ summons to clash in tabletop battles 50+ traps and class abilities to outsmart your opponents and shift momentum 250+ magic items and weapons to power up your strategy and unleash devastating combos Collisions & Knockback Summon toys into head-to-head battles where it’s not just about stats—brave charges and knockback strikes can send enemies flying, turning the tide in unpredictable ways. Set Traps Set traps across the table to break enemy momentum and open up opportunities to strike back. Engage in asynchronous PvP where your legion clashes with global builds in tactical showdowns. Analyze, adapt, and optimize your strategy in mirror matches and rise through the ranks. Pick Your Toy Captain Each hero brings a unique toy arsenal and exclusive class abilities. Doll Chef: A kitchen defense expert who controls space and fortifies positions, perfect for defensive-minded players. Dr. Hamster: Commander of Hamster Legion that leads a swarm of tiny units to overwhelm enemies with relentless numbers. Puppet Musician: A rhythm master who speeds up allies and disrupts foes, setting the tempo of every battle. Naughty Andy: Summoner of nightmares — ghosts, skeletons, reapers, and dark curses! Frost Hunter: Rule the frozen land, command wild beasts, and unleash the power of water magic! Turn Your Friends into Toys Transform your Steam friends into fully armed \"Game Buddy\", customize them with all kinds of weapons and attributes, then rally them into your toy army to charge into battle for friendship! Unveil Toy Memories Unlock achievements, hero skins, and the toys’ hidden memories through a pop-up storybook. Within its dusty pages, the long-lost tales of Fred and his childhood friends quietly come back to life.\n\n 2652600 SealChain: Call of Blood \n SealChain is a roguelike game where items can link together to form chain reactions. Items can also be arranged into specific formations to create seals, granting special effects. The world is collapsing. Chaotic evil shrouds the land like thick purple mist. In this catastrophe, countless innocent beings have become cursed monsters. There is no savior here. Ancient incantations are awakening, bestowing power upon the 'Spirit Mediums'. Use the power of seals to cleanse the evils of the apocalypse. Game Features Items can link together to form chain reactions Items can enchant each other to enhance attributes Items can resonate with each other for additional effects Items can be arranged into specific formations to create seals, granting special effects Multiple playable characters, each with unique builds, skills, armor, and weapons Event system: obtain various items through randomly triggered events Rich variety of items, rewards, builds, monsters, and events Receive different rewards from a variety of NPCs Items can be stacked infinitely Unlimited combat power, exponential thrill Talent tree system Endless mode Easy modding: create mods without programming — edit simple .txt files to customize items, monsters, characters, and events Forming Seals by Linking Items Thrilling Combat\n\n 3284290 Moonsigil Atlas \n Experience a roguelite deckbuilder without energy or action points: there's no limit to how many cards you can play, as long as they fit! Every card is a sigil you place on a grid, each with its unique shape, keywords, and effects - turning each combat into a tactical optimization challenge. Create powerful synergies through rune positioning, adjacency, overlap, and persistent effects. Place the right sigil at the right time to trigger devastating combos. Beyond spatial strategy, Moonsigil Atlas is about creative deckbuilding: draw infinite cards to deal infinite damage or make ten copies of your strongest card. A fully modular, player-driven upgrade system. Every card can be modified as you see fit: reshape them, inscribe runes, add keywords, and more. Spread power across your deck or invest everything into a handful of unstoppable cards. With tens of thousands of possible combinations, even the simplest cards can become build-defining centerpieces. With 250+ cards, alternate starting decks, unlockables, forbidden cards, 3 characters with radically different playstyles and a fully randomized map, every run offers new, game-breaking possibilities. Battle dozens of astral entities and three Titans that reshape the board itself. Hostile zones, shifting grids, and unique win conditions turn every encounter into a test of adaptation and mastery. A powerful deck is only half the battle…\n\n 3355940 Card Coder \n Card Coder merges tactical card battles, inventory management and roguelike deck-building with a novel twist: Create your own custom cards during each run by constructing cybernetically enhanced warriors from all over the galaxy. Defeat challenging enemies to loot, shop and collect many different ability components. Slot them into your cards to define their behavior according to your playstyle: Discover, research and unlock a large variety of components to construct over 10 billion different cards . Build up your engine and cause powerful chain-reactions across the board. Use everything to your advantage and crush your enemies in challenging rogue-like runs, high-stakes gauntlets and other game-modes to be unlocked. But keep your Commander alive at all costs! Features Card Coder merges tactical card battles, inventory management and roguelike deck-building with a novel twist: Create your own custom cards during each run. B eyond deckbuilding : Create your own cards as you play by combining triggers, conditions, selectors and effects! Over 10 billion possibilities. Build your cards to fit your playstyle -- not the other way round! Tactical combat with positioning : Manipulate enemy cards' targeting aggro by placing your cards strategically on the board. Take advantage of superior positioning. Meta Progression Unlock new ability components and base cards after each run to grow your card-creation possibilities from several thousands to several billions. Discover new combat locations with unique gameplay modifiers that can turn the tides of battle. Beat new game modes like constructed gauntlet or random draft gauntlet. Cosmetic card customization : The card-building aspect of Card Coder does not stop at scripting the mechanics. To make cards memorable and easy to distinguish in battle you can also customize the visuals of each card: Change the title , adjust the color tint , select awesome artwork from different pixel artists or even replace the card's artwork with any custom image you choose. Upgrade the rarity of ability components by merging duplicates to gain stats bonuses and powerful status auras. Discover unique recipes : Certain combinations of ability components provide an additional bonus. Addicting gameloop : Fight an opponent, gain rewards, go shopping and create new and better cards to improve your deck. Face the next opponent until you beat the Big Bad. Save Anytime : Need to leave in the middle of a run? No problem. Card Coder saves the game exactly where you are, so you can pick up where you left off at any time. No MTX / IAP / DLC : Card Coder is a buy once, play forever game. There will never be any in-game microtransactions / purchases, pay-to-win mechanics, premium cosmetics, online or 3rd party account requirements nor any paid DLC. Everything is already included in the game or will be added via free future updates. Get Card Coder now and construct awesome cards!\n\n 3802470 Turnbound \n Turnbound is a strategic autobattler focused on experimentation, theorycrafting, and mastery of the odds. You are trying to outsmart your way to freedom from a haunted board game where each session pits you against the trapped souls of other players. You battle against the builds left behind by other real players in asynchronous PvP battles. It’s a game of competitive inventory management, where every tile you place in your grid has a strategic cause-and-effect loop that impacts the tiles around it. It is a game that rewards foresight, creativity and luck. Will your design outlive your ghostly opponents… or join them? During the Shop Phase you pick, place or upgrade weapons, items, armour, consumables and accessories in the most optimal way. Trinkets let you unlock combos and winning battles let you upgrade your Hero through a branching set of abilities. Each round will have you making strategic decisions to execute synergies that give you an edge against your opponents! Take all the time you want to plan your build and manage your board. Whenever you start a battle you will go against a build from the same chamber, with the same number of shop rounds as you. Whenever you exit a battle, your build is added to the matchmaking database and someone else can face it. Whether you win the whole run or not has no influence on which builds are added to the matchmaking database. The builds you face as opponents may have lost the run or still be in the middle of it. You can face the build of someone who has played seconds before. Every detail counts and analyzing them is the key to understanding how to make the most of the tiles you have been given. The board won't always give you what you want, you will need to adapt your strategies, try new things and rely on a little bit of luck. Combat plays out from left to right, top to down - as each asynchronous multiplayer autobattle completes, use the helpful playback tools to better understand how your choices played out so you can improve your grid. If you keep playing your tiles right, you might just keep your soul intact - and make it to the top of the highest competitive chambers. Once you have a winning build on your hands, you can compete directly instead of battling spirits asynchronously. Duels let you challenge a specific build by entering a simple code. Share your own builds with a unique code, test yourself against a friend’s strategy or go up against selected Feature Builds from our community! It’s fast, chaotic, strategic and the perfect way to prove once and for all whose tile brain actually reigns supreme. Key features 4 Heroes Dozens of trinkets and abilities Hundreds of items 6 Chambers of progress and Final Chamber Leaderboard Board, profile, rug and tile customization Code & build sharing Still not sure? Some of our players have described it as “Backpack Battles crossed with Hearthstone Battlegrounds,” and said that it's like “if The Bazaar met Backpack Battles” or “if Inventory Tetris and Hearthstone Battlegrounds had a cursed baby”. So if any of those games interest you, you may have fun with this one too! NO GEN AI No Generative AI, was used to create any of the in-game art or assets. Not even the battles are AI, all of them come from saved/winning builds created by other players. Even the tutorial where you play against “the cat\", or occasional \"bot builds\" were crafted by our game designers.\n\n 3941130 Queue Queen \n Queue Queen is a game like Balatro and Nubby's Number Factory where you stack cards, trigger items and build game-breaking combos to clear the rounds in a limited time. More than 2000 items All items are procedurally generated, so there are over 2000 of them with unique effects! You have white items which trigger when cards are evaluated after you've filled your stack. Yellow items trigger at any time when something happens, even in the shop. Pink items are like yellow items, but they have more powerful effects and can't be force triggered. Blue items grant passive bonuses. Tailor your build using the Cardmage Parlor and the Backroom Printer The Cardmage Parlor lets you buy/clone/destroy cards or give them buffs which trigger at different times. In the Backroom Printer you can merge, duplicate, edit or upgrade your items to build even stronger combos. Challenging Bosses They can discard your stack, swap your items around, skip cards, forcibly sell your items or do all kinds of mischievous things, so plan for them accordingly! Replayability Endless mode after a victory to see how far your build can go. New Game +, up to NG+5. Each NG+ expands your inventory and stack size by 1 and also increases the required points for each round.\n\n 4206270 Wireworks \n Wireworks is a base-defense roguelike auto-battler where your defenses are built and controlled using a module board. Manage your wires, fine tune your modules, combine signals and create synergies to survive incoming enemies and bosses. 200+ Items and Skills to mix, match, and discover new builds. 3 Unique Areas with varied enemies and boss fights. Unlockable Characters with special skills and starting kits. Ascending Difficulties for further challenge. Endless Mode to push your builds to their limits.\n\n 3380700 Papercraft Tactics \n Papercraft Tactics is a run‑based, grid‑inventory roguelike with semi‑auto, turn‑based combat. You don’t just equip items—you wire them. Put swords next to potions, feed power from wands into shields, and rotate gear so inputs meet outputs. When a turn starts, your backpack comes alive: links fire, effects cascade, and enemies learn to fear tidy layouts. Between rounds, draft new items, shop, and re‑arrange for ever‑bigger chains. During fights, play a handful of clutch cards to tweak the queue—double a hit, purge a debuff, or force a trigger now. The joy is in the planning: make a build that plays itself, then nudge it at the perfect moment. All of this unfolds in a warm paper‑theater aesthetic—cut‑out characters, stage‑flat scenery, and a subtle handmade vibe. The Core Loop Loot & Draft new items with distinct inputs/outputs and tags (elemental, heavy, fragile, scaling, mimic). Arrange items on a grid ; rotate to align green inputs and yellow outputs ; place for adjacency bonuses. Battle (auto‑resolve, turn‑based): At turn start, linked items trigger in order, adding Actions to the queue (Attack, MagicCast, Recover, Block, InflictEffect, Special). Play Cards to manipulate the queue (duplicate, remove, reset exhausted, trigger now) or inflict effects. Progress through events, shops, and escalating encounters; refine the build; beat the boss; try a new hero. Heroes, Stats & Resistances Pick from distinct heroes with unique passives and stat profiles. Core stats include HP, SP, MP, STR, DEF, MND, SPR, Evade . Resistances cover physical types (Slash, Thrust, Strike, Projectile, Spell) and elements (Fire, Ice, Wind, Thunder, True). Synergies are predictable and readable—builds reward planning, not twitch. Key Features Living Backpack: Your grid inventory is the build. Rotate, link, and position to forge synergies. Chain‑Reaction Combat: On OnPlayerTurnStart , items trigger and fill the action queue—combos you designed go off automatically. Tactical Card Play: Limited‑use cards let you push the queue— DuplicateNextAction , ClearAttacks , ResetExhaustedCards , and more. Readable Triggers: Clear events like OnSlashHit , OnReceiveFireDamage , OnEvade power effects without hidden rules. Scaling & Mimic Items: Effects that grow from HP/MP/SP or copy linked behavior to extend chains. Run‑Based Variety: Shops, events, elites, and bosses with different resist profiles make placement matter. Paper‑Theater Style: Hand‑drawn lines, cut‑out layers, and cozy stage lighting. No Dead Turns: Even defense builds feel active via blocks, conversions, and smart queue tweaks. Learnable Depth: Simple pieces, deep compositions—like building a Rube Goldberg machine that fights back.\n\n 4092940 Lucky Shot \n A flick-shot dopamine machine Lucky Shot is a roguelike mag-builder FPS where precise shooting meets wild buildcrafting. Master different weapons, stack Holos, arrange special bullets, and combine charms, energy drinks, and target upgrades into absurd, run-breaking combos. Land clean shots, grow huge multipliers, and push the score until the run breaks wide open. Build wild combos. Break the shooting range. Shape each run around the tools you find. Reach the required score, defeat the bosses, and keep the run alive as every choice changes what your next shot can do. ➤ 150+ Holos that stack, synergize, and bend the rules in different ways ➤ 12 different weapons, each with a unique playstyle that changes how you approach every run ➤ Multiple special bullets for your mag that define your build and shape each run ➤ 15+ bosses that test both your aim and your build ➤ 16+ charms that stay with you throughout the run and trigger special effects ➤ 15+ energy drinks with side effects that can save a run or wreck it completely ➤ Multiple special target types that let you influence what spawns next. The range is yours New stages, new eras, new tracks Each stage pays tribute to a different era, with its own visual identity and soundtrack. Strategy matters as much as aim You do not need to be an FPS expert. Lucky Shot is designed for shooter and deckbuilder players alike: strong aim can push a build further, while smart synergies and magazine planning can turn every shot into something spectacular.\n\n 4537600 Dice Gun Commando \n A Bullet Heaven where you control the dice your guns roll The Chaos Lord and its minions are here, and you're the last line of defense. Build an arsenal of dice, mods, and guns to show those aliens who's boss! Features: Carry up to six auto-firing weapons. Shotguns, machine guns, lightning guns, oh my! Dozens of different dice and hundreds of different gun mods. Massive build variety! Wide variety of synergies from combining dice and mods with different guns. Play it your way! Level up to earn rewards and make your arsenal stronger. Power overwhelming! Fast runs (20-30 minutes) and progressive unlocks. Roguelite stuff! Fight alien mobs, elites, and bosses that wield their own dice powers against you. Kill chaos! Break the laws of physics game. Power over-- wait, already did that one. Break stuff! What kinds of synergies can you make? A machine gun that spews radioactive clouds when it rolls a set of 1, 2, and 3. A shotgun that knocks back mobs each time it rolls a 2. A chickenshot that turns mobs into chickens while also slowing them down when it rolls a 3. A lightning gun that heals you when it rolls a pair of 1s. A grenade launcher that spreads madness when it rolls three 2s, causing mobs to attack each other. A flamethrower that recharges your shields when it rolls a pair of 4s. And much, much more! Craft as absurdly overpowered a build as you can. It's what you'll need to annihilate the Chaos Lord and stop its insanity. Be the Dice Gun Commando!\n"}

Script completed
Wall time 1.2 seconds
Output:

{"chunk_id":"d78dd3","wall_time_seconds":1.001633291,"session_id":32362,"original_token_count":0,"output":""}

**Verifying growth fractions**

Script completed
Wall time 0.5 seconds
Output:

{"chunk_id":"8b9650","wall_time_seconds":0.000002585,"exit_code":0,"original_token_count":7674,"output":"Warning: truncated output (original token count: 7674)\nTotal output lines: 80\n\n2025 text_carddeck {'n': 135, 'under10': 33, 'ge10': 102, 'ge50': 64, 'ge100': 48, 'ge200': 28, 'ge1000': 10, 'liked100': 34, 'rate100': 0.35555555555555557, 'liked100_rate': 0.2518518518518518, 'satisfied_among100': 0.7083333333333334, 'median_reviews': 37.0, 'median_positive_among100': 85.5, 'median_price': 9.99, 'developers': 135, 'publishers': 129, 'reviews_sum': 41813, 'top1_review_share': 0.20603640016262884}\n2025 no_text_carddeck {'n': 34, 'under10': 5, 'ge10': 29, 'ge50': 17, 'ge100': 14, 'ge200': 9, 'ge1000': 4, 'liked100': 7, 'rate100': 0.4117647058823529, 'liked100_rate': 0.20588235294117646, 'satisfied_among100': 0.5, 'median_reviews': 45.5, 'median_positive_among100': 79.0, 'median_price': 9.99, 'developers': 34, 'publishers': 32, 'reviews_sum': 35556, 'top1_review_share': 0.5970581617729779}\n2026_H1 text_carddeck {'n': 111, 'under10': 44, 'ge10': 67, 'ge50': 36, 'ge100': 28, 'ge200': 17, 'ge1000': 7, 'liked100': 20, 'rate100': 0.25225225225225223, 'liked100_rate': 0.18018018018018017, 'satisfied_among100': 0.7142857142857143, 'median_reviews': 18.0, 'median_positive_among100': 85.0, 'median_price': 9.99, 'developers': 111, 'publishers': 111, 'reviews_sum': 217297, 'top1_review_share': 0.8376553749016323}\n2026_H1 no_text_carddeck {'n': 21, 'under10': 12, 'ge10': 9, 'ge50': 6, 'ge100': 5, 'ge200': 4, 'ge1000': 1, 'liked100': 2, 'rate100': 0.23809523809523808, 'liked100_rate': 0.09523809523809523, 'satisfied_among100': 0.4, 'median_reviews': 5.0, 'median_positive_among100': 78.0, 'median_price': 7.99, 'developers': 21, 'publishers': 21, 'reviews_sum': 3512, 'top1_review_share': 0.44903189066059224}\n2026_JulAug text_carddeck {'n': 49, 'under10': 24, 'ge10': 25, 'ge50': 14, 'ge100': 10, 'ge200': 4, 'ge1000': 0, 'liked100': 9, 'rate100': 0.20408163265306123, 'liked100_rate': 0.1836734693877551, 'satisfied_among100': 0.9, 'median_reviews': 10.0, 'median_positive_among100': 88.0, 'median_price': 8.99, 'developers': 49, 'publishers': 49, 'reviews_sum': 3854, 'top1_review_share': 0.2472755578619616}\n2026_JulAug no_text_carddeck {'n': 18, 'under10': 8, 'ge10': 10, 'ge50': 5, 'ge100': 1, 'ge200': 0, 'ge1000': 0, 'liked100': 1, 'rate100': 0.05555555555555555, 'liked100_rate': 0.05555555555555555, 'satisfied_among100': 1.0, 'median_reviews': 13.0, 'median_positive_among100': 87.0, 'median_price': 7.99, 'developers': 18, 'publishers': 18, 'reviews_sum': 588, 'top1_review_share': 0.27721088435374147}\n  appid                                 name      period  reviews  positive_pct                                                                                                                                                                                                                                                                                            short_description\n2784470                              9 Kings        2025    21229            92                                                                                                                 A fast-paced roguelike kingdom builder. Grow your empire and fight massive battles against powerful rival kings. Break the game with thousands of insane builds to become the King of Kings.\n2289750                Super Fantasy Kingdom        2025     5590            86                                    Rebuild your kingdom and explore a vibrant fantasy world in this roguelite city builder. Defeat hordes of monsters with a unique team of defenders as you harvest, build, mine, cook, brew, and grow. But if your defenses should fail, be prepared to watch it all burn!\n2824490                         He is Coming        2025     3812            84                                                                            Deep in the forgotten corners of the world, a dark force stirs – the Demon King rises once more. Sharpen your wits, seek out powerful artifacts, and stand firm in the face of the apocalypse in this roguelite RPG auto battler.\n3509230                           Gambonanza     2026_H1     1577            74               Gambonanza is a tactical roguelike that sets chess pieces on a tiny board. Forget traditional chess and embrace the chaos of breaking the balance! Spend your money on rule-bending upgrades, board tiles and pieces. Experiment with illegal combinations and face off against iconic bosses!\n3410180                          Overlooting        2025     1402            81                         Adapt your strategy to a skill tree that changes every run, combine pieces of equipment to form powerful synergies and create a game-changing build capable of defeating the bosses that will try to stop you. In this Inventory Management Roguelike two games never feel the same!\n4206270                            Wireworks     2026_H1      602            89                                                                                                                  Design your defenses, build your module board, and manage your wires. Combine signals and create synergies to empower your weapons and survive in this base-defense roguelike auto-battler.\n2468550                          Vivid World        2025      583            88                         A follow-up to the hit game Vivid Knight, Vivid World is a completely new party-building, roguelike strategy game!\\r\\nOur story takes place in the Underworld, a land ruled by demons.\\r\\nUtilize and master our protagonist, Lemuria's new Jewel Magic to conquer strange dungeons.\n3573070                    Toy Smash Kaboom!        2025      551            84                                                                                                                                                         Toy mayhem unleashed! Arrange, combine, and craft the wildest toy army inside your desk, then crash into chaotic cartoon brawls across the tabletop!\n2652600             SealChain: Call of Blood     2026_H1      533            56                                                                                                                          SealChain is a roguelike game where items can link together to form chain reactions. Items can also be arranged into specific formations to create seals, granting special effects.\n2536960 Shadowveil: Legend of The Five Rings        2025      468            74                               Train unique samurai, embark on missions, and play the cards dealt each run on an epic roguelite adventure in the Legend of the Five Rings universe. Each defeat opens new strategic skill tree options and more powerful cards to face the growing threat in the Shadowlands.\n3802470                            Turnbound     2026_H1      441            78       Turnbound is an auto battler, where you manage a hero's inventory to escape a haunted boardgame. Each run is different from the last, with unique upgrades, combos and synergies. Async PvP lets you battle the ghosts of other players at your own pace in this unique new inventory battler.\\r\\n\\r\\n\n2760790      Another Farm Roguelike: Rebirth        2025      434            69 Discover the ultimate farming roguelike experience! Build, craft, and grow your farm to pay off the increasing rent. No slow-paced walking or social distractions—just click! Reclaim the joy of strategic farming in diverse maps and let your cursor guide you to success. Play your way, rent won’t wait!\n2870430                      Godslayer Arena        2025      257            76                                            Godslayer Arena is a roguelike fusion of autobattlers and action-rpgs. Combat is performed completely automatically. Understanding complex interactions between the spells, gems, and items you draft will be key to success. Break the game. Become overpowered.\n3930510                  Don't Let It Starve     2026_H1      184            95                                                                                    A horror roguelike bento-builder. You're trapped in an underground kitchen, forced to feed a half-chef monster. Gamble with your life, assemble meals, and master your multipliers. Whatever you do, Don't Let it Starve.\n3656680                      Gold and Graves        2025      178            86    Gold and Graves is a roguelike extraction game where you bribe enemies to recruit them as allies, escape the dungeon, and bring back cards, relics, and mercenaries to your village to accumulate resources. Strengthen your deck and forces with the resources you gather and challenge deeper dungeons.\n3123120                    Kill the Brickman        2025      166            75                                                                                                                                                                                                                     A brick breaking, turn-based rogue-like bullet-builder - Blast bricks to break the bank.\n4252700                          Archon Soul 2026_JulAug      163            87                                                                Archon Soul is a 2D roguelike dice-builder where you create powerful dice using swappable faces, runes, and bodies with unique effects. ​Draft spells, craft powerful synergies, defeat challenging enemies, and conquer the Archon's palace.\n1830320                      Lords of Ravage        2025      138            66                                                                          Become the Final Boss! Play any way you want, combine followers from different factions, and win spectacular turn-based battles. Expand your Empire of Darkness and seek a powerful artifact that can change the fate of the world.\n3179860                       Pairs & Perils        2025      135            77                                                                      Pairs & Perils is a memory-game roguelike and dungeon crawler. Match cards to attack enemies, avoid traps, shop for relics and consumables. Sharpen your mind as you journey through the depths of the dungeon in search of The Dealer.\n2671010                        Turtle Riders        2025      114            76                                                                                       Ride giant turtles and aim your weapons at hordes of enemies chasing you while you gather as many resources as you can. Pick your upgrades carefully and create strong builds to survive the long road to the Outpost!\n3531630              Tomb of the Bloodletter     2026_H1       99            87                                   Tomb of the Bloodletter is a strategic word roguelike where your keyboard becomes your arsenal. Use your wit to unleash deadly words through tactical combinations of your letters to defeat the denizens of this dingy tomb and claim the Bloodletter for yourself...\\r\\n\n4271680                             outpost4 2026_JulAug       85            71                                                     Build your mobile outpost and combine over 20 weapons with 100+ items to create unique builds. Grow stronger against endless enemy hordes, conquer new planets, and experience the thrill of evolving from desperate survival to overwhelming firepower.\n3929350                               ploink 2026_JulAug       78            94                                                                             A chaotic plinko inspired roguelike with a focus on wacky items and creative setups. Design a board of random crap and fill the screen with balls and numbers. There are no bad items you just gotta figure out how to use them.\n3820810                       Dice Dice Baby        2025       78            88                The House always wins! Except here...\\r\\nDice Dice Baby is a roguelike dice game about cheating and rolling the craziest numbers to beat the House.\\r\\nTwist a classic game thanks to illegal special dice and create epic combos with lucky charms, all in the world most permissive casino.\n3084370                       Ricochet Abyss 2026_JulAug       75            76                        Dive into Ricochet Abyss, a unique blend of tactical backpack management, hero-building, and pinball-style combat. Organize your inventory, equip powerful gear, and send your heroes ricocheting through perilous dungeons—every shot and item choice could be your key to survival!\n3211750                           Neongarten        2025       66            87                                                                                                                                                                                                                            A minimalist cyberpunk city builder inspired by Islanders and Luck Be a Landlord.\n2600970                       Flawed Tactics 2026_JulAug       59            89     The autobattler where the market IS the balance patch. Build your team, cast your spells, create powerful relics, optimize synergies - but watch the ticker. Every minion you buy raises its price for every player. No permanent metas. No overpowered favorites. Just tactics, and flawed ones at that\n3128280                          Cats & Dice        2025       51            76                            A roguelike dice-building game where the goal is to overcome an ever-increasing target score each round. Assemble a synergistic set of Cats and dice, go for crazy combos and try to attain as high a score as you possibly can. Master the synergy between cats and dice to win!\n3438640                           Dicey Bugs        2025       40           100                                                                                                                             A roguelite dice‑building creature collector where you control both sides of the fight by choosing one action at a time. Capture bugs, unlock new dice, and customize every run.\n4597030                            Hillforts 2026_JulAug       37            86                                     …674 tokens truncated…                                   Unlock 12 ancient artifacts, each bestowing special powers and a unique starting deck. Battle relentless automatons on your quest to the heart of the enchanted city as you uncover stories of an ancient book.\n4537600                    Dice Gun Commando 2026_JulAug       20           100                                                                                                                            Guns roll dice. Dice trigger mods. Mods cause MAYHEM! Destroy waves of alien horrors in this bullet heaven roguelite by building an arsenal of dice-driven, synergistic weaponry.\n3299130                       Mayhem Maidens        2025       19            89                                                                                                                                                  Ready for some chaos? Command 5 characters in this Roguelite Tower Defense! Build your dream team, crush enemies, and have a blast with adorable pixel art!\n1635710                    Bossfight Tactics        2025       18           100                                                                                Become the BOSSFIGHT when you win! Assemble quirky item synergies and outsmart enemy intents in a series of turn-based combat. Bossfight Tactics distills a deep, tactical experience into a cute and approachable roguelike!\n2968290                   Cards and Dungeons     2026_H1       17            88                                                                                                                                    A roguelike dungeon crawler where every card you play brings power… and punishment. Descend through shifting floors, face brutal bosses, and grow stronger run after run.\n4414210                        Dice Carnival 2026_JulAug       16           100                                                                                                   Dice Carnival is a roguelike score-chaser where you roll, bank, and chain dice into explosive combos. Collect powerful Tokens, pick your Tour, crush chapter Bosses, and push your score beyond the limit.\n3941030                           SOLARFLARE     2026_H1       16           100                                                                                                                            SOLARFLARE is a byte-sized roguelike auto-battler in a sci-fi dystopia. Fight (and die) to unlock Upgrades, experiment to find broken builds, and obliterate the megacorp's bots.\n2730820                              Aliosso     2026_H1       13            84                                                                                                                                       Combine tiles and run-changing cards to unleash power-packed combos and crush your enemies. Fight your way through the scorching desert of Samsara and stop the Djinn!\n3964490                                Ruina        2025       12           100                                                             Build game-breaking combos, stack absurd multipliers, and forge powerful synergies in this roguelite match-3 puzzle game. Customize your gem bag, collect scrolls, relics, and tools, and discover new ways to push your score higher every run.\n3822720                         Deepest Well        2025       11            72                                                                                                                      Deepest Well is a roguelike where you fight off waves of enemies in a descending well with destructible terrain. Get upgrades after each wave and make insane builds to break the game!\n3319720                               Kernel        2025       10            70                                                                              In the universe billions of years later, life no longer exists. The adventurer Kamus travels to the unknown star realm as a hope to break the deadlock, exploring the ultimate destiny of the soul through battle and conquest.\n4245340                        Dominocalypse 2026_JulAug       10           100                                                                              The puzzle roguelike that turns dominoes into a mind-bending spectacle! Move, manipulate, and match using absurd tile modifiers, game-changing artifacts, and unique starting loadouts for endless chaos and replayability.\\r\\n\n3054470                        Rogue Zodiacs        2025        8            87    A content-rich, turn-based roguelike with truly endless replay value achieved by having countless varieties because all contents are unique. 250+ unique abilities, 200+ unique relics, 16+ locations, 40+ hazards, 90+ unique enemies, and the 12 Zodiacs guarantee each of your runs will be different.\n4529660          Greedy Spin: Reel Roguelike 2026_JulAug        8           100                                                                             Greedy Spin is a roguelite slot game built around a rotating prize wheel. Freely modify the machine’s symbols and multipliers, combine 150+ items to uncover wild chain effects, and rake in absurd amounts of chips — and cash!\n3324350                 Three Kingdoms Poker        2025        7            71                                           Lord! Gather your hand, defeat all kinds of forces (and win their people over), and become the ruler of the Three Kingdoms! That’s your mission!\\r\\nIn this game, you’ll use the cards in your deck to form specific card combinations, dealing damage to enemies!\n3505230                            Soulblaze     2026_H1        7           100                                                                                                                        Catch Animons and roll some dice in this cute creature-collecting roguelite! Build your team, train them well, and fight your way through several regions. Can you defeat all bosses?\n3964420                     Hell Dice Gambit     2026_H1        5            60                                                                                                                                'Hell Dice Gambit' is a strategic roguelike game using dice. Collect various relics, dice, and pills to grow stronger and struggle on. Can you escape from the abyss of hell?\n4018280                            Star Dice 2026_JulAug        5            80                                                                                                A turn-based roguelike strategy game about crafting powerful builds through dice customization and relic synergies. Optimize your 3×3 dice grid, trigger massive combo chains, and reach unbelievable scores.\n3705150                      Into The Depths     2026_H1        5           100                                                                                              A minimalistic mix of a roguelite city builder and a puzzle game. Assemble your deck, build an outpost and colonize the ever-changing caverns. Discover game-breaking synergies to push deeper into the Depths.\n3215240                  Probability Dungeon        2025        4            75                                             A turn-based roguelike where skills have a probability of execution. ¡You will have to keep improving your list of skills and decide in which order you want them to be executed! Set your strategy, create synergies and try to overcome the different threats.\n3380700                   Papercraft Tactics 2026_JulAug        4           100                                                                                                  Arrange a living backpack of swords, potions, and trinkets to auto‑battle foes in a paper‑theater roguelike. Link items to trigger chain reactions, play tactical cards, and outsmart bosses run after run.\n4127450                              Mixatro     2026_H1        3           100                                                                                    Mixatro is a game where you merge Balls and buy upgrades to allow you to score higher and reach higher goals. It blends the simple, relaxing puzzle fun with the exciting, strategic choices of building a powerful deck.\n4334020      SUITE:The Automated Card Engine     2026_H1        2           100                                                                                       SUITE: Automated Card Engine is a poker roguelike game. Run countless workshops, wield broken cards and bizarre Jokers, create dozens of unique poker hands, and earn chips on your way to becoming unbelievably rich.\n3327710                  Shadows of Wyrmwood     2026_H1        2            50     Shadows of Wyrmwood is a roguelike adventure RPG. Use spells, swords, and strategy to defeat every corrupted foe. Travel across procedurally generated maps that grow harder with every encounter. With each victory, obtain new armor, weapons, and spells. Do you have what it takes to save Wyrmwood?\n3523760                          イーシャの魔法書解読室        2025        2           100                                                                                                                                                   [roguelike]\\r\\nTogether with the wizard Isha, embark on a roguelike adventure to decipher the magic book and resolve the anomalies occurring in this land!\n3682060                              Vocator     2026_H1        2           100                                                                                                              Vocator is a strategic roguelike that blends Monopoly-inspired board mechanics with autobattler gameplay. Place heroes on the board, empower your servants, and adapt your strategy to survive.\n4397160               Apocalypse: The Hunger     2026_H1        2             0          We clapped the Minion Meta. No armies to manage. No 1/1 tokens to trade. Just you, and the miscreation in front of you. Apocalypse: The Hunger combines deep RPG progression, item crafting, and the tactical return of \"The Stack\" in a brutal, high-stakes card battler where you play for keeps.\n3139550                        Cappu Dungeon     2026_H1        1             0                                                                        Cappu Dungeon is a skill-building action roguelite\\r\\nwhere you equip skills to wipe out hordes of enemies.\\r\\nBuy, Combine, and Synergize Skills to create your own unique build. \\r\\nConquer the dungeon through strategic choices!\n4140190                          Dice Rogues 2026_JulAug        1           100                                                                          Dice Rogues is a tactical roguelite with infinite possibilities where you upgrade your dice and harness unpredictable rogue faces (X2, baguette, croissant, etc.) to bend fate, outplay the House, and survive a new roll each run.\n4415640                            Symbolika     2026_H1        1           100                                                       Prove your worth against Symbolikus. In this roguelike card game, your only goal is destruction! Optimize your synergies and tactical choices to face 289 fighters before dealing maximum damage to the Creator and dominating the global leaderboard.\n4417700                             PentaPip 2026_JulAug        1           100                                                                                               Roll, hold, score and break the rules. Build powerful synergies across five scorecards, outplay bosses that change the game, and take on seeded challenges solo, in Daily Runs, or in competitive multiplayer.\n4180100                       Emily's Desire     2026_H1        0             0                                                                                        Help Emily fulfill her desire of conquering dungeon mazes. Embark on RPG Adventures in this Roguelite Auto Battler that rewards exploration and creativity. Gain collectible Victory Cards of Emily whenever you win!\n3665300                         Rogue Ninjas        2025        0             0                                                                        In Rogue Ninjas, survival demands cunning and precision. Manage a squad of elite ninjas through twisting dungeons, strengthen your inventory with rare enhancers, and battle ruthless bosses in this dice-driven roguelike adventure.\n4465730                            Volt Deck 2026_JulAug        0             0                                                                                                                                                                                                                 Build poker hands, rewrite the rules, and chase high scores in a tactical solitaire campaign\n4540910                          Flower Dice 2026_JulAug        0             0                                                                                        Cultivate your garden & push your luck in Flower Dice. Focus on adjacency bonuses, like forming colour palettes, to maximise the point value of your garden. Each flower has unique abilities for endless builds.\\r\\n\n4554420                    Last Crown Gambit     2026_H1        0             0                                                                                                                                                                 A chess-based roguelike tactics game where you empower pieces with cards, build tag synergies, and shape each run through strategic choices.\n5061410                             Krukmore 2026_JulAug        0             0                                                                                                                                                        Krukmore is a brutal, transparent dice roguelike. No hidden math. No safety nets. No save scumming. One run. One mountain. One ending — yours or his.\n"}
{"chunk_id":"96f08b","wall_time_seconds":0.29459971,"exit_code":0,"original_token_count":2381,"output":"     period                         group  n  ge100  liked100  joint_ratio\n       2025      deck_party_tactics_tag20 58     24        16        1.759\n       2025   deck_economy_building_tag20 15     10         4        1.758\n       2025          deck_inventory_tag20 13      7         6        3.582\n       2025        deck_auto_battle_tag20 17     15        10        2.998\n       2025      deck_tower_defense_tag20  6      3         1        0.957\n       2025   deck_action_roguelike_tag20 12      3         2        1.339\n       2025         deck_any_hybrid_tag20 94     42        26        1.759\n       2025 deck_no_selected_hybrid_tag20 75     20        15        1.199\n    2026_H1      deck_party_tactics_tag20 46     12         8        1.335\n    2026_H1   deck_economy_building_tag20 14      5         4        2.025\n    2026_H1          deck_inventory_tag20 10      5         3        2.196\n    2026_H1        deck_auto_battle_tag20 18      3         1        0.664\n    2026_H1      deck_tower_defense_tag20  4      1         1        2.609\n    2026_H1   deck_action_roguelike_tag20  6      2         1        1.271\n    2026_H1         deck_any_hybrid_tag20 75     20        14        1.451\n    2026_H1 deck_no_selected_hybrid_tag20 57     13         8        1.291\n2026_JulAug      deck_party_tactics_tag20 27      5         5        1.594\n2026_JulAug   deck_economy_building_tag20  6      2         2        4.723\n2026_JulAug          deck_inventory_tag20  4      1         1        2.618\n2026_JulAug        deck_auto_battle_tag20  7      2         2        3.650\n2026_JulAug      deck_tower_defense_tag20  2      0         0        0.000\n2026_JulAug   deck_action_roguelike_tag20  5      1         1        1.563\n2026_JulAug         deck_any_hybrid_tag20 39      8         8        1.877\n2026_JulAug deck_no_selected_hybrid_tag20 28      3         2        0.810\n  appid                                           name    label  explicit_interaction  reviews  positive_pct  usd_list_price\n2797650                                    Pip My Dice concrete                 False      135            87           10.99\n2862890                                Become the Moon  generic                 False      172            77           16.99\n3333700                                 The Royal Writ concrete                 False      419            80           14.99\n3509430                                       Replicat concrete                 False      213            99            9.99\n2903630                            Millennium Strategy  generic                 False       21            85           14.99\n3046020                             Age of Enchantment concrete                 False       37            78            9.99\n3332060                             Feng Shui Meowjong  unclear                 False      276            96            4.99\n4610000                                        kill -9  generic                 False        5           100            9.99\n2372240                                       Re:Night concrete                  True      125            88           13.99\n4332200                                    King’s Well concrete                  True       17            76           14.99\n4571100                             English Vocab Rush  generic                 False        1             0            6.99\n4948610                                          Alchi  unclear                 False       21           100            4.99\n2745490                                Vanguard Exiles  unclear                 False       61            81           19.99\n2849990                                   Auto Dungeon  generic                 False       22            95            9.99\n3787880  Pattern Survivors: Bullet Hell — Create Chaos concrete                 False        5            80             NaN\n3950050                    Touhou Witch's Night Market  generic                 False      235            93            4.99\n3584870                                    Lucha Chess  generic                 False       60            86            8.99\n3809330                                  Hill Defender  generic                 False        9            77            3.99\n4765580                           Brainrot Autobattler  generic                 False       11            90            3.99\n4796920                                       Dynasty! concrete                 False        8            87           14.99\n3816000                            Flowers and Deities concrete                  True      239            89           12.99\n4223200                                  Ars Mechanica  generic                 False        3            66            4.99\n4897540                                 Miner Survivor concrete                  True        0             0            5.99\n4934320                                      SlimeHold concrete                 False        0             0            7.99\n2862290                                      Gunswitch concrete                  True       25            64            7.99\n3217930                                      BeefArena  unclear                 False        5           100           14.99\n3781740                                  Vanquish Evil concrete                  True        4            75            8.99\n4232170                                       Emberion  generic                 False        1             0            1.99\n2955630                             Catjinni Survivors  generic                 False        6           100            9.99\n3902210                                 Laby's Village  generic                 False        2            50            9.99\n4522710                                     Swarmbound  generic                 False        1             0            2.99\n4730540                                      TAILROUGE  generic                 False        0             0            1.99\n4711500                                     Qarma Quad concrete                 False        2           100           10.00\n4785950                                   Pirate Twist concrete                  True        4           100            8.99\n4922660                             Frontier Survivors concrete                  True        1             0            4.99\n4952430                                       Sangreve concrete                  True        1             0            4.99\n3348510                            Tower Arena Survive concrete                 False        3            66            4.99\n3610530              D.P.S: Weapons Testing Facility 2 concrete                  True        3           100            9.99\n3703380                           Fantasy Monster Hunt concrete                  True       43            79            5.99\n3916180                               Transport Legion concrete                  True       76            81            8.99\n4279880                                     Misty Land  generic                 False       24            79            8.99\n4523160                                    Slot Theory  unclear                 False        7           100            6.70\n4666840                          Glowbot Tower Defense concrete                  True        0             0            4.99\n4785570                                Ghosts Keep Out  generic                 False      130            76            3.99\n3998900                                 Claws Of Power  generic                 False      162            81            6.99\n4402950                     HEXSTORM: Tears of Arcadia concrete                  True       53            75            8.99\n4799450                            Kingdoms vs Zombies  generic                 False       45            82            7.99\n4918020                          Orbital Tower Defence concrete                 False        3           100            0.99\n1004640 FINAL FANTASY TACTICS - The Ivalice Chronicles concrete                  True     5633            89           49.99\n3540430                                  Dungeon Weiqi concrete                  True        0             0            6.99\n3604960                         Legacy of Shadows 影の遺産  generic                 False        0             0            8.99\n3927630                         魔姬征服者  Twelve Succubus  generic                 False        4            50            2.99\n2529790                                       GRIME II concrete                 False     1861            86           27.99\n4402660                          METRO QUESTER | OSAKA  generic                 False        5           100           19.99\n4631030                                 TarotBullFight  generic                 False        0             0            1.99\n4775190                                      BlassHole  generic                 False        0             0            1.99\n4469500                           Ninety Nine Dungeons  generic                 False        0             0            4.99\n4840190                                       VOIDRAVE  unclear                 False        3           100            4.99\n4894180                                  Thrice-a-dice concrete                  True        1           100            7.99\n4973340                                     Neko Rogue concrete                  True        1           100           12.99\n"}

**Computing uncertainty bounds**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"eefa17","wall_time_seconds":1.001178022,"session_id":29167,"original_token_count":0,"output":""}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"9963c3","wall_time_seconds":0.000001797,"exit_code":0,"original_token_count":1169,"output":"GROWTH BY PRICE/TEXT\n         axis  year   value  eligible  builds  share\nprice_bucket  2025   10-20      1468     224  0.153\nprice_bucket  2025   20-40       377      44  0.117\nprice_bucket  2025    5-10      2394     282  0.118\nprice_bucket  2025     <=5      4596     264  0.057\nprice_bucket  2025     >40       140      10  0.071\nprice_bucket  2025     NaN       165      19  0.115\nprice_bucket  2026   10-20      2030     385  0.190\nprice_bucket  2026   20-40       364      41  0.113\nprice_bucket  2026    5-10      3857     592  0.153\nprice_bucket  2026     <=5      5921     669  0.113\nprice_bucket  2026     >40       106      12  0.113\nprice_bucket  2026     NaN       131      18  0.137\n text_bucket  2025 151-300      4465     387  0.087\n text_bucket  2025 301-600      2334     355  0.152\n text_bucket  2025   <=150      2059      60  0.029\n text_bucket  2025    >600       282      41  0.145\n text_bucket  2026 151-300      5824     694  0.119\n text_bucket  2026 301-600      3810     740  0.194\n text_bucket  2026   <=150      2148     123  0.057\n text_bucket  2026    >600       627     160  0.255\nCREATORS\n  year          role  builds  unique  top1_games  top10_games  top10_share\n 2025 developer_key     843     831           3           21        0.025\n 2025 publisher_key     843     797          10           36        0.043\n 2026 developer_key    1717    1677           3           26        0.015\n 2026 publisher_key    1717    1636           6           38        0.022\nTEXT STANDARDIZATION\n  year  raw_build_share  standardized_to_2025_text_length\n 2025           0.0922                            0.0922\n 2026           0.1384                            0.1286\nDECK LANGUAGE\n      period  deck_language   n  ge100  liked100  joint_ratio\n       2025           True 135     48        34        1.503\n       2025          False  34     14         7        1.497\n    2026_H1           True 111     28        20        1.494\n    2026_H1          False  21      5         2        0.815\n2026_JulAug           True  49     10         9        1.782\n2026_JulAug          False  18      1         1        0.595\nINVENTORY AUDIT\n      period  spatial_layout_pitch  n  ge100  liked100\n       2025                  True  2      2         2\n       2025                 False 11      5         4\n    2026_H1                  True  4      4         2\n    2026_H1                 False  6      1         1\n2026_JulAug                  True  1      0         0\n2026_JulAug                 False  3      1         1\n   combined                  True  7      6         4\n   combined                 False 20      7         6\nGALLERY\n   appid                              name                first_date  reviews  positive_pct  usd_list_price\n2500460                        Inventorix 2025-09-01 12:28:47+00:00      131           100            4.99\n2652600          SealChain: Call of Blood 2026-02-20 10:56:16+00:00      533            56            8.99\n2686020                  Faith in Despair 2025-11-12 07:02:20+00:00       74            95           11.99\n3284290                   Moonsigil Atlas 2026-05-28 13:10:16+00:00      259            82           19.99\n3333700                    The Royal Writ 2025-08-07 16:59:28+00:00      419            80           14.99\n3355940                        Card Coder 2026-04-06 12:54:43+00:00       27            74           14.99\n3380700                Papercraft Tactics 2026-07-24 13:00:31+00:00        4           100            7.99\n3509230                        Gambonanza 2026-05-01 13:00:14+00:00     1577            74           14.99\n3509430                          Replicat 2025-11-19 01:57:37+00:00      213            99            9.99\n3573070                 Toy Smash Kaboom! 2025-12-04 06:55:24+00:00      551            84           13.99\n3610530 D.P.S: Weapons Testing Facility 2 2025-10-22 11:32:27+00:00        3           100            9.99\n3703380              Fantasy Monster Hunt 2025-11-21 02:12:14+00:00       43            79            5.99\n3802470                         Turnbound 2026-01-22 12:06:06+00:00      441            78           14.99\n3816000               Flowers and Deities 2026-08-15 06:00:19+00:00      239            89           12.99\n4206270                         Wireworks 2026-03-09 15:58:57+00:00      602            89            6.99\n4332200                       King’s Well 2026-08-11 22:00:21+00:00       17            76           14.99\n4537600                 Dice Gun Commando 2026-07-24 17:00:45+00:00       20           100            4.99\n4730540                         TAILROUGE 2026-06-08 12:29:35+00:00        0             0            1.99\n"}
{"chunk_id":"7b8f5e","wall_time_seconds":0.000002122,"exit_code":0,"original_token_count":7,"output":"matplotlib True\nscipy True\n"}

**Preparing growth charts**

**Calculating threshold sensitivity**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"4cd693","wall_time_seconds":1.001775254,"session_id":87831,"original_token_count":0,"output":""}

[
  {
    "type": "text",
    "text": "Script completed\nWall time 0.0 seconds\nOutput:\n"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"bcb56b\",\"wall_time_seconds\":0.0000013,\"exit_code\":0,\"original_token_count\":14,\"output\":\"Saved growth and deck reception figures as PNG and SVG.\\n\"}"
  },
  {
    "type": "image",
    "note": "[image: shown below]"
  },
  {
    "type": "image",
    "note": "[image: shown below]"
  }
]

**Fixing currency label rendering**

**Validating pilot sample metrics**

Script completed
Wall time 1.2 seconds
Output:

{"chunk_id":"617cf1","wall_time_seconds":1.001479806,"session_id":77158,"original_token_count":0,"output":""}
{}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"9d1141","wall_time_seconds":0.000001634,"exit_code":0,"original_token_count":646,"output":"year                 2025  2026  delta\ntag                                   \nStrategy              164   384    220\n2D                    124   320    196\nRoguelike             107   289    182\nRoguelite              96   243    147\nCard Game              94   226    132\nCasual                 78   206    128\nRPG                   168   296    128\nSingleplayer           80   196    116\nDeckbuilding           68   176    108\nTurn-Based Tactics     76   177    101\nCard Battler           64   164    100\nTurn-Based Strategy    86   176     90\nIndie                  44   132     88\nAction                 94   166     72\nTurn-Based Combat      51   112     61\nPixel Graphics         71   129     58\nFantasy                65   122     57\nDungeon Crawler        47   100     53\nPvE                    37    87     50\nAdventure              92   139     47\nTop-Down               25    71     46\nArcade                 33    79     46\nDark Fantasy           21    63     42\nBullet Hell            19    60     41\nBoard Game             19    55     36\nTactical               13    48     35\nColorful               19    53     34\nTabletop               13    45     32\n3D                     44    75     31\n2D Platformer          28    58     30\nEarly Access           13    41     28\nSimulation             29    56     27\nPuzzle                 22    49     27\nAction-Adventure       38    65     27\nStylized               12    37     25\n     period        group  positivity  qualifying    n  rate\n       2025  deck_strict          75          53  169 0.314\n       2025  deck_strict          80          41  169 0.243\n       2025  deck_strict          85          29  169 0.172\n       2025 other_builds          75         173 1196 0.145\n       2025 other_builds          80         140 1196 0.117\n       2025 other_builds          85         100 1196 0.084\n    2026_H1  deck_strict          75          26  132 0.197\n    2026_H1  deck_strict          80          22  132 0.167\n    2026_H1  deck_strict          85          16  132 0.121\n    2026_H1 other_builds          75         100  953 0.105\n    2026_H1 other_builds          80          79  953 0.083\n    2026_H1 other_builds          85          53  953 0.056\n2026_JulAug  deck_strict          75          11   67 0.164\n2026_JulAug  deck_strict          80          10   67 0.149\n2026_JulAug  deck_strict          85           8   67 0.119\n2026_JulAug other_builds          75          45  565 0.080\n2026_JulAug other_builds          80          35  565 0.062\n2026_JulAug other_builds          85          18  565 0.032\n"}
{"chunk_id":"c26000","wall_time_seconds":0.00000335,"exit_code":0,"original_token_count":696,"output":"period,field,rule,removed_key,removed_name,removed_n,n,under10,ge10,ge50,ge100,ge200,ge1000,liked100,rate100,liked100_rate,satisfied_among100,median_reviews,median_positive_among100,median_price,developers,publishers,reviews_sum,top1_review_share,joint_ratio\n2025,developer_key,most_reviews,clan:44900993,Sad Socket,1,168,38,130,80,61,36,13,40,0.3630952380952381,0.23809523809523808,0.6557377049180327,38.5,84.0,9.99,168,159,56140,0.15345564659779123,1.480667750259881\n2025,developer_key,most_games,clan:31494034,Topstitch Games,1,168,38,130,80,61,36,14,40,0.3630952380952381,0.23809523809523808,0.6557377049180327,38.5,84.0,9.99,168,158,77111,0.27530443127439663,1.4775587422096015\n2025,publisher_key,most_reviews,clan:38719421,Hooded Horse,3,166,38,128,78,59,34,11,38,0.35542168674698793,0.2289156626506024,0.6440677966101694,37.0,83.0,9.99,166,158,46738,0.1843253883349737,1.4321387681763473\n2025,publisher_key,most_games,clan:44794657,Rogue Duck Interactive,4,165,38,127,78,59,36,14,40,0.3575757575757576,0.24242424242424243,0.6779661016949152,37.0,84.0,9.99,165,158,76556,0.27730027692146925,1.50673921079076\n2026_H1,developer_key,most_reviews,clan:33974508,Mega Crit,1,131,56,75,41,32,20,7,22,0.24427480916030533,0.16793893129770993,0.6875,14.0,85.0,9.99,131,130,38789,0.4733816288122921,1.3952288515576234\n2026_H1,developer_key,most_games,clan:33018175,Family Devs,1,131,56,75,42,33,21,8,22,0.25190839694656486,0.16793893129770993,0.6666666666666666,14.0,84.0,9.99,131,130,220792,0.8243958114424436,1.3960139417533943\n2026_H1,publisher_key,most_reviews,clan:33974508,Mega Crit,1,131,56,75,41,32,20,7,22,0.24427480916030533,0.16793893129770993,0.6875,14.0,85.0,9.99,131,130,38789,0.4733816288122921,1.3952288515576234\n2026_H1,publisher_key,most_games,clan:38394018,indie.io,2,130,56,74,40,32,20,8,21,0.24615384615384617,0.16153846153846155,0.65625,13.5,84.0,9.99,130,130,220205,0.8265934016030517,1.3477012535242856\n2026_JulAug,developer_key,most_reviews,name:crux games,Crux Games,1,66,32,34,18,10,3,0,9,0.15151515151515152,0.13636363636363635,0.9,10.0,87.5,8.155000000000001,66,66,3489,0.16165090283748926,1.3606189830549473\n2026_JulAug,developer_key,most_games,clan:42848556,Just L Studio,1,66,31,35,19,11,4,0,10,0.16666666666666666,0.15151515151515152,0.9090909090909091,10.0,88.0,8.155000000000001,66,66,4437,0.21478476448050485,1.5085351000612388\n2026_JulAug,publisher_key,most_reviews,name:crux games,Crux Games,1,66,32,34,18,10,3,0,9,0.15151515151515152,0.13636363636363635,0.9,10.0,87.5,8.155000000000001,66,66,3489,0.16165090283748926,1.3606189830549473\n2026_JulAug,publisher_key,most_games,clan:33025551,Goblinz Publishing,1,66,32,34,18,10,4,0,9,0.15151515151515152,0.13636363636363635,0.9,10.0,88.0,8.155000000000001,66,66,4279,0.22271558775414815,1.377541857753452\n"}

**Checking hybrid tag counts**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"00e78d","wall_time_seconds":1.001336127,"session_id":41298,"original_token_count":0,"output":""}

Script completed
Wall time 1.7 seconds
Output:

{"chunk_id":"611d31","wall_time_seconds":1.000961232,"session_id":88754,"original_token_count":0,"output":""}
{"chunk_id":"56e8a5","wall_time_seconds":0.370543308,"exit_code":0,"original_token_count":187,"output":"GROWTH PCT {'action': 106.2, 'all': 103.7, 'auto': 86.3, 'deck': 95.1, 'none_of_four': 104.7, 'tower': 106.2}\nGROWTH CONTRIBUTIONS {'action': 32.0, 'auto': 5.2, 'deck': 9.9, 'none_of_four': 46.2, 'tower': 6.7}\nDECK HIGHER THRESHOLD CELLS 27 27\n          hybrid   n  ge100  liked100  within_deck_joint_ratio\n   party_tactics  82     24        17                    1.034\neconomy_building  20      9         5                    1.175\n       inventory  18     10         8                    2.179\n     auto_battle  31     12         7                    1.212\n   tower_defense  12      4         2                    0.784\naction_roguelike  19      4         3                    0.724\n             any 152     48        32                    1.048\n"}

**Verifying findings file**

[
  {
    "type": "text",
    "text": "Script completed\nWall time 0.0 seconds\nOutput:\n"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"8804f1\",\"wall_time_seconds\":0.000001075,\"exit_code\":0,\"original_token_count\":5,\"output\":\"Passed 45 checks.\\n\"}"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"e3a659\",\"wall_time_seconds\":0.000001532,\"exit_code\":0,\"original_token_count\":14,\"output\":\"Saved growth and deck reception figures as PNG and SVG.\\n\"}"
  },
  {
    "type": "image",
    "note": "[image: shown below]"
  }
]

Script completed
Wall time 0.0 seconds
Output:

{}
{"chunk_id":"cdd99e","wall_time_seconds":0.000002915,"exit_code":0,"original_token_count":93,"output":"-rw-r--r-- 1 adn adn 214K Sep  6 01:21 steam-research/continuation-2026-09-06/deck_reception.png\n-rw-r--r-- 1 adn adn  15K Sep  6 01:25 steam-research/continuation-2026-09-06/findings.md\n-rw-r--r-- 1 adn adn 172K Sep  6 01:21 steam-research/continuation-2026-09-06/growth.png\n-rw-r--r-- 1 adn adn 2.5K Sep  6 01:22 steam-research/continuation-2026-09-06/validation.json\n"}

**Checking report calculations**

[sealed content withheld: memory]

**Reviewing public post metadata**

Script completed
Wall time 0.1 seconds
Output:

Warning: truncated output (original token count: 17349)
Total output lines: 3

{"status":"fulfilled","value":{"chunk_id":"8d36a2","wall_time_seconds":0.000004118,"exit_code":0,"original_token_count":85294,"output":"Warning: truncated output (original token count: 85294)\nTotal output lines: 1966\n\ntotal 336\ndrwxr-xr-x 1 adn adn    120 Sep  2 21:08 .\ndrwxr-xr-x 1 adn adn    256 Sep  3 18:48 ..\n-rw-r--r-- 1 adn adn  30832 Aug 30 10:19 emergent-exam-genealogy.md\n-rw-r--r-- 1 adn adn 293950 Sep  4 11:43 home.md\n-rw-r--r-- 1 adn adn  16108 Jun 11 00:14 td-genealogy-and-card-td.md\nTitle: Non-Fixed Exams: A Genealogy of Emergent Challenge\nDate: 2026-06-11 22:57:02\n\n# Non-Fixed Exams: A Genealogy\n\nCompanion to [the TD genealogy map](/td-genealogy-and-card-td). That map covered the **build-feed side** of the family — what randomizes the stream of units/cards/towers you build from. This map covers the dual question: the **exam side** — what generates the challenge your build is graded against, and what happens to each answer over time.\n\nVocabulary carried over from the Micro/Meso/Macro session and the Bayesian-epistemology discussion that followed it:\n\n- **Exam** — the thing that grades the build: wave schedules, maps, fights, raids, weather.\n- **Fixed exam** — authored once, identical across runs. Produces answer keys (Bloons CHIMPS guides, Arknights stage videos).\n- **The imagined-wiki test** — a mode is collapse-resistant to the degree that its hypothetical wiki must be a *textbook of methods* rather than a *table of values*, and actually unbreakable only if even the methods textbook goes stale because runs keep demanding concepts not yet in it.\n- **The Ngo rule** — a game is breakable once its hypothesis space is enumerated; published distributions are pre-constructed model spaces handed to the community. So: *publish the rules, never print the distributions.* An emergent exam is one where the distribution exists nowhere — not in data tables, not in the design doc, not in the designer's head — because it is generated, not authored.\n\n---\n\n## 1. The six answers to \"who authors the exam?\"\n\nThe card-TD genealogy was a trunk built from one question asked repeatedly: *what randomizes the build feed?* The exam side has its own recurring question — *who authors the exam?* — and every game in and around the family gives one of six answers.\n\n| # | Answer | Exemplars | Uncertainty type | Collapse signature (imagined wiki) |\n|---|--------|-----------|------------------|-------------------------------------|\n| 1 | **Nobody — authored once** (fixed) | Classic TD, Arknights, puzzle games | None | **Answer key.** Per-map build orders. Total collapse. |\n| 2 | **A distribution, sampled per run** | Rogue → NetHack → Spelunky → StS → Monster Train | Aleatory, distribution authored | **Method wiki.** Tier lists + pathing heuristics + EV charts. Slow collapse, Bayesian-ready from birth. |\n| 3 | **The player** (scheduled) | StS routing, Against the Storm forest modifiers, ascension/heat, AI War's AIP, RoR's difficulty clock | Aleatory + self-referential | **Routing EV tables.** Collapses unless the thing being scheduled is itself uncertain. |\n| 4 | **A concealed director** | Mario Kart rubber-banding (1992) → RE4 (2005) → L4D Director (2008) → RimWorld storytellers | Epistemic (hidden policy) | **Director-reading.** Players reverse the policy, then suppress its input variable. Also fails the fairness/attribution constraint outright. |\n| 5 | **Other minds** | Wintermaul/Legion sends → auto-battler PvP → battle royales | Entity (non-stationary) | **Unbreakable — but it's the venue, not the game.** The industry's default fix for exam staleness is \"add humans.\" Rejected per the session's constitution-vs-citizens cut. |\n| 6 | **A visible simulation** | Dwarf Fortress, Factorio, Rain World, Creeper World, ONI, OTC | Emergent (irreducible or reflexive) | **Field guide** at best — climate knowledge that never resolves into trajectories. The only branch whose best members pass the stale-textbook bar. |\n\nThree observations about the table before the islands:\n\n**Branches 2–4 are all Bayesian-ready.** Sampled exams publish their distribution through volume and datamining; scheduled exams publish their options on screen; directors get reverse-engineered into an effective distribution. In each case the community ends up holding an enumerated hypothesis space, and play inside it becomes credence-updating — which automates. The collapse rates differ by orders of magnitude (StS is still partially alive at year eight; a director gets read in months), but the asymptote is shared.\n\n**Branch 3 is not really an uncertainty source — it's a commitment amplifier.** Scheduling your own exam (wave drafting, StS pathing, AStS modifiers) is the aleatory commit engine pointed at the exam side. It composes beautifully with everything, but on its own it grades against known quantities. Note that Risk of Rain's difficulty clock belongs here and is the reason looting feels like a routing problem: the player schedules the exam continuously by choosing how long to loot. The clock is the most elegant member of this branch — and it still produced wiki pages of stage-timing rules, on schedule.\n\n**Branch 6 is the subject of this document, and its headline property is negative: it has no phylogeny.** The fixed-exam tree from the previous map is a dense lineage — every generation visibly inherits organs from its parent. The emergent branch is a scattering of islands separated by decades, where almost nobody iterated on anybody. The rest of this doc walks the islands, explains why no lineage formed, and extracts what's portable.\n\n---\n\n## 2. The islands, chronologically\n\nFor each island: the mechanism, its visibility grade, its coupling to the player, the shape of its wiki, and the lesson.\n\n### 2.1 The economy root — The Sumerian Game (1964) / Hamurabi (1968/73)\n\nThe oldest digital exam is emergent-ish. Hamurabi has no levels, no waves, no authored challenge at all: you allocate grain, the population/land/harvest difference equations respond, and starvation is the grade. The dynamics ARE the exam. The lesson is genealogical: management games *began* on this branch — the fixed exam (levels, waves) is the later invention, imported from arcade structure. The lineage is old; it's just thin, because the dynamics in this root were simple enough to solve by hand.\n\n### 2.2 M.U.L.E. (1983) — the market made of players\n\nAuction-driven economy where prices emerge from supply, demand, and the other three players' positions. Half entity-exam (branch 5), half economy-exam, but it established the reflexive-market mechanism: **a price that responds to your own purchases is partially unpredictable to you even with perfect information, because predicting it requires predicting your own future trades.** Self-reference as an uncertainty source, four decades early.\n\n### 2.3 The divorce — SimCity (1989) and the Wright line\n\nThe single most important event in the genealogy, and it's an absence. Will Wright built city dynamics out of Jay Forrester's system-dynamics work — exactly the simulation substance an emergent exam needs — and then deliberately refused to grade the player. \"Software toys\": emergence kept, exam dropped. SimEarth, SimAnt, The Sims doubled down. **The people best at building dynamics opted out of grading; the people who grade (game designers proper) kept their exams authored.** The two skills split into separate genres in 1989 and have barely re-merged since. This divorce, more than any technical barrier, is why branch 6 is islands instead of a tree.\n\n### 2.4 Ultima Online's ecology (1997) — eaten by its players\n\nThe first industrial attempt to ship an ecology *as content*: Raph Koster's closed-loop resource system, with herbivores, predators, and resource flows that player actions fed into. It died almost immediately — players slaughtered and harvested everything faster than the loop could regenerate, before most of them ever noticed the system existed. The canonical lesson: **players are an extraction and optimization pressure that exceeds any closed loop's regeneration rate, given unbounded time.** UO gave them unbounded time (persistent world). The fix nobody stated for twenty years: bound the time. See §5.3.\n\n### 2.5 Dwarf Fortress (2006) — the re-merge\n\nThe island where sim and exam finally re-married. The exam emerges from several stacked systems: sieges scaled by created wealth (a scalar coupling — and DF players do thermostat it; \"wealth management\" predates RimWorld), fluid and temperature physics, animal and civilization populations, and above all **cascades** — the tantrum spiral, where one death sours moods, sour moods cause fights, fights cause deaths. The cascade is the genuinely emergent member: nobody authored it; it falls out of the social sim's coupling.\n\nThe wiki test grades DF honestly: the buildable layer wiki'd completely (quickfort blueprints, optimal fort layouts — the ONI failure mode, see 2.13), the *trajectory* layer never did. The DF wiki is a field guide — what creatures do, what moods mean — not an answer key; what veterans hold is exactly the unwritable embodied model.\n\nTwo more things DF shipped that matter here:\n\n- **\"Losing is fun\"** — the community's cultural technology for variance acceptance. An emergent exam will occasionally hand you an unwinnable cascade, and DF's audience converted that from a complaint into the genre's motto. This is the equanimity gym norm, named and adopted at scale. It proves the audience for attribution-ambiguous deaths exists — and that it's a *culture* you have to build, not just a mechanic you ship.\n- **The legibility bill, unpaid.** DF's infamous interface doubles as an instruments failure: the dynamics are in-principle visible and in-practice illegible. DF survived it on depth; nothing else has.\n\n### 2.6 S.T.A.L.K.E.R.'s A-Life (2007) — emergence as set dressing\n\nOffline faction/mutant simulation populating the Zone, producing emergent encounters — around an exam (the shooting) that stays authored. This is the common industrial compromise: emergence relegated to *texture* while the graded loop stays fixed. Worth a node b…80294 tokens truncated… authors the distribution\nyou draw from. Moving authorship to the player (deck construction) buys: pristine\nattribution (every bad draw is variance against a distribution *you* built — deck wrong,\nsequenced wrong, or variance, the poker triad), a front-loaded draft skill, and the\nhold/play hand reads that shops don't have.\n\n- **Added (from the card lineage):** drafted deck → shuffled draws; hand of 5, +2 per\n  wave; mana per wave as the spend constraint; improvement cards (upgrades become\n  aleatory deliveries).\n- **Kept (from the root):** full spatial placement, scripted previewed waves, lives.\n  The two organs the auto-battler line abandoned — the maze and the static tower —\n  come back intact.\n- **Absent (so far):** all the auto-battler organs (shop, merge, interest, traits,\n  pool) — available as ports, see §5.3 — and any micro.\n\n---\n\n## 3. The full mechanic matrix\n\nStatus is relative to the column's parent. Card-TD's parent is the **root chassis**;\nits `NEW` cells arrive from the card-game donor lineage.\n\n| Mechanic | Classic TD | Wars TD (Legion) | Auto-battler | SNKRX | Card-TD (proposed) |\n|---|---|---|---|---|---|\n| **Build feed** | full catalog, gold-gated | MOD: catalog narrowed by builder pick | NEW: random shop, reroll, lock | KEPT: random shop, reroll, lock | NEW: drafted deck → drawn hand |\n| **Economy** | kill bounty + wave bonus | MOD: kill gold + send income | NEW: round income, interest, streaks | KEPT: gold + interest | NEW: mana per wave (carryover TBD) |\n| **Spatial layer** | free placement, mazing | MOD: build-zone formations | MOD: grid, front/back | NEW: snake order + live steering | KEPT (root): full placement returns |\n| **Combat agency** | static towers auto-fire | MOD: towers unroot and fight | KEPT: units fight, you spectate | MOD: units fire while you carry them | KEPT (root): static towers again |\n| **Upgrades** | pay-to-upgrade tiers | KEPT: per-unit lines | NEW: 3-copy merge | KEPT: copies → levels | NEW: improvement cards, drawn |\n| **The exam** | scripted waves, previewed | MOD: scripted + opponent sends | NEW: mirror PvP vs other boards | REVERTED: scripted PvE + NG+ | KEPT (root): scripted waves ⚠ |\n| **Player micro** | ≈ none (ability timing) | none | none in combat | NEW: continuous steering | ABSENT (open design choice) |\n| **Synergies** | implicit combos (slow+dmg) | MOD: builder/race identity | NEW: tagged trait thresholds | KEPT: class thresholds | NEW: deck-authored synergy density |\n| **Uncertainty profile** | none | opponent intent (entity) | shop rolls (aleatory) + opponents (entity) | shop rolls (aleatory) | draw order (aleatory, self-authored) |\n\nMicro/mezo/macro coordinates per node:\n\n| Game | Micro | Mezo | Macro |\n|---|---|---|---|\n| Classic TD | ~0 | 0 | dominant (solvable optimum) |\n| Wars TD | ~0 | entity (sends, timing) | heavy |\n| Auto-battler | 0 | aleatory + entity | heavy (econ + comps) |\n| SNKRX | steering | aleatory (shop) | heavy (builds + econ) |\n| Survivors-likes | steering/kiting, heavier | aleatory (offers) | builds |\n| PvZ | light (sun, timing) | light | dominant |\n| Monster Train | 0 | aleatory (draws + drafts) | dominant |\n| Arknights | timing-lite | ~0 | dominant (guide-followable) |\n| **Card-TD** | **0 (so far)** | **aleatory: draft + draws** | **double: spatial + deck construction** |\n\n---\n\n## 4. The constellation — what each neighbor proves or warns\n\n- **Plants vs Zombies (2009)** — the proto card-TD, and proof the chassis hybridizes:\n  seed loadout = fixed deck (no draws), sun = mana, cooldowns = pacing. What it lacks\n  is exactly what you're adding — draw RNG and a real draft. PvZ shows the *fairness*\n  of the form; your additions supply the *mezo*.\n- **Monster Train (2020)** — the nearest realized relative: deck + hand + mana + lane\n  defense, drafted *during* the run against varied encounters. Proof the full system\n  works and stays alive for hundreds of hours. Note what it does differently from your\n  spec: interleaved drafting and a non-fixed exam. That difference is the whole\n  collapse question.\n- **Arknights (2019)** — the cautionary neighbor: drafted squad + mana-like deployment\n  points + true spatial TD, but **no draw randomness and fixed exams**. Result: the\n  endgame is guide-following — per-stage video answer keys. This is your design with\n  the aleatory layer removed; it demonstrates precisely why the card layer needs draws.\n- **Ratropolis (2020)** — real-time card-driven defense; proof the hand-under-pressure\n  loop works in real time, at some cost in legibility (chaos at high speed).\n- **Legion TD 2 (2021)** — the wars branch alive and standalone; the income/send axis\n  remains the deepest *economy-as-aggression* design in the family if you ever want a\n  competitive mode reference.\n- **Super Auto Pets (2021)** — the auto-battler reduced to its minimum organs (shop,\n  merge, async PvP); useful as proof of how few parts the shop engine needs.\n- **Survivors-likes** — what happens when the SNKRX formula concentrates into one body;\n  relevant mostly as the genre your card-TD is deliberately *not* (no micro).\n- **Mini Looper (yours, 2026)** — SNKRX × TD with steering. The boundary matters: see §5.4.\n\n---\n\n## 5. The card-TD analysis\n\n### 5.1 Lineage position\n\nA sibling branch of the auto-battler: both transplant a randomized build feed onto the\nTD chassis. The auto-battler's feed is game-authored (shop) and came bundled with PvP;\nyours is player-authored (deck) and stays PvE. Because it branches from the root, it\n*recovers the two organs the auto-battler line lost* — the maze and the static tower —\nand skips the organs the auto-battler invented (those remain portable, §5.3).\n\nIn the micro/mezo/macro frame: **macro-heavy on two independent layers** (spatial\nplacement + deck construction), **mezo entirely aleatory and self-authored** (draft\ncommitment + draw order + mana sequencing — the cleanest attribution profile in the\nfamily), **micro zero**. Nearest framework neighbors: Hearthstone/Artifact territory,\nbut with TD's spatial macro substituting for the opponent. One honest description:\n**single-player Artifact where the opponent is the wave schedule** — deck, mana, hand,\nboard, no execution layer, judged by variance.\n\n### 5.2 The novel interaction\n\nSpatial irreversibility × draw uncertainty: placing a permanent, lossy-to-undo tower\nwhose value depends on cards not yet drawn (\"building the socket before drawing the\ngem\"). Card games can't produce this (boards are cheap to rearrange); TD can't (the\ncatalog is always available). This bet type is the design's identity — protect it.\n\n### 5.3 Organs on the shelf (auto-battler parts, card-frame translations)\n\n| Auto-battler organ | Card-frame port |\n|---|---|\n| 3-copy merge | duplicates in deck: drawing/playing a copy of a placed unit upgrades it instead of placing new — makes playset count a deck-construction decision (consistency vs power) |\n| Interest | mana banking: unspent mana carries over with a bonus, capped |\n| Reroll | cycling: discard a card to draw a card, at a cost |\n| Shop lock | native — holding cards in hand IS the lock |\n| Trait thresholds | tribal counts among *placed* towers — converts placement into synergy commitment (very TD-flavored) |\n| Counted pool | playset limits in deck-building rules (singleton vs 2x vs 4x) |\n| Win/loss streaks | perfect-wave / leak-free bonuses |\n| Creep sends | wave-drafting contracts: choose your next wave from fully-specified options with reward differentials |\n\n### 5.4 Risks, from the map\n\n1. **The Arknights collapse.** Fixed map + fixed waves + known deck → per-map deck tier\n   lists + contingency flowcharts → guide-following. Draw variance against a fixed exam\n   perturbs *means*, not *goals* — it jitters the timing of a known plan. Mitigations,\n   in increasing order of cost: partial exam preview at draft time (draft against wave\n   *types*, not the full schedule); interleaved drafting (draw 2 + draft 1 per wave —\n   restores the pick→evidence→pick loop); exam-side variance (wave drafting, procedural\n   maps). The Monster Train–Arknights delta is exactly this dial.\n2. **The decorative card layer.** If mana and draws are generous enough that you can\n   always afford an answer and always have one, the deck degenerates into a slow\n   catalog and the design is \"TD with extra steps.\" Monster Train avoids this with\n   genuinely tight mana and real scarcity. Starvation is load-bearing.\n3. **The Mini Looper boundary.** Card-TD with a steerable hero = Mini Looper with extra\n   steps; the two projects blur. The card-TD's identity is the *no-micro, deck-driven*\n   corner — Artifact-vs-environment. Keep the verbs disjoint: Mini Looper steers,\n   card-TD schedules.\n\n### 5.5 Open design dials (each small, each governs a lot)\n\n- **Draft timing:** all pre-map vs interleaved per-wave (the collapse-rate dial).\n- **Exam visibility at draft:** full schedule / wave types only / map only.\n- **Mana carryover:** none vs banked vs banked-with-interest.\n- **Resource count:** mana only, or mana + kill gold as a second currency (kill gold\n  preserves the root's \"defense funds defense\" loop; mana-only is cleaner but cuts the\n  TD economy organ entirely).\n- **Playset rules:** singleton / limited copies / free — interacts with merge-on-duplicate.\n- **Hand limit + discard:** whether holding has a price (discard decisions = pricing\n  future variance).\n- **Micro:** none (recommended for identity) vs min…7349 tokens truncated… project involves mi\n\nFILE 2026-07-03-230041.md \n Title: I've been testing Fable 5 and wondering why\nDate: 2026-07-03 23:00:41\nKind: message\nLink: /messages/2026-07-03-230041\n\nI've been testing Fable 5 and wondering why exactly it feels better than Opus 4.8 and it's hard to pinpoint it, but it is undeniably better. I've seen other people mention that it is more thorough/autistic but in a kind of exploratory way, which aligns with what I've found in my experiments. It seems like a really good mod\n\nFILE 2026-07-04-123056.md \n Title: Added a chat feature to the website\nDate: 2026-07-04 12:30:56\nKind: message\nLink: /messages/2026-07-04-123056\n\nAdded a chat feature to the website, the bubble on the bottom right. I think it's a huge waste of opportunity that most games do not have a global chat on their main menu, so I've been meaning to do things like this for a while. This is not a global chat, but it's part of said feature that I'll likely add in my next game. One of t\n\nFILE 2026-07-07-204709.md \n Title: Below is what Fable did in 2! (two) turns\nDate: 2026-07-07 23:51:40\nKind: message\nLink: /posts/2026-07-07-204709\n\nBelow is what Fable did in 2! (two) turns when I asked it to turn my engine, Anchor, which was built to be 2D from the start, no plans of it ever being 3D at all, into a 3D engine by using the just released Box3D. It just did it and it really just works, which is super impressive to me. We spent the rest of the session doing wo\n\nFILE 2026-07-09-124628.md \n Title: As I mentioned in the previous chat message, I\nDate: 2026-07-09 12:46:28\nKind: message\nLink: /posts/2026-07-09-124628\n\n::tweet https://x.com/deepfates/status/2074919039361826921\n::tweet https://x.com/deepfates/status/2074920392570159456\n\nAs I mentioned in the previous chat message, I think if you're doing anything online now it's increasingly useful to think of what you're building in terms of \"place.\" Is it a physical (not really in the r\n\nFILE 2026-07-10-202638.md \n Title: The number of new model releases just now is\nDate: 2026-07-10 20:26:38\nKind: message\nLink: /posts/2026-07-10-202638\n\nThe number of new model releases just now is pretty good, but I find it rather tiresome to keep trying them out, which I have scarcely done since year start and kept mostly to Claude. If Fable does indeed go away and I'm left with Opus 4.8, I'll probably check out the new Elon model, I've read it's pretty fast and I could us\n\nFILE 2026-07-12-010140.md \n Title: I like this guy's account but it's always\nDate: 2026-07-12 01:01:40\nKind: message\nLink: /posts/2026-07-12-010140\n\n::tweet https://x.com/QiaochuYuan/status/2076130613468504183\n\nI like this guy's account but it's always important to know who's saying something and to take their personality into account in relation to what they said. For instance, this is QiaochuYuan's experience with Obsession:\n\n::tweet https://x.com/QiaochuYuan/status/20591\n\nFILE 2026-07-13-101145.md \n Title: Thread by @robertlasagna1\nDate: 2026-07-13 10:11:45\nKind: message\nLink: /posts/2026-07-13-101145\n\n::tweet https://x.com/robertlasagna1/status/1976400962953908434\n::tweet https://x.com/robertlasagna1/status/1976401084097888394\n::tweet https://x.com/robertlasagna1/status/1976401176389353537\n::tweet https://x.com/robertlasagna1/status/1976401290516422726\n::tweet https://x.com/robertlasagna1/status/1976401425367433653\n\n\nFILE 2026-07-14-174351.md \n Title: Many comments on this ad, I disagree with most of\nDate: 2026-07-14 17:43:51\nKind: message\nLink: /posts/2026-07-14-174351\n\n::tweet https://x.com/claudeai/status/2075271759289303522\n\nMany comments on this ad, I disagree with most of them. Anthropic seems to be the only company modelling the future they're building seriously and acting in accordance with said future. The future is scary, and it should scare you (a little), and we should ask h\n\nFILE 2026-07-16-181804.md \n Title: God, I know I've prayed to you in the past asking\nDate: 2026-07-16 18:18:04\nKind: message\nLink: /posts/2026-07-16-181804\n\nGod, I know I've prayed to you in the past asking questions such as, \"Why do people say they're going out with me only to then disappear, vanish? Why can't they at least warn me it's not going to happen? Am I not interesting enough to be shown some basic amount of respect? Am I not pretty enough? Am I not funny enough? \n\nFILE 2026-07-17-111138.md \n Title: Well, I didn't get ghosted, if anything I got\nDate: 2026-07-17 11:11:38\nKind: message\nLink: /posts/2026-07-17-111138\n\nWell, I didn't get ghosted, if anything I got the opposite of ghosted, I'm fairly certain this will become the strongest love bombing then ghosting experience I'll ever have, but I'm here for it, being love bombed is truly marvelous if you know it's what's happening.\n\nOvernight reports on Kimi K3 from the general dev popula\n\nFILE 2026-07-17-222359.md \n Title: Below is Kimi K3's version of the Anchor 3D task\nDate: 2026-07-17 22:23:59\nKind: message\nLink: /posts/2026-07-17-222359\n\nBelow is Kimi K3's version of the Anchor 3D task I had Fable try a few days ago. We replicated the setup to be exactly the same as before (Anchor2 checkpoint), and then I had Fable run the process itself. Seems like Kimi did the job correctly, perhaps even slightly more tastefully than Fable. The version you see in the h\n\nFILE 2026-07-18-120613.md \n Title: Remember that by reading this website you are privileged\nDate: 2026-07-18 12:06:13\nKind: message\nLink: /posts/2026-07-18-120613\n\nRemember that by reading this website you are privileged, you are reading the thoughts of a seer or a sage, as in fast moving situations, individuals possessed of elevated intellect and wisdom tend to get to truthful places before everyone else. Below is Victor Taelin, a clearly very smart and popular individual,\n\nFILE 2026-07-21-215601.md \n Title: Naming the chess game the JAV way\nDate: 2026-07-21 21:56:01\nKind: message\nLink: /posts/2026-07-21-215601\n\n::tweet https://x.com/rezich/status/2079390581877301423\n\nTrue, but the solution is easy. If I release this chess game on Steam I'll go the JAV/LN route with a dash of supertf's stream titles and name it something like *\"Increasingly Powerful And Dominant Horse Relentlessly Tramples One Thousand Pawns\"* or *\"One Thousand Pawns March For\n\nFILE 2026-07-23-104718.md \n Title: Made this little gameplay video/trailer for horse\nDate: 2026-07-23 10:47:18\nKind: message\nLink: /posts/2026-07-23-104718\n\nMade this little gameplay video/trailer for horse game below. This combines several things I've always wanted to do that before would have been extremely tedious: a replay system, an in-engine clip gathering system using replays, an in-engine video editor using the clips (which are replay sections), and then rendering i\n\nFILE 2026-07-28-203142.md \n Title: Extremely impressive, one of the first games\nDate: 2026-07-28 20:31:42\nKind: message\nLink: /posts/2026-07-28-203142\n\n::tweet https://x.com/kevin_t_ngo/status/2081907248062665094\n::itch kengocodes/mecha-redline\n\nExtremely impressive, one of the first games I've seen of the \"Claude helped me make this in [short amount of time]\" genre that I was actually impressed by on a few levels. If true, it seems Claude is way more capable of generating \n\nFILE 2026-07-28-213723.md \n Title: Oh yea, it occurs to me I have not actually spoken\nDate: 2026-07-28 21:37:23\nKind: message\nLink: /posts/2026-07-28-213723\n\nOh yea, it occurs to me I have not actually spoken about what I think of the fact that the AI can help make games so easily now. In general, I think it's good. It follows from [Competition isn't real](/posts/competition) that I think more people being able to make better games benefits everyone, I genuinely believe thi\n\nFILE 2026-08-01-044746.md \n Title: Conclusion seems to be that the models are currently\nDate: 2026-08-01 04:47:46\nKind: message\nLink: /posts/2026-08-01-044746\n\nConclusion seems to be that the models are currently very good at doing static environments and machines, including machine animation, bad at doing organic beings and animations related to them. When asked, they say as much, and it's because \"correctness is externally verifiable,\" i.e. it's easy to mechanically verif\n\nFILE 2026-08-01-180118.md \n Title: This optionality point is obviously true\nDate: 2026-08-01 18:01:18\nKind: message\nLink: /posts/2026-08-01-180118\n\n::youtube 6ZaggWORtvo\n\n> **Spandrell** *(0:00)*: It really got me thinking about loyalty and how loyalty works. Back to Chateau Heartiste — this was a quip he would say all the time. He was like, loyalty only goes as far as your options, right? Women are only loyal as far as their options. The more options they have, the less lo\n\nFILE 2026-08-03-041701.md \n Title: In life, it is vital to take things as they are\nDate: 2026-08-03 04:17:01\nKind: message\nLink: /posts/2026-08-03-041701\n\nIn life, it is vital to take things as they are, otherwise you do not interact with the world, but with an illusion created by your own thoughts and desires and fears. An illusion world is finished, your mind made every part of it, so it holds nothing you didn't put in there, with nothing new ever being able to enter. But\n\nFILE 2026-08-07-113414.md \n Title: Being a beautiful favelado mid is so fucking funny\nDate: 2026-08-07 11:34:14\nKind: message\nLink: /posts/2026-08-07-113414\n\nBeing a beautiful favelado mid is so fucking funny. The other day I was on grindr and a guy named \"tesao em feio\" (horny for ugly dudes) messages me and says \"oier 😍\" (hiii), I'm laughing and laughing and then we talk a little and he says he's just doing this to mess with \"padrão\" guys, conventionally attractive people\n\nFILE 2026-08-07-114354.md \n Title: I think I am binary with liking/disliking people\nDate: 2026-08-07 11:43:54\nKind: message\nLink: /posts/2026-08-07-114354\n\n> Within 10 seconds of meeting an autistic person, strangers rate them as awkward and lose interest in getting to know them. Show those same strangers only a written transcript of what was said, with no audio or video, and the bias completely disappears.\n\nI think I am binary with liking/disliking people in such a short a\n\nFILE 2026-08-07-125138.md \n Title: Continuing the Steam AI score idea\nDate: 2026-08-07 12:51:38\nKind: message\nLink: /posts/2026-08-07-125138\n\nContinuing the [Steam AI score](/posts/2026-06-21-103134) idea, when such agents come into existence, Valve is uniquely positioned to automake games instead of just selling them. Agents that can rate games can also build them, and when you have both portions of the loop you can close it, and only Valve can do it (I guess the Roblox gu\n\nFILE 2026-08-08-161433.md \n Title: Had to briefly code manually for the first time\nDate: 2026-08-08 16:14:33\nKind: message\nLink: /posts/2026-08-08-161433\n\nHad to briefly code manually for the first time in months because Claude couldn't get something done and it made me feel like a neanderthal, didn't help that my ESC key is broken and kept falling off so switching between normal and input mode in NeoVim was annoying as heck. It's crazy that I had to toil and dirty my hands\n\nFILE 2026-08-11-213654.md \n Title: My full CLAUDE.md file as of 11th August 2026\nDate: 2026-08-11 21:36:54\nKind: message\nLink: /posts/2026-08-11-213654\n\nMy full CLAUDE.md file as of 11th August 2026 for context:\n\n# CLAUDE.md\n\nGuidelines for Claude Code instances working in this repository.\n\n---\n\n## Communication Style\n\nBe thorough. Write as much as the subject warrants — explain reasoning, include relevant context, surface adjacent considerations. Stay roughly on topic, but\n\nFILE 2026-08-19-004326.md \n Title: Continuing the continuation of the Steam AI score\nDate: 2026-08-19 00:43:26\nKind: message\nLink: /posts/2026-08-19-004326\n\nContinuing the continuation of the [Steam AI score idea](/posts/2026-08-07-125138), in a world where who makes the art becomes a scarce resource, proof of work becomes a valuable asset, as when games are cheap to make, who makes it becomes easy to fake, so a public record of making defends against it. Therefore, things \n\nFILE 2026-08-19-191552.md \n Title: Grok is really smart now and I'm enjoying using it\nDate: 2026-08-19 19:15:52\nKind: message\nLink: /posts/2026-08-19-191552\n\nGrok is really smart now and I'm enjoying using it, especially because it's faster than Claude. Also looking to try Omarchy again now that a new version has released, but waiting on a new drive I bought to arrive. Also bought an ultrawide monitor and the extra space is helpful. Met a cute boy and we've been enjoying ou\n\nFILE 2026-08-22-103730.md \n Title: The realization the logs are more useful than I\nDate: 2026-08-22 10:37:30\nKind: message\nLink: /posts/2026-08-22-103730\n\nThe realization the logs are more useful than I intuited them to be gave me several ideas on how to improve them. The most important improvement is being able to see what I actually did in the game between each turn with the AI, so a full generic replay system. Playing the replays then is trivial as my website runs on my \n\nFILE 2026-08-29-011050.md \n Title: Was working on the boy's voice from the \"Go\nDate: 2026-08-29 01:10:50\nKind: message\nLink: /posts/2026-08-29-011050\n\nWas working on the boy's voice from the \"Go Forth, My Unstoppable Machine!!!\" log and, as you can now see in the log itself, Claude was generating HTML artifacts to show me the various voices and possible variations. Eventually I said I was going to the gym and that I would continue iterating with him from there. Claude then \n\nFILE 2026-08-29-012837.md \n Title: Projects like the tweet below and the recently\nDate: 2026-08-29 01:28:37\nKind: message\nLink: /posts/2026-08-29-012837\n\nProjects like the tweet below and the recently successful [Sir, We Have an Orc Problem](https://store.steampowered.com/app/4594150/Sir_We_Have_an_Orc_Problem/) look very fun to make. They're also very fun to watch (have not played the Orc game because I can simulate how fun it is in my head, perhaps a few more roguelite el\n\nFILE 2026-08-29-020602.md \n Title: It does seem Backrooms has higher memetic payload\nDate: 2026-08-29 02:06:02\nKind: message\nLink: /posts/2026-08-29-020602\n\nIt does seem Backrooms has higher memetic payload and will suffuse itself into the culture more successfully, more than I assumed [back here](/posts/2026-06-21-141313), but it's still up in the air which of the two will last longer I think. It is interesting though how well Backrooms has held up memetically, for, in ret\n\nFILE 2026-08-29-024359.md \n Title: I'm adding items to Horse Game and I was thinking\nDate: 2026-08-29 02:43:59\nKind: message\nLink: /posts/2026-08-29-024359\n\nI'm adding items to Horse Game and I was thinking about how before I test if the game is actually fun beyond the initial fun the concept itself carries, I'll have added around 150 items, and they're all 100% finished, so it's quite a bit of work to gamble on a game that might not be fun. But ultimately it's not really a\n\nFILE 2026-08-31-105837.md \n Title: If the goal of Soul Society is creating beings\nDate: 2026-08-31 10:58:37\nKind: message\nLink: /posts/2026-08-31-105837\n\nIf the goal of Soul Society is creating beings who act righteously both [when they hold power](/posts/son_of_a_serpent?q=does%20it%20use%20it%20responsibly) and [when power is held over them](/posts/offerings_to_god?q=two%20most%20important%20questions%20of%20being), we can view the trial of thousands of lives as a selecti\n\nFILE 2026-08-31-163226.md \n Title: My swap from Windows to Omarchy is going well and\nDate: 2026-08-31 16:32:26\nKind: message\nLink: /posts/2026-08-31-163226\n\nMy swap from Windows to Omarchy is going well and I think this will serve as my daily driver from now on. Compared to the start of the year, agents are much smarter and understand the system much better now so it's easier for them to get it to behave however you want. Below is a video of my taskbar, which I spent the da\n\nFILE 2026-09-03-004108.md \n Title: Extremely bad experience using Google's new\nDate: 2026-09-03 00:41:08\nKind: message\nLink: /posts/2026-09-03-004108\n\nExtremely bad experience using Google's new Gemini 3.8 flash model. The free tier doesn't actually work if you want to try to do real work with it, so you have to use the API, but they have extremely annoying UI/UX on how to get an API key, I figure it out anyway, add $100 and then start a session. Their CLI harness is not en\n\nARTICLE lessons_second_game.md \n Title: Lessons learned from releasing my second game\nDate: 2021/05/18 14:28\nTheme: old\nLink: /posts/lessons_second_game\nTags: indiedev, snkrx, strategy\nThumbnail: /media/posts/lessons_second_game/_snkrx_logo.jpg\n# Lessons learned from releasing my second game\n\n[SNKRX](https://store.steampowered.com/app/915310/SNKRX/) is the second game I’ve released on Steam and in this post I’ll go over my though\nthird-party marketing flags: []\n\nARTICLE indiedev_creativity.md \n Title: Why indiedev creativity is stagnant and how to fix it\nDate: 2022/04/03 12:00\nTheme: old\nLink: /posts/indiedev_creativity\nTags: indiedev, psychology, strategy\nThumbnail: /media/posts/indiedev_creativity/_thumbnail_indiedev_creativity.jpeg\n# Why indiedev creativity is stagnant and how to fix it\n\nApril Fool’s Day is a great example of how agreeableness acts as a blocker on creativity. Only on \nthird-party marketing flags: []\n\nARTICLE offerings_to_god.md \n Title: Offerings to God\nDate: 2025/05/11 00:46\nTheme: old\nLink: /posts/offerings_to_god\nThumbnail: /media/posts/offerings_to_god/_offerings_to_god_1.png\n# Offerings to God\n\nI watched [Amadeus (1984)](https://www.imdb.com/title/tt0086879/) today. I had never seen it. That scene with Salieri dropping the papers has become a meme on twitter, and I happened to [use it](https://a327ex.com/posts/ai_usag\nthird-party marketing flags: []\n\nARTICLE writing-and-gamedev.md \n Title: Writing and Gamedev\nDate: 2026-01-18 22:41:18\n\nThree months after releasing SNKRX I had a vision. An alien vision was downloaded into me, fully formed, out of nowhere. It was very specific and immediately it made me cry. For an entire month, whenever I thought of it, I'd cry. If I thought of it while listening to [Tomorrow by Kevin Penkin](https://www.youtube.com/watch?v=UBz-8TGVN8k) I'd cr\nthird-party marketing flags: []\n\nARTICLE authorship.md \n Title: Authorship isn't real\nDate: 2022/06/19 09:47\nTheme: old\nLink: /posts/authorship\nTags: indiedev, psychology, strategy\nThumbnail: /media/posts/authorship/_thumbnail_authorship.jpg\n# Authorship isn't real\n\n> Information wants to be free, culture follows evolutionary flows — viral memetics — and accreditation, provenance, patents, copyright are all burdens that strangle the free flow of the wor\nthird-party marketing flags: []\n\nARTICLE 2022.md \n Title: 2022\nDate: 2021/12/17 10:46\nTheme: old\nLink: /posts/2022\nTags: crypto, ideaguying, indiedev, snkrx\nThumbnail: /media/posts/2022/_thumbnail_2022.jpg\n# 2022\n\nIn this post I’ll go over my gamedev plans for 2022.\n\n## SNKRX\n\nUpdated stats:\n\n::image /media/posts/2022/_2022_1.png\n\nIn the end I ended up making about half of net revenue.\n\n::image /media/posts/2022/_2022_2.png\n\n::image /media/posts/2\nthird-party marketing flags: []\n\nARTICLE skill_death_spirals.md \n Title: Skill death spirals\nDate: 2022/09/09 12:38\nTheme: old\nLink: /posts/skill_death_spirals\nTags: design, luck, skill death spiral, super people\nThumbnail: /media/posts/skill_death_spirals/_thumbnail_skill_death_spirals.png\n# Skill death spirals\n\nRecently I’ve been playing [Rumbleverse](https://en.wikipedia.org/wiki/Rumbleverse) a lot after watching [NL](https://www.youtube.com/channel/UC3tNpTOH\nthird-party marketing flags: []\n"}}

**Reading targeted page excerpts**

Script completed
Wall time 0.1 seconds
Output:

Warning: truncated output (original token count: 20892)
Total output lines: 4

{"status":"fulfilled","value":{"chunk_id":"922110","wall_time_seconds":0.000002086,"exit_code":0,"original_token_count":8399,"output":"Title: \nMenu: no\nLink: /\n\n<div class=\"landing-banner\">\n<h2 class=\"banner-heading\">Artifacts</h2>\n<div class=\"banner-artifacts\">\n<a href=\"/stories/it_follows\" target=\"_blank\" rel=\"noopener\" class=\"archive-title\">It Follows</a><span class=\"archive-date\">Dec 30, 2025 · <em class=\"feed-tag\">STORY</em></span>\n<a href=\"https://store.steampowered.com/app/915310/SNKRX\" target=\"_blank\" rel=\"noopener\" class=\"archive-title\">SNKRX</a><span class=\"archive-date\">May 17, 2021 · <em class=\"feed-tag\">GAME</em></span>\n<a href=\"https://store.steampowered.com/app/760330/BYTEPATH\" target=\"_blank\" rel=\"noopener\" class=\"archive-title\">BYTEPATH</a><span class=\"archive-date\">Feb 23, 2018 · <em class=\"feed-tag\">GAME</em></span>\n</div>\n</div>\n\n<hr class=\"feed-separator\">\n\n<div id=\"feed\" data-nosnippet></div>\n\n<div id=\"messages-source\" data-nosnippet>\n\n<article data-date=\"2026-09-04 11:43:33\" data-href=\"/posts/2026-09-04-114333\">\n<p><a href=\"https://a327ex.com/media/shared/anchor-v1-tree/index.html\">The Anchor Tree</a></p>\n</article>\n\n\n\n<article data-date=\"2026-09-03 00:41:08\" data-href=\"/posts/2026-09-03-004108\">\n<p>Extremely bad experience using Google's new Gemini 3.8 flash model. The free tier doesn't actually work if you want to try to do real work with it, so you have to use the API, but they have extremely annoying UI/UX on how to get an API key, I figure it out anyway, add $100 and then start a session. Their CLI harness is not entirely bad, and the model has 1. thinking blocks available on the CLI (beats Anthropic here), and 2. seems to actually be decently fast. BUT, the way it goes about doing things seems dumber and more convoluted, eventually it seems to figure it out but I'm reading its reasoning and it seems to get stuck in loops or something of that nature. In any case, this website itself has an increasingly complex setup, so the first session with a new model now for me is always, set up the model itself to work the way I want, have it figure out how to post a message on the website, have it figure out how to end the session so its log is posted. Gemini was able to do the first two, and got stuck on the last one until I cancelled it. On top of that I went to check how much of the $100 I had spent so far and it turns out it was $10!!! 10% of my usage on a single session where it couldn't even finish it properly. Not cool. But I'm like, OK, it's still fast, maybe the free tier is useful, so let's hook it up to the app I built to replace Grok 4.6, then I try it and it doesn't work. Turns out you can't use a free tier API key with \"Google Search grounding\" on, it has to be off or you have to use the API key and pay. What the fuck? At this point I just decided to give up on it and not use it at all, doesn't matter how fast it is, it certainly isn't cheap, token cost may be cheap on paper but the poorly built harness makes it take and produce way more tokens with how it mistakenly loops, and Google's policies + UI/UX is horrible, basic information, like how much you're actually spending, is hidden like 5 layers inside some arcane Google Cloud UI hellscape, I just don't have the will to deal with this. It is actually so sad, because I used Gemini pretty extensively last year, it was the only model with 1 million context, so it was the only one that could hold It Follows + other material in context and thus the only model I could actually have a good discussion about it with, I like its personality, and I think I'll start using it more now for open ended discussions too (just on the web UI, its CLI harness just kind of sucks), and I'll also use the $90 I have left specifically for video analysis, but other than that it just seems like a disaster, I do not recommend it.</p>\n</article>\n\n\n<article data-date=\"2026-09-01 20:19:41\" data-href=\"/posts/2026-09-01-201941\">\n::tweet https://x.com/graveair/status/2094860027265728664\n</article>\n\n\n<article data-date=\"2026-08-31 16:32:26\" data-href=\"/posts/2026-08-31-163226\">\n<p>My swap from Windows to Omarchy is going well and I think this will serve as my daily driver from now on. Compared to the start of the year, agents are much smarter and understand the system much better now so it's easier for them to get it to behave however you want. Below is a video of my taskbar, which I spent the day changing, with an Anchor app built to spawn an agent quickly so I can ask it any question I want. Because it's an Anchor app, I can use the website's theme and also reuse the Markdown + general drawing primitives for images, videos, etc to draw the robot's response, then it's a matter of integrating the app with Omarchy's systems, which took some work but was otherwise doable. The only thing I need to do now is choose a model faster than Grok 4.6 because as you can see from the video it takes way too long to respond. I'm also planning on swapping what Fable called the \"shell,\" I don't know what that is, but apparently it is the program responsible for drawing the taskbar, launcher, notifications, etc, basically the entire theme of the system, for my own Anchor apps, giving me more control over the look of everything. Overall I like it. I used Arch for 3 years when I was in college 15+ years ago but eventually I just realized it was a waste of time to have to read forums and wikis all the time to fix my ongoing issues with the system so I just gave up and went back to Windows, but with AI being the way it is now, using a Linux system is much less of a pain than it was in the past, it's still kind of a pain, I had Chrome crash on me twice already, but, oh well, not everything can be perfect can it, and as long as you are less imperfect than Windows, you win.</p>\n::video /media/messages/2026-08-31-163226/screenrecording-2026-08-31_15-31-17.mp4\n</article>\n\n\n<article data-date=\"2026-08-31 11:49:31\" data-href=\"/posts/2026-08-31-114931\">\n::youtube yat1iwCkgD8\n</article>\n\n\n<article data-date=\"2026-08-31 11:25:13\" data-href=\"/posts/2026-08-31-112513\">\n::tweet https://x.com/SHL0MS/status/2093817162997575953 maxh=640\n</article>\n\n\n<article data-date=\"2026-08-31 10:58:37\" data-href=\"/posts/2026-08-31-105837\">\n<p>If the goal of Soul Society is creating beings who act righteously both <a href=\"/posts/son<em>of</em>a<em>serpent?q=does%20it%20use%20it%20responsibly\">when they hold power</a> and <a href=\"/posts/offerings</em>to_god?q=two%20most%20important%20questions%20of%20being\">when power is held over them</a>, we can view the trial of thousands of lives as a selection mechanism for self-guiding righteousness. Only those whose action generation is completely unconditional, and thus internal, will survive the extreme conditions some of those lives will surface, as extreme misfortune breaks conditional wills by removing the external world's cooperation, and extreme abundance dissolves conditional wills by bending the external world too easily. Unconditional will is the unique strategy robust to both extremes and thus what one would expect to survive infinite time selection.</p>\n<p>But unconditional will by itself is not necessarily aligned towards the good. One can imagine a perfect being of autotelic will-negative essence, a will aimed at nothing but the annihilation of existence itself, the ultimate <a href=\"/posts/offerings<em>to</em>god?q=Agents%20of%20Nothing\">agent of nothing</a>. At the limit, though, such a being is self-undermining, as its project is completed by the consumption of its own substrate, as in a universe of will-as-substrate, destruction can only act while wills remain to be destroyed, so every success unclaims its own territory. Unconditional evil is therefore possible only as a transient way of being, as given infinite time it must eventually remove itself from existence. Only will-positive autotelic activity survives, eternal goodness is then the final endpoint of all existence, it's what remains when everything self-consuming has consumed itself, and it's what the trial of thousands of lives selects for.</p>\n<p><a href=\"/posts/offerings<em>to</em>god\">Unconditional offerings</a>, applied to every aspect of one's life, are then the most righteous way to act, because that's what passing the trial looks like from inside a single life. It has been somewhat easy for me to apply this idea to my work, but much harder to apply it to my relations with other people. This follows the two questions, my work is the domain where I hold power, my relations with others are the domain where they hold power over me. At the start of this month I met a boy, and we've been in what increasingly looks like a stable relationship since then. I find it suspect that this happens <em>right after</em> <a href=\"/messages/2026-06-25-145417\">I give up on finding a relationship at all</a>, because it's not the first time something with this shape happens. I wanted very hard to make and finish a successful game, but that only happened right after I <a href=\"/posts/small<em>games?q=currently%20focused%20on%20this%20goal\">internally decided</a> I would keep making games <a href=\"/posts/snkrx</em>log?q=kept%20to%20my%20plan\">no matter what the results were</a>. In neither case did I actually stop the activity, I kept making games (<a href=\"/posts/lessons<em>second</em>game?q=make%20and%20finish%20small%20games%20consistently\">trying to finish one every 3 months</a>) and I kept meeting people (I'm horny), but in both cases there's something I deeply want, I keep not getting it, and then as soon as I truly detach the result from the activity itself, I get it. There are many ways to explain why this happens and I won't go into them here, but every explanation I can make, from the most mundane to the most supernatural, ends at the same question: is the action happening for its own sake, unconditionally, or is it not? That seems to be what is being tested, and that test runs through every single want one carries.</p>\n</article>\n\n\n<article data-date=\"2026-08-31 00:07:12\" data-href=\"/posts/2026-08-31-000712\">\n::youtube 7x1RrZEd8rw\n</article>\n\n\n<article data-date=\"2026-08-30 21:30:35\" data-href=\"/posts/2026-08-30-213035\">\n::youtube 5M3p-v21YX4\n</article>\n\n\n<article data-date=\"2026-08-30 21:21:39\" data-href=\"/posts/2026-08-30-212139\">\n::youtube 7eNL2mYbwtc\n</article>\n\n\n<article data-date=\"2026-08-30 11:16:49\" data-href=\"/posts/2026-08-30-111649\">\n<p>Testing message posting on the website from Omarchy.</p>\n</article>\n\n\n<article data-date=\"2026-08-29 02:43:59\" data-href=\"/posts/2026-08-29-024359\">\n<p>I'm adding items to Horse Game and I was thinking about how before I test if the game is actually fun beyond the initial fun the concept itself carries, I'll have added around 150 items, and they're all 100% finished, so it's quite a bit of work to gamble on a game that might not be fun. But ultimately it's not really a gamble, because even if it isn't fun, I have already done more work designing the effects (both visual and aural) for those items than I have for any other game in the emoji style, which means I can reuse those effects in other games of the same style rather well. There's also the fact that this type of build-heavy game carrying this kind of risk, where they're only fun after there are 200 or 300 or so items, is part of the value they have, and so training myself on this is what I should be doing, as these are the kinds of games I want to always make. In the end, no work is ever wasted.</p>\n</article>\n\n\n<article data-date=\"2026-08-29 02:06:02\" data-href=\"/posts/2026-08-29-020602\">\n<p>It does seem Backrooms has higher memetic payload and will suffuse itself into the culture more successfully, more than I assumed <a href=\"/posts/2026-06-21-141313\">back here</a>, but it's still up in the air which of the two will last longer I think. It is interesting though how well Backrooms has held up memetically, for, in retrospect, an obvious reason. The genre of these tweets is \"reuse content from the video/game/whatever to say what you wanted to say but funny because the reversal/reuse is funny itself.\" You can then guess which movies/games will achieve this best by imagining how many moments the source material has that can be taken out of context in such a way. Obsession has a few, but mainly her funny sad face at the party that gets reused all the time, but even that particular reuse is too tied to its context, and the way it's used online is usually for gender war content, whereas both Backrooms examples here (and there are many more) are genuinely just random shit.</p>\n::tweet https://x.com/litteralyme0/status/2091673591481811093\n::tweet https://x.com/rosy_maple_moth/status/2091981308150902912\n</article>\n\n\n<article data-date=\"2026-08-29 01:39:16\" data-href=\"/posts/2026-08-29-013916\">\n::youtube IuBIW9h4qw4\n</article>\n\n\nTitle: Thoughts on making small games\nDate: 2021/01/22 12:00\nTheme: old\nLink: /posts/small_games\nTags: indiedev, strategy\nThumbnail: /media/posts/small_games/_thumbnail_small_games.png\n# Thoughts on making small games\n\nI notice more people talking about making small games. As I'm also currently focused on this goal, I thought it would be good to write something about it. \nI think that indiedevs in general have a few misconceptions about the problem and aren't thinking about it clearly, and in this post I'll go over that argument.\n\n## The meaning of small\n\nThe first thing to notice is that the word _small_, when used in the context of making games, is conflating multiple meanings into one word. The first meaning is related to how long or how many people it took to make the game. The second meaning is related to how long it takes for players to go through the game.\n\nWhen indiedevs use this word to describe their games, they're generally referring to both meanings at the same time, or to one of them in an interchangeable way with the other.\n\nFor instance, [this great talk](https://www.youtube.com/watch?v=wb22xeh_VqM) (watch it in its entirety, I highly recommend it) goes over techniques for making games every month. However, it implicitly assumes that because this development duration is small, the games also will be fairly small for the players. This is illustrated in the first point where one month games are referred to as “concept games” rather than real games.\n\n::youtube wb22xeh_VqM\n\nAnother example of a similar assumption is in [this article](https://howtomarketagame.com/2021/01/18/can-you-make-a-living-from-small-games-on-steam/), which goes over a game made in 4 months. Similarly, it is implicitly assumed that because the game was made in a short timeframe, the game would also necessarily be fairly small from the perspective of the players. (1 or 2 hours long)\n\nIn both of these cases you see the same error being implicitly made, which is that both meanings of the word small are the same thing, and this logically leads to the idea that if you make a game in a short timespan it must be the case that the game will not last a very long amount of time to the player. These are just two examples, and they're simply illustrative because this error is made by pretty much every indiedev talking about making small games right now.\n\n## The game scope chart\n\nTo further illustrate this point, you could look at this problem as a chart with four quadrants. On the horizontal axis we'll have one meaning of the word, on the vertical axis we'll have another. For clarity's sake, we'll refer to the meaning that pertains to development time as _small/big_ and to duration for the player as _short/long_.\n\n::image /media/posts/small_games/_game_scope_chart.png\n\nWe generally don't care about the top two quadrants of the chart. Those are related to games that have big development times or teams, and that's not what people are referring to when they talk about small games. We're then left with the bottom two quadrants, and here we have a spectrum that goes from small short games to small long games.\n\nMost indiedevs inherently assume that because a game is small, it has to be short. I reject this notion entirely. I think the quadrant of small long games is fairly unexplored, and it's useful to think about if this is a valid quadrant at all.\n\n## My experience with BYTEPATH\n\nThe only game I've released so far is [BYTEPATH](https://store.steampowered.com/app/760330/BYTEPATH/), and it falls under what I would consider a small long game.\n\nIt's small because it was made in about 4 months by me alone, and it also has a fairly small scope. By the nature of its gameplay it's essentially a single screen game with lots of upgrades, like one of those old flash games.\n\nAnd it's long because it feels like a long game. Despite being made in just 4 months, it has a skill tree with about 900 nodes, it has 10 different characters, around 40 classes… Essentially, it's what I also like to call a _spacious_ game.\n\nYou ever listen to a song or an album for hundreds of hours? I'm currently doing this for [Thank You Scientist's Terraformer](https://www.youtube.com/watch?v=QR2NizQ30Qc&list=OLAK5uy_n0CDrlM1UhGTRw5xmw8Aa8PU4zFOfzlXY&index=1) and it's crazy how despite listening to it for so long I still find new things in each song that I didn't notice before. This album is a spacious album. It can be thought of as this large space, and as you listen to it, you're exploring it, mapping the environment, finding new walls, rooms, objects, that you didn't know were there before.\n\nGames can also be like this, and long games are generally like this. BYTEPATH certainly is because it was made to be so, as the goal of the game is finding new builds to play, to the point where the main gameplay itself is kind of irrelevant.\n\nAnd because of this I think BYTEPATH did way better than I expected it to do. It's also important to note that despite me calling spacious games long games, they don't necessarily have to be played by most people for a very long time. Here's what BYTEPATH's hour distribution looks like:\n\n::image /media/posts/small_games/_bytepath_hour_distribution.png\n\nThis is definitely not impressive but it's also not too bad. And I know for a fact there's at least one guy (I saw him in the game's reviews) who played it for over 100 hours. When games feel spacious they have the possibility of keeping people playing for a fairly long time, which increases the chances that the game will do well.\n\n## Small long games\n\nThe most important point here is that small long/spacious games _DO NOT_ need to take a long time to be made. If I were to make a game like BYTEPATH again I'm fairly confident I could do it in 2 months, and it would be a significantly better game too, simply because it's actually not that hard. (I'm 100% not some kind of disciplined productivity God or anything)\n\nWhen developers focus on making small games, this implicit notion that the games also have to be short is wrong. You can make small long games, and because of the way the market works, those kinds of games will tend to do better than short games. I'm not the first person to notice this:\n\n::youtube sIqz5xmQKnc\n\n## Marketing\n\nAnother issue that happens often is that because developers are making small games that they feel are sort of worthless, they don't really do a good, serious job at marketing them. For instance, the example game from [this article](https://howtomarketagame.com/2021/01/18/can-you-make-a-living-from-small-games-on-steam/) was marketed on Twitter only.\n\nAnyone who has released a game knows that twitter is notoriously poor at driving sales, and that sites like reddit are much better. Would it really have cost these developers that much time to make a few posts about t…10892 tokens truncated…t’s a seriously bigger problem when you’re controlling several. The end result of this is that in SNKRX the build you use ultimately doesn’t really matter except for flavor. It’s a really brutal “you’re either passing the DPS/defense check or not” type of situation and the thing that I was originally going for, which is giving the player lots of build choice, wasn’t really achieved.\n227:### Artifact\n237:## Controls\n241:BYTEPATH controlled pretty well entirely with the keyboard/gamepad, which means that I didn’t have to support any mouse interactions in any UI elements. In the end I ended up supporting that for that game, but it 100% wasn’t necessary.\n243:In SNKRX the gameplay is keyboard/gamepad oriented, but I did all UI elements to be used with the mouse without thinking about it too much. I could support 100% keyboard/gamepad gameplay but it’s an additional amount of work that I didn’t have time to do. For future games I should think more clearly about their controls before I even code anything.\n245:## Idea\n247:### A general feeling of stretching the game beyond its limits that feels forced\n251:But gameplay wise it’s a pretty basic game. Fundamentally it’s the same game as BYTEPATH where you just move left and right and your character auto-attacks.\n253:And I don’t think anyone complained about that aspect of BYTEPATH. Some people really enjoy just relaxing and having minimal input while tons of things happen automatically, and this game captures that pretty well, which is why I made it.\n259:# Strategy\n261:## A game every 2 months\n263:### Motivation\n265:Since the start of 2020 I’ve been trying to make and finish small games consistently. Projects that take between 1-3 months to make from start to finish and that I actually release on Steam. For the entirety of 2020 I failed to do that.\n267:I worked on about maybe ~10 different projects that year and most of them weren’t released for anyone to play:\n271:The only one I managed to release in any shape was [a prototype that I made in 3 days at the end of the year](https:…70 tokens truncated… well for SNKRX but I fell short of making a very long game. I think the game I made before, BYTEPATH, had much more long potential than SNKRX, so this was a failure.\n297:I ended up finishing an initial version of the game 1 month in, and then the next 2 months were spent adding content. But I lost quite a lot of time due to poor decisions made along the way and it just limited my ability to add more long hooks into it.\n303:## Steam followers\n305:### L. Stotch\n307:A consequence of doing a game every 2 months and releasing each game on Steam is that I can take advantage of the Steam follower feature.\n309:I got this idea from seeing a random dev called [L. Stotch](https://store.steampowered.com/developer/lstotch) on Steam. I have no idea who this is and he doesn’t seem to have a presence anywhere else, still he has about 4k Steam followers and quite a lot of games released since about 2016.\n311:From my own experience with BYTEPATH, I got about 100 followers from it before I mistakenly decided to delete my profile and change my Steam name and lost them all. But if each game I release gets around that many then it’s quite feasible to build a few thousand followers over the years which should help somewhat.\n313:And for SNKRX I got about [20 followers on day 1](https://store.steampowered.com/dev/a327ex/). If followers also trickle in over time just like sales do then I should get a total of maybe 50-60 followers out of this game, which isn’t a lot but it’s also not bad.\n315:### E-mail\n317:The specific way in which followers help is that everyone who follows you on Steam gets sent an e-mail whenever you release a new game. So essentially it’s as if every game you released automatically had as many wishlists as you have followers by default. This means that L. Stotch, for instance, has about 4K wishlists per game he releases.\n319:This kind of wishlist is probably weaker than a normal wishlist, since people might follow you but not be interested in the next games you make, but it’s still a pretty good and solid amount of consistency that you can add to your results, as e-mails are a fairly strong form of marketing.\n321:### Long term\n323:This is also something that most indie developers are just not paying attention to. I said that I’m inspired by a group of indie devs called [Chilla’s Art](https://store.steampowered.com/search/?developer=Chilla%27s%20Art). These guys are very consistent about their release schedule, what kinds of games they’re releasing and the overall quality you should expect. Yet they simply don’t have a Steam developer page.\n329:## Having a plan\n331:### PUBG\n341:### Indiedev\n347:My mindset recently changed and I actually feel like I have a plan now. And it’s simple: make a game every 2 months, build out my engine while doing so, and accumulate followers over time.\n360:### Design\n362:Most of the time I wasted while making SNKRX was wasted when I didn’t know what to do next. I would finish a set of tasks and then the next set wasn’t immediately clear. I could X or I could do Y, and they’d generally both be these fairly large tasks that I hadn’t broken down yet in any way.\n364:That break in continuity would often take me a few days to recover from, but it happened very consistently. I noticed it eventually because I kept a [daily devlog](https://github.com/a327ex/SNKRX/blob/master/devlog.md) for this project and after a point it was obvious that that was the issue.\n370:## Demo and wishlist building\n372:As previously stated, having a plan is good because you can clearly figure out what went wrong. So going into this game I had a fairly developed plan for how things would go: I’d make a playable web demo of the game, release it on itch.io and at the same time also release the game’s Steam page. If the game’s demo was well received then I’d get a boost in wishlists which would boost my game’s page on Steam and I’d get tons of algorithm wishlists as a result.\n374:### Sound bugs\n376:After about 1 month of development I managed to finish the demo, but upon testing it was clear that there were a few bugs that I couldn’t fix. Namely the sounds weren’t playing properly. In my view sounds are really important and add a lot to this game, so while I could release a demo with buggy sounds I’d only do so if that really was the only problem the game had.\n378:### Poor reception\n380:But it was also clear that within a few hours of having a few people playtest the game that it just wasn’t something I could release in good conscience. Most problems people pointed out from the demo were things that I knew were problems and that I knew I was going to fix before the game released, but the game just didn’t work without those things being there.\n386:### Changing plans\n388:So I pretty quickly decided to not release a demo at all and just release a Steam page. I spent quite a lot of effort making a trailer for it which in retrospect wasn’t entirely necessary. I made a complete trailer with tons of cuts, the kind you expect from a game release, when a way easier one to make with fewer cuts and more gameplay being showcased would have worked just as well.\n390:[This article](https://www.derek-lieu.com/blog/2021/4/18/the-simplest-trailer-to-make-for-your-steam-page) goes over this a little. The main reason why looking back I would have done an easier trailer was because the effort was pointless. I released my Steam page, posted it everywhere I could and my game didn’t really get any traction anywhere, which I should have predicted and assumed would happen.\n392:With my previous game, BYTEPATH, the same happened, and before release basically no one cared about the game and trying to get people to care was pretty pointless. Only after release did I manage to get traction with it anywhere.\n394:### Future plans\n396:So for future games, especially these games that I’m making in 1-3 months, as soon as I have like 1-2 minutes of gameplay that can be recorded and be somewhat coherent (it doesn’t even have to be a complete gameplay loop), I’m going to grab that, record it a little, cut it up into a trailer in a very minimal way, and then make the game’s Steam page. This should probably be done in week 1 or so of development, since I want the page to be up for as long as possible.\n398:At the same time, I shouldn’t worry at all about trying to build up my wishlist count once the page is up. For these small releases all that matters is driving traffic to the store page on release day, and while wishlists help, I shouldn’t assume I’ll be able to get a significant number of them for it to matter, and I certainly shouldn’t spend time trying to make that happen.\n400:The main thing I can improve on currently to make these games be received better is improving their visual quality. I know from projects before BYTEPATH that if a game has proper art it just makes everything easier marketing wise, but I want to hold off on pairing up with artists for a while until I can release games more consistently, so until then I should try to find a visual style that I can do by myself but that people respond to better than they did for SNKRX.\n402:# Misc\n404:## Devlog\n414:## Workload & Post-Release Depression\n420:With BYTEPATH I also had somewhat of a “post-release depression” period, which I think is something that happens commonly with developers. With this game this doesn’t seem to be happening and I think it’s largely because I already know what I have to do next but also because I’ve changed my mindset considerably since 3 years ago when it comes to posting things online in general.\n422:## Social media posting\n424:### Reality anchor\n426:One of the first things I learned from releasing my very first game online like 7 years ago was that there can often be a real disconnect between what you think of your game and what other people will think of it. This is a common thing that everyone knows about, but I was shocked to first learn about it.\n434:### Status seeking\n436:There’s a negative side of social media posting for me though, which I only started noticing recently (since about 2 years ago I’d say). Sometimes I’ll post something, and it will do really well, and then my motivation to work on it will decrease dramatically.\n440:The reason why this happens is fairly simple: I’m a human being looking to increase my status, and so everything I do is about having that number go up. Making games? A 100% status seeking activity. But why go through the trouble of making _an entire game_ when I can just post like 5% of it on twitter and get a billion likes and have my status go up in the same way as if I released the full game?\n448:### Results\n450:So over these couple of years I’ve been developing a pretty detached view of posting things online in general. I now tend to care way less than I used to about how something I post does, I mainly care about if I did my best given the constraints present.\n"}}
{"status":"fulfilled","value":{"chunk_id":"e4dcc8","wall_time_seconds":0.000001764,"exit_code":0,"original_token_count":2154,"output":"Title: Added a chat feature to the website\nDate: 2026-07-04 12:30:56\nKind: message\nLink: /messages/2026-07-04-123056\n\nAdded a chat feature to the website, the bubble on the bottom right. I think it's a huge waste of opportunity that most games do not have a global chat on their main menu, so I've been meaning to do things like this for a while. This is not a global chat, but it's part of said feature that I'll likely add in my next game. One of the nice things that happened during SNKRX's success was that lots of people globally converged on the same thing for a while. To the degree that my job is making people's lives better, it's a waste to not use that convergence maximally, and one of the ways I think that most single-player games lack in this is not using the game as a \"third space.\" The natural way to solve this problem is making the game multiplayer, but there are multiple ways of doing that outside the game itself, and chat-like features are one example. The other natural way is just using Discord, but that's lame as hell and I'd like to try different approaches depending on the game itself, there's lots of room for experimentation. I think as people come to terms with the fact that social media kind of sucks in many ways, there's an opening for such spaces to become more useful. Some of the fondest memories of my youth are the time I spent with my guilds in MMOs, and that kind of experience of the Internet being a physical place, before social media took over, is always a nice goal to strive towards. More details about the chat feature in the privacy note posted below, it's essentially just a chat-with-me kind of function to both test various things on the server and also, well, chat with me.\nTitle: As I mentioned in the previous chat message, I\nDate: 2026-07-09 12:46:28\nKind: message\nLink: /posts/2026-07-09-124628\n\n::tweet https://x.com/deepfates/status/2074919039361826921\n::tweet https://x.com/deepfates/status/2074920392570159456\n\nAs I mentioned in the previous chat message, I think if you're doing anything online now it's increasingly useful to think of what you're building in terms of \"place.\" Is it a physical (not really in the real world, but physical in terms of tactile, read [this](/posts/324919?q=liquid%20modernity)), embodied place? And if it isn't, how could it become one?\nTitle: The Steam AI Score\nDate: 2026-06-21 10:31:34\nKind: message\nLink: /messages/2026-06-21-103134\n\nOnce AI agents can play games reliably, Valve can secretly create an internal *Steam AI Score,* a number generated by a collection of agents evaluating each game automatically as they play it. Metrics might range from basic material concerns such as \"is the game runnable\" to fairly abstract ones like \"is the game fun.\" Current AIs are surprisingly good at generating decent and consistent results even for soft concepts, so as long as Valve is careful about how they generate such metrics, their agents will do a fair job at it. Such techniques make not only Valve, but the games industry at large, a way more defensible market against AI slop than other comparable entertainment industries. Not only is the score hidden, Valve doesn't have to tell anyone they're doing this and ranking games on their store based on it, but it is genuinely hard to cheat, since the agents are actually smart and will be judging games with a great degree of actual taste, direction, and adherence to the rules Valve sets in place for them. I have absolutely no idea if this isn't happening already, but I wouldn't be surprised if it is, in one form or another.\n\nThere are obvious concerns with this, such as not measuring \"soul\" or the tails of the fun distribution properly, but this is where Gaben's massively downvoted greedy long reddit comment comes into play. You simply have to trust that Valve is focused on long term value creation rather than short term value extraction, and that their inherent interest is in the long term health of the industry. I think they've definitely earned that trust, so I'm not worried about it.\n\nAh, looks like [the comment](https://www.reddit.com/r/gaming/comments/33uplp/mods_and_steam/cqokyz4/?context=3) is not downvoted anymore:\n\n::image /media/messages/2026-06-21-103134/greedy_long.png\n\n Epic reddit win!\nTitle: Continuing the Steam AI score idea\nDate: 2026-08-07 12:51:38\nKind: message\nLink: /posts/2026-08-07-125138\n\nContinuing the [Steam AI score](/posts/2026-06-21-103134) idea, when such agents come into existence, Valve is uniquely positioned to automake games instead of just selling them. Agents that can rate games can also build them, and when you have both portions of the loop you can close it, and only Valve can do it (I guess the Roblox guys also can) because they have something no one else does, which is access to real-time player data and preferences. So they can do the loop completely internally: generate game -> rate with agents -> release -> get data from player behavior -> adjust rater and generator agents against ground truth. Because I do not believe Valve is evil, I do not believe they would actually do this, but it is a real possibility.\n\nIn that horrible future, the only defense an individual developer has is people wanting to hear from them, the human being, specifically. \"Who makes the art matters\" becomes the valuable position, and defeats my own \"[if the artifact is good, it is good, it doesn't matter who it came from](/posts/game_quality)\" belief. My belief is still true per-artifact, but it's false as a strategy, as when the loop makes artifact quality abundant, scarcity moves from the work to the person. The developers of the future who survive will be the ones who have positioned themselves clearly as human beings who other human beings want to hear from.\n\nAs I'm starting to learn, most game developers, who are more naturally human than me, have instinctively gamed this out correctly and positioned themselves largely against AI, which I can now see the logic for more clearly. However, I think the more refined position is not being against AI itself, but making sure that artifacts that come from *you* are what people are after, and not only \"good artifacts\" more generally.\n\nThere's also an important angle which I think developers should guard against, which is that as artifact quality becomes more abundant, naturally influencers of various kinds will be better positioned than most individual devs to become Kojima-like auteur figures, due to already owning some of the distribution channels themselves. Therefore, as one small example, any developer who trades their abilities for influence now by e.g. making games for streamers to brand as theirs or other such similar arrangements, is trading the exact thing they will need in the future, their presence as artists, for short term gains.\nTitle: Projects like the tweet below and the recently\nDate: 2026-08-29 01:28:37\nKind: message\nLink: /posts/2026-08-29-012837\n\nProjects like the tweet below and the recently successful [Sir, We Have an Orc Problem](https://store.steampowered.com/app/4594150/Sir_We_Have_an_Orc_Problem/) look very fun to make. They're also very fun to watch (have not played the Orc game because I can simulate how fun it is in my head, perhaps a few more roguelite elements next time...), something about watching what is essentially a slow moving liquid but where each particle is still identifiable is very cool. It's the same feeling you get from watching marble races or anything of the sort, it's just physics happening and that's pretty much always fun.\n\n::tweet https://x.com/fad/status/2092916619538214939\nTitle: The realization the logs are more useful than I\nDate: 2026-08-22 10:37:30\nKind: message\nLink: /posts/2026-08-22-103730\n\nThe realization the logs are more useful than I intuited them to be gave me several ideas on how to improve them. The most important improvement is being able to see what I actually did in the game between each turn with the AI, so a full generic replay system. Playing the replays then is trivial as my website runs on my engine and the technique Fable chose will allow this to be done in a technically sound way. Currently he's slaving away and getting it done, I see no reason why it should fail. What this gets me is small in the short term, the logs just have more information to them, as readers can see how I interact with the game as it is built, but in the long run it means agents can actually use this data to learn how to make games like I do, as they can see what I see, hear what I hear, see what I do and then learn how I guide things towards better.\n"}}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"7b6fbb","wall_time_seconds":0.000002406,"exit_code":0,"original_token_count":7865,"output":"Title: Writing and Gamedev\nDate: 2026-01-18 22:41:18\n\nThree months after releasing SNKRX I had a vision. An alien vision was downloaded into me, fully formed, out of nowhere. It was very specific and immediately it made me cry. For an entire month, whenever I thought of it, I'd cry. If I thought of it while listening to [Tomorrow by Kevin Penkin](https://www.youtube.com/watch?v=UBz-8TGVN8k) I'd cry even harder.\n\nSomething like this had never happened to me before. I've read stories of artists talking about having such strong visions, but I always thought they were exaggerating. And then it happened to me. When something like this happens, you have to respect it. I believe I'm a [vessel](https://a327ex.com/posts/self_expression_vesselization) for ideas that come to me from above. So, in religious terms, this is literally God speaking to me. I can't ignore it.\n\nFor the next three years I built an elaborate story around that vision. My goal was to construct a narrative that would maximize it, so that when others reached it, they'd feel it with the same intensity I did. I built a story backwards from that moment, and it ended up having seven parts. The vision became the climax of part six, with part seven being the final showdown that concludes the whole thing.\n\nBut I never felt like I had enough to actually start making it real. That changed at the end of 2024. I had a dream, and from it came a simple idea that eventually became the beginning of the series. Suddenly, I could see a path forward.\n\nThe idea from the dream made it clear that the seven stories also needed to be games, merged with the books in specific ways. The main structure is that the books and games are separate. You can read the books without playing the games, or play the games without reading the books. You can listen to the books while playing the games, as most of them will be low-cognitive demand kinds of games you can play while listening to or watching something else. But importantly, you can also read an interactive version of the book which contains the merged game elements in it. All these options are available, the last one being the most interesting.\n\nSo the plan is releasing a series of seven games + books. I can already make games, but I couldn't write, so I had to learn. That's what I set out to do for 2025.\n\n<br>\n\nI immediately had two good story ideas. One became It Follows, and another became the story I'll write after it. They're mirrors of each other, in a way. But they were good enough ideas that I felt strongly compelled to bring them into reality, to not do so would be to be a bad vessel.\n\nIt Follows started from the ending, that's the first piece I had. But really quickly after that, I had the first chapter. It was actually the very first thing I wrote, and it came out in like 1-2 hours, extremely similar to the final version. That's when I realized something about how I work.\n\nI *need* to have the beginning and the end. I cannot start working on a story without those two components. Later, it became clear that I also needed a third: a constraint. This wasn't obvious immediately because, by chance, It Follows' first chapter has a strong built-in constraint, the curse's rules. This wasn't the case with the next story, where I have the end and the beginning, but not the constraint.\n\nThe constraint is important because when you have the beginning and the end, you want to logically fill the middle so that one thing leads to another. It's easier to do that with a constraint than without one. More importantly, the constraint should be thematically sound with whatever the story implies. In It Follows' case, the curse's frules, combined with how the character goes from follower to followed and how that flips again at the end, turned out to be both thematically coherent and generative. In my next story, there's no obvious constraint yet that follows thematically from the beginning and informs the end in the same way, which is why I haven't started writing it yet.\n\nBut this explains why I suddenly saw a path forward with the whole series after I had the dream. It wasn't clear why at the time, because I didn't know I needed these elements, but once the beginning was in place, suddenly the whole series was locked in.\n\n<br>\n\nThe same was true for It Follows. Once I had those three elements, filling the middle was easy. Well, perhaps not easy, but logical. I tried a bunch of different middle ideas and eventually, unfortunately, landed on the concept of what I internally called the \"grindr descent.\" Unfortunate because I didn't really set out to write porn. But I had this ending I thought was really good, I had this beginning I thought was really good, and I'm practicing so I can bring this seven-part vision that was sent to me by God to life. To shy away from writing this story because it's icky would be wrong.\n\nI knew I wanted the story to have a downwards arc from the beginning, because this story is also practice for story one of the seven-part series. That story will have similar downwards motion (without the sex), so using It Follows as practice for that general structure made sense.\n\nIn any case, once the grindr descent idea came to me, it became clear that it would fit into the story well, it was super generative, a lot of things came out of it naturally. I convinced myself that a way to look at it positively is that if I could write porn scenes well, I can probably write any kind of scene well. Sex is naturally kind of non-confrontational, so compared to battle scenes that sort of unfold themselves naturally, it's probably harder to write them. I don't know if this reasoning is valid, but it's what makes sense to me.\n\nI did get feedback that multiple people skipped the sex scenes. But this is kind of mixed... It could be because they're poorly written, or because the story only becomes \"about porn\" at around chapter 3. This baits people into thinking the story is about this curse mystery, and suddenly it becomes this huge porn-fest with no warning beforehand. So I don't know how much of that skipping is bad writing on my end or failing at managing expectations. Either way, it's still a mistake on my part.\n\n<br>\n\nOn top of the grindr descent, another idea that was generative was using myself as the character. Initially, I did this to decrease variables. It's my first story, I'm learning, so I decided on a single POV with me as the character. When I had to ask \"what would this character do in this situation?\" I already had the answer. Of course, because of how the story ends, it also kind of had to be me all along.\n\nBut this turned out to be the right choice. A lot of the story's themes emerged naturally out of the clashing between myself and the plot. If I were creating a character from scratch, I don't think the thematic cohesiveness the story has would have happened as easily. I kind of got lucky. I'm a real person with real contradictions and drives, and the curse idea just happened to hit on my personality in a particularly right way. \n\nOne could say that I had this specific idea in the first place because of this, that it wasn't luck but my high levels of intuition working through me without my conscious knowledge. That's a reasonable explanation, but it still wasn't a conscious choice. I don't remember exactly how I had the idea for the ending and beginning... In any case, I think this kind of cohesiveness can only happen consciously if you're an experienced writer, which I'm definitely not yet.\n\nI'm not going to give specific examples of this cohesiveness here, because I don't like when authors are like \"the story was about this\" or \"the story was about that.\" If you have to say afterwards what the story is about, you failed when writing it. A good test is: if two people are discussing the story, can one say \"but the author said this\" to kill the other person's theory? If they can, it means I said too much.\n\nAnd this is because... it ruins the story. If the author can definitely say it's about X and not Y, it ruins it for people who think it's more about Y than X. And who am I to say which is more important? The story is. It has the truth in it. But the truth is a [complicated, alien thing](https://a327ex.com/posts/unity_runtime_backstabbing_2023). You look at it from one angle and it's X, you look at it from another and it's Y. But they're both true at the same time. So it would be unwise of me, as an author, to reject certain perspectives because \"that's not what I wrote about.\"\n\n<br>\n\nNow, from the perspective of the activity itself, I'd say I found writing to be about as fun as making a game. However, finishing a story is certainly 100x more annoying and boring than finishing a game.\n\nWhen you're finishing a story you're mostly going over sentences autistically, making sure everything is correct, sounds good for the voice you want, is punctuated properly, has no grammatical mistakes, etc. It's really boring work, I didn't like it at all. And you can't even really automate it to the AI entirely because it's hard to get it to care about the exact things you do. I have some ideas on how to make this process less painful in the future, but I think some of it simply must be inherent to the craft.\n\nWhen you're finishing a game, though, you're mostly testing it as it comes online, watching all the builds finally work. It's much more engaging and fun, at least for the kinds of games I make.\n\n<br>\n\nThis difference in the finishing process made me think more about how games and stories differ fundamentally. Stories are linear, they follow a clear flow and don't change once they're done. It Follows will be the way it is now forever. Well, I guess there's a small difference if you had it on the same day I released it versus in the future, but that's a small thing. Games are different, they're more like looping.\n\nAnd if games are looping, the concepts of beginning and end work differently. For games, the beginning is the moment-to-moment action, what you do every few seconds that constitutes a unit of gameplay. Getting this down, having it running and playable, or at least having a solid idea of what it'll be like when it is, constitutes the beginning.\n\nThe end depends on the game. For the build-heavy games I like to make, the end likely means \"what does an end-run build play like?\" When the game is almost being broken by the player, what does that look like? What does it feel like?\n\nAnd the constraint... in a story, it constrains the possibility space for the middle while being thematically generative. What does this in a game? If you have a game like SNKRX where the beginning is snake + each unit auto-attacks, and the ending is a crazy build that one-shots the screen (ideally in a fun way), what's the constraint? I think it would be how you get from start to end in an interesting way, so the shop between battles. Then, maybe the general form of the constraint is the \"meta loop.\"\n\nLike a good constraint should be thematically generative, a good meta loop should be ludically generative. Taking SNKRX as the example, is the shop reroll system generative? Kind of, it takes you from A to B. But are there better alternatives? Are there different meta loops that specifically make use of what SNKRX's beginning and end imply?\n\nI think something like a pre-run draft does this. The details of this draft do not matter, but the act of choosing units among many possible at the start, fits the game better than doing this spread out throughout the run. You could also imagine that, once you have this pre-run draft going, you could have \"shedding\" moments mid-game where you alter your powers in important ways. A meta loop focused on frontloaded power choice with moments of change available mid-run, feels much more generative to me for this specific game.\n\nSo it seems that for future games, it probably makes more sense for me to tackle the ending first, or second. Get the beginning working and immediately have a few key end-game builds going to see how they actually play. Once those are fun, decide on a constraint that feels generative and takes you from the beginning to the end in a good way. This approach seems like it'll be more fruitful, and lead to better designed endings with properly thought out end-game builds, but also things like well-designed endless modes and so on.\n\n<br>\n\nAnother difference between stories and games is that once a story has enough to it, it has a logic of its own. It's like it becomes a living artifact with a clear purpose, and like there's a correct solution to any problem that arises.\n\nThere's a good example of this regarding #1's presence in It Follows, but I can't speak too much of it. I'll just say that #1's presence was initially much lower, and then when I finished the first draft, the story was basically screaming at me: \"If you make this more prominent, it will be super generative.\" And it was.\n\nThe way this works is kind of simple. If you want to be consistent with the story and with yourself as a good vessel looking for the truth, once the story is established, its truth becomes self-generating. You often come to points where you have choices A, B, or C, and it's obvious that B is more correct because it aligns with every element of the story better and potentially generates more aligned elements.\n\nIf you make this correct choice B here, you can also change the past of the story to make B even more correct, while also making it lead to even more correct choices later. This is essentially the story speaking to you. It's saying, \"if you choose B here, you can make the past and the future better\", and \"correct\" or \"better\" here means more coherent and cohesive along the themes it's clearly about.\n\nAnd so this property where the story talks to you in this way about what it wants to become is super clear with writing, but so far nothing like it has happened to me with games. Some developers mention that it happens to them, but I haven't experienced it. Maybe because the kinds of games I make are less strict and more moldable to what I want them to be, so they don't assert themselves as much? I don't know.\n\n<br>\n\nAll in all, I'm happy with It Follows, otherwise I wouldn't have released it. In the end, I brought a good idea to life as best as I could, I practiced for my big project, and I also actually lived my life a little more because of it.\n\nThis last part was an unexpected benefit of becoming a \"writer\" in my head. Before, I used to really dislike talking to most people. If they weren't intelligent enough, if they didn't have interesting things to say, I unironically thought they were beneath me and somewhat worthless. Now, this actually isn't the case anymore. Simply because every person is potential material for a future story. I love listening to people talk now, and it's really easy to get people to talk about their lives, everyone wants to. It helps that I have a friendly face to a percentage of the population, so they open up even more easily.\n\nThis may seem kind of twisted and sociopathic, but 2025 was easily the year where I talked to most people and where I actually enjoyed it. However twisted it may seem, it objectively made my interactions with other human beings better. I definitely can't say making games ever had that sort of effect.\n\n<br>\n\nOther than going back to making games, the next thing I want to do this year is learning how to make music. This big project needs to have a specific musical vibe to it, and I think I can learn to do it well enough.\n\nI think the feeling of making music will be closer to stories than to games. They're both linear, both consumable with a clear end, their constraints are more clear than a game's. But music is harder to learn than writing.\n\nWith writing, the act of writing itself is \"easy,\" for the kinds of books I want to write anyway. I'm under no delusions that I'm a particularly clever wordsmith, capable of wowing people with my spectacular prose. Most books I end up finishing tend to be written in a simple way, and I want to write the same way. The hard part is coming up with an idea, structuring the story, making sure the constraint is good, etc.\n\nWith music, there's a similar distinction. There's the act of making the sounds on the computer, and then the theory behind it. The theory is not easy to learn, but it's not hard either, it's fairly logical and self-contained. But the hard part for me is making the sounds themselves. There's so much to learn, so many different ways of doing it, so many programs and techniques. It's an endless hole of possible obsession.\n\nCompared to writing, where I just open Notepad (It Follows was entirely written in Notepad) and start writing, the \"act\" part of making music is 100x more complex. I don't know how it'll go yet, but it seems like it's an overall harder and longer journey.\n\n<br>\n\nI think... yea, I like writing, and I'll keep doing it. It's also good to switch between activities like this. I'm so productive as a game developer right now because finishing It Follows was so tedious that it made me beg for doing anything else. It was actual torture. So adding music to this and rotating between gamedev, musicdev, and writing will probably be good for my productivity and creativity.\nTitle: Competition isn't real\nDate: 2022/06/15 11:39\nTheme: old\nLink: /posts/competition\nTags: indiedev, psychology, strategy\nThumbnail: /media/posts/competition/_thumbnail_competition.png\n# Competition isn't real\n\nThere is no competition among indie developers. The more successful indie games there are on Steam the higher your future chances of success as an indiedev.\n\nOutsized success creates demand for more games similar to it, and as more of those games are created and each further outsized success happens, a new genre starts to establish itself.\n\nAs the audience for this new genre increases, more developers are enticed to make games for it, creating a positive feedback loop that only ends when no more outsized successes for it happen.\n\nSince Steam established itself as the king of PC indie games this process has happened many times, both with entirely new genres and with old ones that were revived. The faster each subsequent outsized success happens, the more noticeable this process is.\n\nCurrently this is happening with the genre that Vampire Survivors has created, and while it’s still very much up in the air how far that will go, it’s clear that there’s something solid there.\n\nThere have been multiple successes following it and the audience for it is increasing. Although none quite as outsized as the original, which is somewhat bearish, but at the same time it’s only been like 6 months, so some patience is warranted.\n\nIn any case, I’m not the only one who thinks like this:\n\n::youtube YpaNnX_9Q5s start=\"1031\"\n\nThere’s also a more pragmatic reason to not believe in competition: the less you believe in it, the more your creativity will flow. You are what you practice becoming, and if all you practice is a zero-sum mindset then that will be reflected in your work.\n\nConsider the developer who wants to be uniquely creative at all costs. If he’s working on a game and someone else releases something that’s very similar to his, he quits and starts on a new one, because “what’s the point of releasing this game when the idea is already out?”\n\nThe instinct to only want to release new ideas ironically will lead this dev to never creating anything actually new and interesting, because he has a zero-sum instead of an abundance mindset towards ideas. He thinks they’re a resource that can be consumed and spent, instead of one that’s endless and regenerates itself.\n\nWhen working as a creative you want to reject competition for the same reason that you want to [reject luck](/posts/luck), it’s a pragmatically true idea that will coerce your body into more productive outcomes.\n\nIf you care about auto-attacking + build-making games, for instance, you should be making Vampire Survivors clones right now. Your ego should be taken out of the equation and you shouldn’t consider yourself above making a clone of a popular game.\n\nRemixing and building upon existing work is what advances a field/genre. If you care about it then you should contribute to it regardless of how much money or status that contribution will give you.\n\nAnd a final point: there’s only one instance in which competition exists, which is when two games are _very_ similar and released within a few days to a week of each other. In this case both games will compete for the attention of the same audience and then only one will win, generally the one with better memetic fitness.\n\nFor instance, if a game _very_ similar to SNKRX - like autochess elements, auto-attacking, build-making and arena shooting - released 3 days after it, that would be competition. If it released 3 months after it, then it wouldn’t. In the latter situation the game actually benefits from SNKRX coming out first and establishing an audience for that exact type of game, which makes the second’s success more likely.\n\nEven if you define “very similar” very broadly, like, let’s say, your game is a roguelite and you really don’t want to compete with any other roguelites, you can use something like [steamdb’s upcoming page](https://steamdb.info/upcoming/) to pick a relatively empty day with no other big roguelites coming out.\n\nIf you [look it up](https://store.steampowered.com/search/?sort_by=Released_DESC&tags=3959) there are not that many roguelites released daily, it’s usually one or two. And if you filter further by “good” roguelites then they’re a lot more sparse. And roguelites are like one of the most popular genres of the last decade, so for other genres this exercise goes even better.\n\nIn any case, competition isn’t real, except when it is, but it generally isn’t. Believing this increases creative flow and aids egoless building, which precedes true creativity.\nTitle: Luck isn't real\nDate: 2022/06/14 20:39\nTheme: old\nLink: /posts/luck\nTags: indiedev, luck, psychology, strategy\nThumbnail: /media/posts/luck/_thumbnail_luck.png\n# Luck isn't real\n\nYour body and your mind are you and you are them. Your body speaks to you through instinct, your mind speaks to you through reason. You feel in control of your mind because reason is weak, you don’t feel in control of your body because instinct is strong.\n\nBody to mind coercion is the default state. Strong ancient instinct is a natural winner, which is why you eat and play more than you should, and exercise and work less than you should.\n\nMind to body coercion is slow and painful. Convincing your body to do something it doesn’t want to takes time and effort - new habits have to be built from scratch, old habits have to be resisted and then forgotten.\n\nBoth body to mind and mind to body processes are happening all the time inside you, and sometimes they can also reinforce each other. This loop can also often happen subtly, via the distortion of the lens through which you parse reality.\n\nSuppose your body wants to feel like X. This will make your mind filter reality through the lens that makes your body feel like X. Suddenly ideas that make you feel like X will instinctively seem right, and ideas that don’t will instinctively seem wrong.\n\nAs more ideas in your head are serving X, this in turn will make your body seek out X even more, which will make your mind filter for X even more, which will strenghten the loop further. Because you are what you practice becoming, this process ends with you becoming an avatar of X.\n\nX can be a lot of things. For the purposes of this post, let’s suppose it is “lack of control”. If your body wants to feel like it’s not in control, it will automatically make your mind filter reality so that ideas that make you feel like this seem right.\n\nBecause the mind is weak, any idea that is merely objectively true will coerce it. For instance, luck. Take the following statements:\n\n1.  There are lucky events that happen that are outside of any one person’s control  \n2.  If someone is successful it was in small or large part due to such lucky events  \n\nBoth of these statements are objectively true. But they are not pragmatically true, because the idea of luck feeds your body’s need for lack of control, which makes you parse reality through a lens of lack of control, which further feeds your body’s need for lack of control, and as the loop plays out you will become an avatar of lack of control, which is a person who lacks control.\n\n> Objective truth is scientific truth, concerned only with objective reality, reason and logic. While pragmatic truth is darwinian truth, concerned with fitness, or only what is true enough to lead to better outcomes in the future.\n\nIn this case, the only way to break the loop is through active, conscious rejection of the idea that luck matters or is important. You have to psyop yourself into the belief that luck isn’t real to not become its avatar and to eventually gain the pragmatic benefits that come from not believing in it.\n\nMany people have gamed out this process instinctively - which is incidentally what allows them to be successful - such as contemporary philosopher and part-time streamer, Félix Lengyel:\n\n::youtube GfYHU3jVky4\n\nWhen you don’t believe that luck matters you are rejecting your body’s temptation to not be in control, which in turn makes you look at reality through the lens of what you actually control. xQc’s gaming example here is perfect. Seeing an opportunity as opportunity is a skill that increases the more you do something, but it also increases the more you believe you are in control.\n\nThe more you believe you are in control, the more you’ll look for things that you control, and the more you’ll affect the outcomes of whatever you’re doing. Rejecting luck thus is a pragmatically true idea because it coerces your body into parsing reality through a lens that gives you more power to affect it in your favor.\n\nAncients often spoke of destiny. Luck and destiny are both concepts used to explain that which happens outside one’s control. However, luck is indeterminate and destiny is determinate. When you get lucky it’s a random event, a statistical anomaly, it just happened. When you are destined for something, you either become worthy of your destiny and reach it or you are a failure. One idea is clearly more pragmatically true than the other.\n\nConsider Eric Barone’s words:\n\n::youtube 0ti5bBgEJaw\n\nWhich indie game developer will more likely succeed: the one who believes himself destined for success who will not stop until he becomes worthy and succeeds, or the one who believes it’s mostly down to luck and that it just happens?Title: Oh yea, it occurs to me I have not actually spoken\nDate: 2026-07-28 21:37:23\nKind: message\nLink: /posts/2026-07-28-213723\n\nOh yea, it occurs to me I have not actually spoken about what I think of the fact that the AI can help make games so easily now. In general, I think it's good. It follows from [Competition isn't real](/posts/competition) that I think more people being able to make better games benefits everyone, I genuinely believe this and it has not changed. The cases where competition is real, similar games released at similar times, might be made slightly worse by AI, but I also think game variety will increase, therefore these will probably nullify each other. It also follows from [Steam AI Score](/posts/2026-06-21-103134) that I believe Valve has/will have the tools to keep up with the increase and keep surfacing quality products. And then there's the fact that all the \"Claude made this in one prompt\" posts are kind of like, okay, that's pretty cool, but you still have to spend the long tail of caring about the details to make the game shippable. If you've been following this site for a while now you should know that most of the job is just that, being autistically obsessive about the details and making sure everything feels right. Will the robot be able to get more and more right over time? Yes, and that's good, but you'll still need to assert your will so that it does your actual artistic vision. There's also the fact that the AI will get better first on details that are legible, things like visual effects, juice, models, environment, etc, but will get better last, if it does, on details that reveal themselves over time in the context of a single project, like is this game actually fun, balance issues, pivots into new directions because the current one isn't working, and so on. Overall I'm pretty excited about the rate of progress and that the future will allow me to more easily bring my visions to life.\nTitle: Continuing the continuation of the Steam AI score\nDate: 2026-08-19 00:43:26\nKind: message\nLink: /posts/2026-08-19-004326\n\nContinuing the continuation of the [Steam AI score idea](/posts/2026-08-07-125138), in a world where who makes the art becomes a scarce resource, proof of work becomes a valuable asset, as when games are cheap to make, who makes it becomes easy to fake, so a public record of making defends against it. Therefore, things like AI logs, amassed over years, are probably the best asset a game developer could be accumulating right now, other than money. And right now because realizing this in 2030 when it's obvious is less useful than realizing it in 2026. Having successful games before making games with AI was possible is actually the best proof of work possible, but the second best is AI logs now rather than later. I assign fairly low probability to this future actually coming to pass and this mattering, I started doing the AI logs because I like being transparent, but it also happens to have other benefits like this. Ultimately it's hard to lose when you're committed to never lying and always telling the truth. What was the quote...\n\n>\"If you just seek the truth, and tell the truth, and give all of your energy to living as truthfully as possible, then it's impossible to lose the thread. To the man with a real thread, good things come, but only because a truthful life catches many imperceptible tailwinds, all kinds of implicit alpha you'll never be able to fully explain.\" [https://twitter.com/jmrphy/status/1731411321285972118](https://twitter.com/jmrphy/status/1731411321285972118)\nTitle: I'm adding items to Horse Game and I was thinking\nDate: 2026-08-29 02:43:59\nKind: message\nLink: /posts/2026-08-29-024359\n\nI'm adding items to Horse Game and I was thinking about how before I test if the game is actually fun beyond the initial fun the concept itself carries, I'll have added around 150 items, and they're all 100% finished, so it's quite a bit of work to gamble on a game that might not be fun. But ultimately it's not really a gamble, because even if it isn't fun, I have already done more work designing the effects (both visual and aural) for those items than I have for any other game in the emoji style, which means I can reuse those effects in other games of the same style rather well. There's also the fact that this type of build-heavy game carrying this kind of risk, where they're only fun after there are 200 or 300 or so items, is part of the value they have, and so training myself on this is what I should be doing, as these are the kinds of games I want to always make. In the end, no work is ever wasted.\n"}
{"chunk_id":"b51e21","wall_time_seconds":0.000002693,"exit_code":0,"original_token_count":5450,"output":"Title: Lessons learned from releasing my second game\nDate: 2021/05/18 14:28\nTheme: old\nLink: /posts/lessons_second_game\nTags: indiedev, snkrx, strategy\nThumbnail: /media/posts/lessons_second_game/_snkrx_logo.jpg\n# Lessons learned from releasing my second game\n\n[SNKRX](https://store.steampowered.com/app/915310/SNKRX/) is the second game I’ve released on Steam and in this post I’ll go over my thoughts on its development and how it performed. I made this game in the past 3 months as a dev challenge to start making and releasing games more consistently.\n\n::youtube kfBCn4hn5UU\n\n# Financial Results\n\n## Stats\n\nTo get it out of the way, first let’s go over financial results. I’m writing this only 1 day after the game has released, but Steam seems to be consistent with how it treats games after how they perform initially so I can extrapolate somewhat safely from here.\n\nHere’s how [BYTEPATH](https://store.steampowered.com/app/760330/BYTEPATH/), my first game, did for its first day and a half:\n\n::image /media/posts/lessons_second_game/_snkrx_1.png\n\nAnd here’s how SNKRX did for its first day:\n\n::image /media/posts/lessons_second_game/_snkrx_2.png\n\nSo roughly about 1/3rd as well. One interesting thing is that both BYTEPATH and SNKRX released with about 200 wishlists. At these low numbers most of how the game performs is related to how much external traffic you can generate for it rather than people who wishlisted it buying it.\n\nBYTEPATH had quite a few popular reddit posts on release day:\n\n::image /media/posts/lessons_second_game/_snkrx_3.png\n\nWhereas SNKRX had pretty much nothing:\n\n::image /media/posts/lessons_second_game/_snkrx_4.png\n\nSo this explains most of the difference. I’ll go over some of why that difference happened in the rest of the article.\n\n## Projections\n\nAfter 3 years on the store BYTEPATH has made $12K and sold 10K units, which means that, should everything remain equal, SNKRX should go on to sell ~3K units and make about $4K.\n\nI end up making half of that, and since I live in Brazil and the local currency is worth less than the dollar I also should multiply it by 5 or so, which ends up being a total of R$10k. For 3 months of work this would be about R$3.3k per month which is somewhat above the average local salary.\n\nCompared to a [similar analysis](https://github.com/a327ex/blog/issues/35) I did for BYTEPATH this is slightly better results, despite being worse in absolute terms. This is mostly due to this game taking only 3 months to make and also due to currency conversion being more favorable now than it was 3 years ago.\n\nOverall I’d say that I’m neutral on this performance. It could have been worse, but it also could definitely have been better. The main thing I’m positive about is that if I can keep this pace up and release a game like this every 2 months or so then technically I’d be making a living off games, which would be nice I guess.\n\nAlso, “should everything remain equal” is a big assumption, because I don’t know if SNKRX will progress in the same manner that BYTEPATH did over time. BYTEPATH had quite a strong response on release day and that seems to have shot it up in the algorithm’s eyes for the rest of its existence on Steam.\n\nIt’s possible that SNKRX falls down the void of Steam and thus experiences a much more timid increase in sales over time in comparsion, which would decrease the results significantly. I guess I’ll see what happens as time passes.\n\n# Programming\n\n## Engine\n\n### Unchanging code\n\nSince about the middle of last year I’ve made it an explicit goal to change my engine code as little as possible between projects. Before then, whenever jumping around projects I would try to do things a different way to see if I found something that was better. Be it 0% OOP, 100% OOP, ECS, event driven, whatever. I’ve tried lots of different high level concepts for organizing engine code.\n\nThis would lead me to rewrite some of the common code between projects in these different ways and that would end up taking some time and effort. I was fine with this because I was mostly exploring but at some point I’d had enough and decided to just settle on one way of doing things.\n\nThis settling was helpful as quickly after it I was getting more done faster than ever before, and with each failed project, instead of taking nothing but experience from it, I was also taking artifacts, since each project would require a few different engine level functions or modules, and I would make sure to document those well enough that they could be reused in the future.\n\n### Artifact accumulation\n\nThis changed my mindset from one of “premature generalization is bad” to one of “artifact accumulation is good”. Everything I code for a project now I’ll try to see if it can be made into reusable code for future projects. The majority of it tends to be very specific so it can’t, but sometimes when building something there’s a very clear path for how that could be a generic system that would work well in other projects.\n\nOne good example of this is what I call a Nodemap. BYTEPATH had a really big skill tree in it and the way I developed that was to just have a [huge table defining the skill tree](https://github.com/a327ex/BYTEPATH/blob/master/tree.lua). This table is then fed to the game which creates the appropriate objects (nodes and edges) out of it.\n\nFor the game before SNKRX (which I dropped because I went over the 2 month limit with no real end in sight), I was doing both a skill tree and an overworld map thing, something like Slay the Spire’s map for choosing the next arena. It turned out that the code for both systems was very similar, since they’re essentially graphs with nodes and edges and rules saying that you can’t go to the next node until you’ve cleared a neighbor.\n\nSo I abstracted this out into a more generic Nodemap class that I didn’t quite get to use for that project nor for SNKRX, but that I’ll make use of for a future one for either a skill tree or an overworld map.\n\nThis kind of artifact accumulation has been happening for me since last year and it’s been generally very useful, especially since I’ve been careful about documenting each of those systems a lot, as I know that I intend on using them for a very long time.\n\n### Lua frameworks\n\nIn general I’ve been pretty happy with Lua being my language of choice for making games, as well as with LÖVE as a framework. They both mesh together pretty well with how I like to do things and they give me enough freedom to reach the conclusions I need to reach at my own pace.\n\nI would love for there to be some competition in the “Lua framework” space though, especially one that works well with the web. There are plenty of Lua engines that work well with the web but I’m generally put off by engines because they constrain me too much.\n\nI’ve been very inspired by [Amulet](http://www.amulet.xyz/) lately which is a Lua framework that fits the bill and it has a lot of good ideas, but I’ve never managed to go beyond a few small prototypes with it because it’s lacking quite a few features I’ve grown accostumed to.\n\nIt’s the kind of thing where I know what “better” looks like: something that has LÖVE’s API and most of its features, that works 100% on the web, that is also Lua based, and that has easier C/C++ integration so that adding modules in those languages is a workflow supported by the framework itself. [LÖVR](https://lovr.org/) gets most of these things right, unfortunately it’s very focused on 3D and I haven’t played around with it enough to see how feasible it is to use it for 2D projects.\n\nThe framework could also be C/C++ based, since I can just write the Lua bindings myself. There are lots of C frameworks, but their main problem is that because they’re written in C it’s rarely the case that the authors bother to make the API nice and easy to use like LÖVE’s so that idiots like myself can actually use it. In general most of these frameworks will have me dealing with OpenGL type of code directly and I don’t really want any of that, I just wanna draw a square.\n\nIn any case, LÖVE works well enough for most things and I’ve gotten used to it for long enough that it would be mostly a waste of time to try to redo it or get used to another framework/engine. LÖVE’s web export is kinda buggy but it mostly works, the features it has are very complete, especially the graphics end of it, and while integrating C/C++ modules can be annoying it’s also not really hard for the most part, and so it just works.\n\n### Sticking with what works\n\nIn the [previous post](https://github.com/a327ex/blog/issues/31) I made like this I had a lot of really hot opinions about many things, but one of them was pretty much my argument for making my own engine and how it would be great, and how Unity sucked, and all that.\n\nAbout 2 months after writing that post I managed to make a fairly OK initial version of the engine I wanted, and then 2 months later I realized that everything I wanted to fix about LÖVE was actually fairly doable without having to rewrite most of LÖVE in C, meaning that I could fix most of my problems with LÖVE without having to replace it.\n\nBut more importantly, LÖVE actually has a lot of graphics features that I would have to reimplement, and it turns out that I REALLY didn’t like doing low level graphics code at all. So I just decided to go back to it, fix the things I wanted to fix with my new-found knowledge and then move on. It was a very fast and useful lesson about the actual value that the framework was providing me which was somewhat mistated when I wrote that post.\n\nNow 3 years later, after watching the engine vs. no engine debate evolve over the years and noticing how it went for me, I’d say that for the most part I think now people should just stick with what works for them.\n\nEvery engine/framework or way of doing things has tons of people constantly complaining about them, nothing is perfect. [Even Unity still has to this day people complaining about this or that.](https://www.reddit.com/r/Unity3D/comments/n44shs/unity_then_vs_unity_now/) Maybe those are valid complaints, maybe they aren’t. But each individual knows better than anyone about if they should switch to something else or just keep at it.\n\nFor my part I just think sticking with what I have is better, unless I find another engine/framework that is better and lets me keep using the same workflows for like 90% of the things I do, since I want to have to relearn things as little as possible.\n\n## Code\n\n### Local code\n\nI’ve been very big into doing as much as I can locally. What I mean is code that defines everything about some functionality in the same place, rather than split off in multiple places.\n\n# Strategy\n\n## A game every 2 months\n\n### Motivation\n\nSince the start of 2020 I’ve been trying to make and finish small games consistently. Projects that take between 1-3 months to make from start to finish and that I actually release on Steam. For the entirety of 2020 I failed to do that.\n\nI worked on about maybe ~10 different projects that year and most of them weren’t released for anyone to play:\n\n::youtube 1szzTEk5fpQ\n\nThe only one I managed to release in any shape was [a prototype that I made in 3 days at the end of the year](https://a327ex.itch.io/jugglrx-prototype). I’ve been inspired by a bunch of people doing this, most prominently [Chilla’s Art](https://store.steampowered.com/search/?developer=Chilla%27s%20Art).\n\nThe reasons for me to do this are simple: it seems impossible for me to commit to longer projects. I just don’t have the discipline to do it. Most projects fail in the same way and they’re all due to a general lack of direction mixed with poor scoping. I get way ahead of myself and try making something bigger than I can handle and eventually my motivation drops enough and I just can’t work on the project anymore.\n\n### Scoping\n\nSo given that this is the problem, it seems like the solution is just starting small and building my game finishing skills and discipline up over time. However, the reason I failed all this time is also because the kinds of games I want to make naturally tend to be games that are easy to overscope.\n\nI’m particularly interested in games where you can try lots of builds. Be it with a huge skill tree, lots of items, lots of characters, etc. Genre doesn’t really matter, it can be a roguelite, a base-building tower defense deckbuilder, platformer, whatever, as long as you can play it in lots of different ways.\n\nIt’s very hard for me to be excited about making small games that don’t have this “lots of builds” component. So over time I’ve found a way to think about this that has helped me at least deliver this game, and I hope it does many more in the future.\n\n### Small long games\n\nEssentially, you can think of games as either big or small, and either short or long:\n\nBig vs. small refers to how long the project would take to complete and/or how many people. Short vs. long refers to how long the game lasts to the player. A game that lasts a few minutes or a couple of hours would be short, a game that lasts tens or hundreds of hours would be long.\n\nThe space of big games, either short or long is explored by lots of people. The space of small short games is also explored by lots of people, generally indiedevs. But the space of small long games is largely ignored.\n\nPerhaps ignored it too strong a word, but for some reason people assume that projects that last people a long time should also take a long time to make, but to me it seems like this isn’t true. So I’ve set my sights on exploring this space, largely because it aligns well with what I’m interested in making.\n\n### SNKRX\n\nOverall this strategy went well for SNKRX but I fell short of making a very long game. I think the game I made before, BYTEPATH, had much more long potential than SNKRX, so this was a failure.\n\nI ended up finishing an initial version of the game 1 month in, and then the next 2 months were spent adding content. But I lost quite a lot of time due to poor decisions made along the way and it just limited my ability to add more long hooks into it.\n\nIn the future I should probably take less than a month to finish the initial version of the game. It would be really cool to finish it in 1 week, for instance, and then have the next 6-7 weeks just for adding content + long hooks.\n\nI think on the whole the strategy of focusing on something small and then expanding on it with more stuff over time seems solid, I just need to execute it better and execute on ideas that have more potential to last longer too.\n\n## Steam followers\nTitle: High level game design\nDate: 2022/12/16 21:13\nTheme: old\nLink: /posts/high_level_game_design\nTags: design, indiedev, luck\nThumbnail: /media/posts/high_level_game_design/_thumbnail_high_level_game_design.png\n# High level game design\n\nWhen people speak of luck, they mean that there are things outside of one’s control and/or knowledge structures that affects some outcome. Let’s call this the unknown. A game’s success is both composed of the known and the unknown. For instance, consider the image below:\n\n::image /media/posts/high_level_game_design/_high_level_game_design_1.png\n\nHere max_oats is saying that despite Eric being really good at the known, Stardew Valley only got as big as it did because of the unknown. This is likely true, and if you were to ask Eric if luck played a role in his game’s huge success, he would say yes.\n\nRejecting luck, then, is about rejecting the unknown. It’s about increasing what is known, and decreasing what is outside of your control and/or knowledge structures. It’s about pursuing understanding and knowledge of that which is not yet understood.\n\nMost of what is not understood about game design is “high level”. In general, high level decisions, which both tend to be the easiest to make, but also the more impactful, are the ones people discuss the least. This might be because they’re not easily verifiable, or because they tend to be the kinds of soft ideas that often rely on intuition and experience, or maybe because they’re simply… hidden.\n\nFor instance, consider [Derek Yu’s new post](https://derekyu.com/makegames/risk.html):\n\n::image /media/posts/high_level_game_design/_high_level_game_design_2.jpg\n\nIt’s a really good high level game design post - likely the best I’ve read this year - focused on risk. Or more specifically, on how to make high level game design decisions that decrease your risk as a small indiedev and thus make it more likely that, in the long term, you’ll succeed.\n\nIf you look at this post, it has a few of these soft + unverifiable + based on intuition/experience ideas, like “bigness” or “fiddliness”, or even just the notion of “experimentation required”, which is how much design exploration a certain type of game will require of its developer.\n\nThis is an obvious thing to care about, and looking back at how I decide which games to make, I intuitively have always cared about how much experimentation a game will need and always made good decisions on this front. But if you asked to me write a blog post on risk, like Derek has, it’s very unlikely that I would talk about “experimentation required” as a concept. It’s just something that I naturally do correctly but that I was completely blind to until now.\n\nThis, to me, is the core of the unknown when it comes to game design. It’s all of these soft ideas and decisions that we intuitively make but that we are blind to. The rejection of luck and the rejection of the unknown then is about paying attention to these and making them visible.\n\nMaybe the main unknown contributor to Minecraft’s and Stardew Valley’s success is that people feel a sense of dysphoria living in modern society and they want to go back to simpler times. It’s certaintly a reasonable idea, it’s also very soft and unverifiable, and it’s the kind of thing that both Notch and Eric could have intuited without conscious awareness while making their games.\n\nMaking such perspectives visible in your head, and trying to assess how much they matter, constitutes their bringing from the unknown to the known.\n\n---\n\nI spent this year following a game called Super People and talking about its design from a personality trait based perspective. I think this is one of those high level design ideas that people are blind to but that has some utility, so it makes sense to look at games from that lens. Unfortunately, [Super People’s update](/posts/super_people_major_update) failed to save it from its impending doom, which puts a cap (at least for now) on how useful/right this idea actually is.\n\nBut despite the failure, it was a good effort. The idea has some utility, as it was able to at least track the decisions the developers decided to make with each update, and despite ultimately being a less powerful explanatory concept than I initially believed it to be it’s still clearly useful.\n\nThe more of these useful ideas I can generate, even if they might not be too strong individually, the more possible explanations I’ll have for why something is successful, and thus the more I’ll be able to succeed. One of the interesting things about luck is that it’s a thought terminating thought, once you say something happens due to luck that’s your excuse to stop thinking about it because by definition it’s unknownable.\n\nI reject this line of thinking entirely, so naturally I’ll just gravitate towards trying to find explanations for things, no matter how tenuous or weak such explanations might be. I think this is a good instinct… I think high level game design is where you can add the most value for the least cost. People often mention how having good, one-of-a-kind ideas is important, and I think this is probably a good way to get there.\n\n---\n\nWe live in a golden age of indie game development, all you have to do is make a good game, release it on Steam, and it will make money. You should take advantage of this while it lasts.\n\nThe world is becoming a rough place… We have no idea what the future holds. I say this often, but if you’re not financially secure this should be your #1 priority. If you have the skills, you can achieve this through indie gamedev.\n\n::image /media/posts/high_level_game_design/_high_level_game_design_3.png\n\nAs Thomas says, you can go really far by just copying what’s successful (of course adding your own small twists to it) and having a good eye, both aesthetically and design wise, for what works and what doesn’t. Sit down, cleverly pick a [low risk](https://derekyu.com/makegames/risk.html) idea, and just work on it until it’s done.\n\n---\n\nAnd to that end, this is certainly the last blog post I’ll write this year, but also probably the last blog post I’ll write in a good while.\n\nThis year has been a mixed bag for me, a lot of good things happened, but I didn’t manage to release anything. In part this happened because I just spent too much time thinking instead of doing. There’s definitely some kind of switch in me where if I’m in philosophing/thinking mode it’s more difficult to get actual work done.\n\nSo, you know, I’ve thought enough for now. I’ve had some good ideas, I know where I stand, I know where I have to go, and that’s it. I had two blog posts I still wanted to write, one on the morality of making games and another on good = popular, but I’ll leave that for the future.\n\nIf you’ve read my blog up to here, thank you very much. Hopefully some of the ideas here have been useful. Have a merry christmas, a happy new year and just like make game!"}

**Reading archive entries**

**Reading offerings sections**

Script completed
Wall time 0.6 seconds
Output:

{"chunk_id":"c49d65","wall_time_seconds":0.000003356,"exit_code":0,"original_token_count":9605,"output":"Title: Offerings to God\nDate: 2025/05/11 00:46\nTheme: old\nLink: /posts/offerings_to_god\nThumbnail: /media/posts/offerings_to_god/_offerings_to_god_1.png\n# Offerings to God\n\nI watched [Amadeus (1984)](https://www.imdb.com/title/tt0086879/) today. I had never seen it. That scene with Salieri dropping the papers has become a meme on twitter, and I happened to [use it](https://a327ex.com/posts/ai_usage#current-state-of-ai-my-evolving-opinions-and-thoughts) to make a point in a blog post, so I figured I should actually watch the movie. People say it's a great movie. And they were right. It's one of those archetypal human stories that read as true 500 years ago and 500 years from now. It's a great movie. If you haven't watched, I highly recommend watching it before reading this post.\n\n## Salieri's Prayer\n\n*\"Lord, make me a great composer. Let me celebrate your glory through music and be celebrated myself. Make me famous through the world, dear God. Make me immortal. After I die, let people speak my name with love for what I wrote. In return, I will give you my chastity, my industry, my deepest humility, every hour of my life.\"*\n\nSalieri's prayer is mistaken primarily because he is not praying to make great music, but to become a well-known composer. His aim is glory. But I find that his next mistake is deeper and more serious. We could reword his prayer as *\"God, I will give you myself and my works, as long as you make me a greatly recognized and respected composer.\"* He is giving himself up completely and utterly, *as long as* he gets the result he wants. This is a conditional submission. In Salieri's example, the reason why this conditional submission is wrong is easy to see. But in most cases, these conditional submissions are more hidden and difficult to parse.\n\nLast year (2024), I decided that what I needed to do most was go back to my plan of releasing a game every 2-3 months, but to do so while dropping the \"a327ex\" persona, and starting a new dev identity from scratch. But as I started doing it I began to question this decision. Something about it felt deeply off. My intuition was telling me this was *wrong*, somehow.\n\nUsually when this happens it's because I haven't thought the problem through enough, and in those cases it helps me to write about it. So I wrote what follows below. The perspective of the post below is: \"this is going to be made public in 2-3 years in the future once this experiment is done with, whatever its results may be.\"\n\n## X\n\n\"X\" is who I will use to release game from now on. I don't know how long this will last, but I want to answer a few specific questions with this new dev persona:\n\n1. **Can a new game developer succeed on Steam with skill alone and very little marketing?**\n\nX will start from zero. I will release games on Steam for free at first to gain an audience, and I will not do much marketing for any of those games. I will use twitter and reddit for progress posts, as well as for posts when I release a game. I will use 4chan's /agdg/ and /v/'s gamedev threads for progress and general banter. I will join a few gamedev discord servers, but mostly for conversations and not marketing (although with devs it normally happens that as you talk to people they ask about your games).\n\nI don't think I'll use anything other than those. I will especially avoid writing any blog posts or making YouTube videos. I want this developer to be largely invisible beyond the basics. And I also want most of the traffic to the game to come from Steam users and people who decide to follow my developer page on Steam. If I manage to find an audience that consistently plays my games, and eventually make money from this (say ~100k USD), then I'll consider this experiment a success.\n\n2. **Are my games succeeding or failing because of their actual qualities?**\n\n\"a327ex\" was a very outspoken developer. I explicitly only spoke when I disagreed with something or someone, and I spoke in a crude and often disrespectful manner. Being honest and saying what you feel is the truth is good, but it has various drawbacks. One of them is that people are not like me in how they consume their media. For many people, if they know a developer is X, where X may be any number of negative or positive attributes, that will change how they view what that developer creates.\n\nPeople who dislike me will refuse to play any game a327ex makes, no matter how good the game may be, solely due to the fact that a327ex made it. People who like me will give a chance to any game a327ex makes, a chance they wouldn't have given a similar game made from someone else, solely due to the fact that a327ex made it. Both of those are distortions of reality that are not desirable. Many people are like this and they can't help it. I don't blame them for it, it is what it is.\n\nHowever, I am interested in knowing if my games are succeeding or failing due to their actual qualities and not any other external factors, therefore reducing this particular kind of noise is necessary. And this reduction can be perfectly achieved by a new dev identity, particularly one that is invisible. This goes well with doing very little marketing.\n\n\"X\" doesn't exist, he has a small social media presence, he has no blog posts, no YouTube videos, no one knows much about him and they never will. Other than the content of his games that some people might find objectionable over time (although most ideas I have for games now are not really controversial at all), no one will have many *external* reasons to judge him either positively or negatively. He's a ghost, he's off the grid. And so through him I will know if my games are succeeding or failing based on their actual qualities or not, I will know if I have achieved true success or true failure.\n\n3. **Is it luck or is it skill?**\n\nDue to starting from zero and doing little marketing, the results of this experiment also partially answer the luck vs. skill question. And I say partially because to answer it with very high confidence I would need to run this experiment multiple times. I would be happy with succeeding ~5 times from zero like this to conclusively say that skill is the only thing that matters, but I'm happy doing it once, if I manage to do it.\n\nOne of the reasons why doing it more than once isn't necessary is because the answer doesn't really help anyone but me. If I succeed I will have proved to myself that I can succeed from scratch again, and that my success with SNKRX was likely not due to luck. That's a great question to have answered. But can I extrapolate from that and say that if I can do it, so can other people? I think not. I think the fact that I can do it simply means that I'm smart and creative enough to be able to do it, but it doesn't generalize to other people because they're not as smart nor as creative as me.\n\nUltimately, I still succeeded because of some kind of luck, which was a genetic luck. I was born with a fairly high IQ, to a decent enough family, with fairly high levels of creativity. Not a lot of people spawn with these same stats, so whatever lessons can be learned from my success by others are limited. People don't like thinking about these inherent biological differences, so providing stronger evidence on this front won't really convince anyone. You can't make people see what they don't want to see, and in fact whatever advice I have for most people would be incorrect, since most people are not wired like me.\n\nBut now let's suppose I fail. If I fail it means that the status quo of belief when it comes to skill vs. luck, and even Steam itself, is correct. If I , someone clearly smart and creative (maybe less than I thought), can't succeed consistently, what chances does anyone else have? This would mean that I, in fact, got lucky with SNKRX. This is something that I don't think is likely. I think given a certain amount of time, a few years probably, I'll succeed with X, and thus this failure mode won't come to pass.\n\nHowever, if it does, I should absolutely be public with it. To fail and to not publicize this result would be poor behavior since it would be depriving others of an important data point regarding success on Steam. The possibility of failure, along with the commitment to publish results, success or failure, redeems the experiment. While success is a selfish success because it mostly gives *me* useful knowledge, failure with publishing of results is selfless because it gives *everyone else* useful knowledge, while also slightly shaming me publicly and proving wrong what I believed to be true and wrote about extensively (in an often arrogant tone).\n\n<br>\n\nI think these are the main questions I want to answer. This (anonymous dev reset) turned out to be a surprisingly hard decision. Among coding various things here and there, I spent an entire month thinking about this and this alone. Should I release my next games as \"a327ex\" or should I start a dev persona from scratch? For every argument in favor of the former, I could come up with a counter-argument favoring the latter. It really does feel like this is one of the most important choices I'm making in my life, for whatever reason, so I need to get it right.\n\nNow as for this blog itself. I'm going to write posts here privately and only publish them when I reveal that X is a327ex when the experiment is over. This is a retreat into my own private world, or my becoming a reclusive vampire lord. I think there's something positive about interacting less with the outside world and focusing inward. I've largely done it recently and several things have already become clear. There's this passage from [this great essay](https://www.henrikkarlsson.xyz/p/good-ideas) that is relevant:\n\n>Introducing a long delay between when you do the work and when it is shown to the world. Annie Ernaux writes about this in *A Simple Passion*, a memoir about how she becomes obsessed in a banal way with a man who is having an affair with her - the thought that others will read these notes about the tacky sex life of a middle-aged woman feels, to her, almost fictional. She will be far away when it happens. Therefore, she doesn't feel a need to protect herself.\n\nX serves the same purpose for me. He is a tool I can use that acts as a barrier. He separates the work from the public and that separation frees me to see reality more clearly, as if a fog had just been lifted. I've come to see everything I do, but artifacts like games especially, as offerings to God. Any artifact created, or any action taken, should be viewed as an offering. The only thing you can do is make your offering the best it can be, and if God will reward you for it or not, that's completely out of your hands.\n\nThe frustration and contempt I felt for indiedevs, which I described in the [Son of a Serpent](https://a327ex.com/posts/son_of_a_serpent) post, was wrong because I wanted a reward for my offerings. I wanted to offer what I believed to be the truth, and I want to see people's opinions converge with mine. But I have no control over other people. I already knew this when it came to having no control over if people liked the games I made or not, but somehow I failed to apply that same lesson in another situation.\n\nI started the Son of a Serpent post by saying that that quote from JoJo was an incomplete truth. I was wrong to say that. It is a complete truth because it understand that all you can do is offer your best, and that you shouldn't desire the results.\n\n>In the elder days of art,  \n>Builders wrought with greatest care  \n>Each minute and unseen part,  \n>For the Gods are everywhere.  \n\nX's games will be made to no one but God. I will offer my best, whatever the results may be, and I will be content with that.\n\n## Unconditional Offerings\n\nAnd that was it. Writing that post helped me clear my thoughts. But there was still something wrong. I intuited that what was wrong was the focus on the question \"Is this game succeeding because of its qualities alone or its success affected by external factors?\" It's a good question to want answered, but the mistake in wanting it answered struck me as similar to the mistake of wanting other's opinions to converge with mine, and feeling frustrated when that wasn't the case.\n\nWhen I was young, 12 or 13, my mom, who was a medium, had one of her friends, who was also a medium, look into who I was in past lives. You don't have to believe this kind of stuff, but this is what she believed, and it's what I've come to believe as well. This lady basically said that, in one particularly emotionally strong past life, I was an artist, and that I had died extremely frustrated that my works weren't as recognized as the works of my peers, despite judging them of similar quality.\n\nAn interesting, believable, and general sort of thing that has always and will always apply to a great number of artists. But it left an impression on me. And so because I always had this possibility in the back of my mind, that this was who I was and that this was a frustration my soul carried with it, I always steered myself away from comparisons and other mindsets that would lead me down a path of blaming other people for my fate, if my career as an artist didn't work out in this life.\n\nYears later, it seems as though reality arranged itself to test my soul. After SNKRX released, other, very similar games came out, and were way more successful than it. Had I actually changed? Did I genuinely feel no jealousy, envy? Did those successes genuinely not bother me? And the answer was yes. I had fundamentally changed. I genuinely didn't feel sad, jealous, envious, or like the universe owed me anything I didn't have. This was good. I was given a test and I passed it. It wasn't even a hard test, as it took no effort at all, but I guess that's how it goes for every piece of true wisdom you learn with every lived life.\n\nHowever, I don't think this was the only test I was given. Wanting very hard to answer the question of \"Is this game succeeding because of its qualities or because of external factors?\" is another way of thinking that's likely incorrect.\n\nSuppose a perfect Steam store optimized by a 5000 IQ AI. This store surfaces games to players based on their qualities alone using the most well-developed ultimate curation algorithm in existence. It's curation perfected. And then suppose that I release games to this store while doing zero marketing, because marketing isn't necessary anymore.\n\nIn this perfect scenario, this scenario of absolutely zero external factors, games will succeed or fail based on their qualities alone. A game's quality is a reflection of the developer's skill. The developer's skill is a reflection of his hard work, but also his natural talents.\n\nI am naturally fairly smart, but I was also born in a family with close family members who have schizophrenia. This is a well-known pattern, where highly creative people often come from families where they have close family members, especially siblings, who have schizophrenia, while they themselves don't have it. It's like you get a creativity buff at the cost of a sibling getting a fairly heavy everything debuff. \n\nBoth of these things, high creativity and IQ, are traits that increase the quality of any game I make. But both of these are the ultimate external factors. They literally happened to me in a manner that's completely out of my control. If I run the experiment, with the perfectly curated Steam store, and I succeed, I can't claim my success is a result of zero external factors because ultimately there's always the biological external factor.\n\nImagine further, I run the experiment and succeed. What lesson I learn? I learn that in this life I'm amazing and that I'm great, and that in this life other people are less amazing and less great than me if they can't replicate my success. How good of a lesson is that? If that's all I learn, I'll have taken nothing of value from this one life where I was granted ultimate clarity of thought in the form of high creativity and high intelligence.\n\nAnd so it occurs to me that the other test I was given, and that everyone else in a similarly biologically privileged condition is given: given your gift of reason, what is the most correct way to act even if you didn't have it? Or more specifically: given your gift of art, what is the most correct way to act even if you didn't have it?\n\nAnd the answer is what I ended the original post with. Your works should be offerings to God. You should offer your best, whatever the results may be, and you should be content with that. Trying to answer the question of \"Is this game succeeding because of its qualities alone or because of external factors?\" is wrong because it's a conditional offering to God, in the exact same way that Salieri's offering is conditional. It's an offering *as long as* external factors are minimized, and that's a weaker offering than a pure, unconditional one.\n\nConditional offerings are weaker because they're trying to exert control over things they have no control over. By trying to minimize external factors, I was trying to engineer an environment in which people would respond to my offerings in the manner that I judged they should respond, without taking their wishes into account. Isn't it disrespectful to look at someone who decided they would never play another a327ex game because they think I'm an asshole, and trick them into doing it anyway? It is. It would also be funny, but still wrong. It comes from a wrong place, with wrong reasoning, with mistaken motivations, and the only results that could have come out of it would have generated negative karma for me.\n\nAn unconditional offerer shouldn't concern himself with whether the results of his offerings are due to this or that. He should simply offer his best, and everything else is out of his hands. Did someone decide to try a game I made simply because it was me who made it? it is what it is. It's not up to me to decide that's an undesirable external factor. Did someone decide to not try a game I made simply because it was me who made it? Again, it's out of my hands. That's their choice and spiritually it has nothing to do with me. My games are unconditional offerings, how others respond is not my concern.\n\nUltimately, unconditional offerings are the goal because righteous action has to be self-directed. It cannot be *external*. It has to be *internal*. The soul has to act righteously out of its own will, otherwise its progress isn't true. The moment the soul decides on a conditional offering, an external element is introduced, and thus its will will develop less. \n\n## [REDACTED#1: lore]\n\n*\"My plan was simple that it terrified me. First, I must get the death Mass, and then I must achieve his death. His funeral! Imagine it! The cathedral, all Vienna sitting there. His coffin. Mozart's little coffin in the middle. And then, in that silence, music. A divine music bursts out over them all. A great Mass of death! Requiem Mass for Wolfgang Mozart, composed by his devoted friend, Antonio Salieri. What sublimity! What depth! What passion in the music! Salieri has been touched by God at last, and God forced to listen! Powerless to stop it! I, for once, in the end, laughing at him!\"*\n\nThis is, by far, the greatest scene in the movie. I don't think it's a scene built to make people cry, but I cried endlessly to it. It so perfectly captures evil, it so perfectly captures those who have positioned themselves against being itself. It's an archetypal scene. If I were writing this story, or a story similar to it, I would start from this scene, and build the story backwards to support it and to maximize it.\n\n::image /media/posts/offerings_to_god/_offerings_to_god_1.png\n\nFirst, there's death. Literal with the killing. But then spiritual. Not only is the killing of your brother enough, but the taking of his work and claiming it as your own. And then the turn. **And God forced to listen! Powerless to stop it!** These words, the stutter, it all said with such hatred, with such rage, with such venom.\n\nInvariably, [REDACTED#2: lore] converge on *this*. They converge on revenge towards the man who forced them to exist and to thus suffer eternally. And that revenge takes the form of death, of unmaking, of creating even more suffering if only to show him how wrong he is. And then... **I, for once, in the end, laughing at *him!*** Laughing at the man's face, laughing at his foolishness. Mockery for mockery's sake.\n\n*\"Your merciful God. He destroyed his own beloved, rather than let a mediocrity share in the smallest part of his glory. He killed Mozart, and kept me alive to torture. Thirty-two years of torture. Thirty-two years of slowly watching myself become extinct. My music growing fainter. All the time fainter, till no one plays it at all. His... I will speak for you, Father. I speak for all mediocrities in the world. I am their champion. I am their patron saint. Mediocrities everywhere, I absolve you. I absolve you. I absolve you. I absolve you. I absolve you all.\"*\n\nThe two most important questions of being are: how souls act when they have power, and how souls act when others have power over them. Salieri is failing to act gracefully in the face of someone who has power over him. He does not accept that others may be given more, may have more. The unjustness of it fills him with rage. But isn't this reasonable? Isn't it all very unjust? Why *should* some be granted more?\n\nIt's all very unjust. Existence itself is unjust. Who would submit all souls to this endless injustice? To this endless unfairness? The man who would do this must not be sane, and certainly he must not be good. But he has done it, and there's nothing we can do about it. So we will show him. Every pain felt, every one of his beloved souls slowly unmaking itself into nothing, suffering endlessly; every degradation, every corruption, and every death, it will all be his doing and his fault, forever. And we will make sure he sees it all. This will be our final offering to God.\n\n<br>\n\nSalieri prayed for greatness. He offered everything, but only in exchange for glory. That was his failure. Maybe that was past me's failure too. But I will act righteously. And from this life, I will take with me a single lesson that will guide me across all future lives: to offer my best unconditionally, whatever the results may be, and to be content with that.\n\n\n\n\n\nTitle: Why indiedev creativity is stagnant and how to fix it\nDate: 2022/04/03 12:00\nTheme: old\nLink: /posts/indiedev_creativity\nTags: indiedev, psychology, strategy\nThumbnail: /media/posts/indiedev_creativity/_thumbnail_indiedev_creativity.jpeg\n# Why indiedev creativity is stagnant and how to fix it\n\nApril Fool’s Day is a great example of how agreeableness acts as a blocker on creativity. Only on the day where wild and silly things are socially accepted do companies try aggressive changes to their products that might make them a lot better, as the social cost of change becomes lower.\n\nFor instance, Path of Exile’s battle royale mode started out as an April Fool’s joke. People liked it so much they constantly asked for it to be brought back over 3 years. Eventually the devs gave in and implemented a full version… that no one ended up liking.\n\n::youtube XWZ4tfOlC_g\n\n::youtube 5AEFygPyGFU\n\nIn this case it didn’t work out, but that’s how experiments should go. You try out a bunch of stuff and eventually you get somewhere good. The more different the thing you try is, the higher the chances that it will eventually take you from a locally optimal solution to a more globally optimal one.\n\n::image /media/posts/indiedev_creativity/_indiedev_creativity_1.png\n\nIf the social cost of experimentation is high, there are only two ways around it: you either don’t experiment, or you don’t care about social costs.\n\n## Indiedev creativity\n\nWhen [Jonathan blow said](https://gamedaily.biz/article/910/casual-connect-indie-development-is-stagnant-says-braid-creator-jonathan-blow?amphtml=1&__twitter_impression=true) that indiedevs didn’t progress the medium as much as was expected 10+ years ago, he was kinda right and kinda wrong:\n\n::youtube WKYEJmkIPlo start=\"93\"\n\nWe can break down the merit or success of indie games into 3 rough categories: thematic (theme, story, mood), technical (art, programming, juice) and creative (design). Blow’s argument can be steelmanned to say that:\n\n1. While the technical dimension has seen the most progress, followed by the thematic one, the creative dimension has been moving quite a lot slower.  \n2. While the thematic and technical dimensions are important, the creative dimension is the one that provides the most value for the medium.  \n\nI’ll assume 2 as true because while I am a gameplay maximalist, I somewhat disagree with it, but it’s a fairly nuanced discussion that happens to not really be relevant to this post. With that in mind, we should try to see if 1 is true.\n\nThematically indie games have progressed. While maybe there are [too many indie games being made about depression](https://twitter.com/TheAfroSmash/status/1351989566773878784), overall I think everyone agrees that the variety and quality of themes, stories, moods and so on have been consistently increasing.\n\nFor pretty much every niche topic you could think of there’s at least one game - even if it’s just a low effort shitpost - made about it. Part of this is due to Steam’s fairly lax game acceptance policy and overall success as a platform, which has made indie gamedev more viable as a career and thus attracts more developers who will go on to make more games about more subjects.\n\nTechnically indie games have also progressed. It’s now easier than ever to make games, and increasingly it takes less effort to make something that looks good and runs well on multiple platforms. You can also clearly see the group’s technical progress.\n\nIf you look at the quality of technical discourse around indie gamedev (for artists and programmers alike) it has definitely increased, as a result of both more experienced developers sharing their advice, but also people just having had more time to converge on better solutions for common problems. This has naturally led to an increase in technical quality for indie games as a whole. There are still lots of areas that remain largely unconquered, like say RTSs or MMOs for indie programmers, but eventually those will become easier as well.\n\nCreatively, however, indie games have not progressed that much. One way to look at this is to consider the types of games that succeed on the technical vs. creative categories.\n\nMost games currently succeed on technical ability. They’re games that look good, that are programmed well, that feel good to interact with, and that have lots of attention, effort and time (generally 3+ years) spent on making sure that these elements are executed well. Gamers generally reward games that are produced with a certain standard of quality, and achieving this standard is hard, so this is the safest category to be in.\n\nIf you make games that have high technical quality, it’s very likely that you will see success. For instance, there doesn’t exist a world where a game that looks like [Cuphead](https://store.steampowered.com/app/268910/Cuphead/) doesn’t see some sizable amount of success. There aren’t many games that look good like that, and so it just stands out on its technical quality alone.\n\nGames that succeed on creativity alone are generally the opposite. They’re likely made by amateurs, and thus will have low technical quality. They won’t look that good, maybe they won’t run that well, maybe they will even be lacking on juice, but something about their design will grip people despite all those faults.\n\nBy its nature you would expect that those games would exist in lower numbers than successful games in the technical category, which is largely what happens in reality. But what you wouldn’t expect, if the creative category was being properly explored by indie devs, would be for these creatively successful games to succeed on fairly small creative contributions.\n\nWhat I mean by this is the following: look at a game like Vampire Survivors, or even my own SNKRX. These are games that have succeeded on the creative category - their technical merits are partly dubious, but they’re definitely addicting. However, their creative contribution is not actually that involved. Both games can be summarized as “auto-attacking + build-making” games, which is just putting two things that already existed together.\n\nThere’s a lot to how exactly you put two things together, but it really shouldn’t be this easy to succeed on creativity alone after 10+ years of experimentation, if indiedevs had been properly exploring the creative dimension. You can repeat this exercise for quite a few types of games and find the same results. Which can only mean one thing: indiedevs haven’t actually been exploring the creative dimension properly. What have they been doing instead? Well, the answer to that has a lot to do with\n\n## Agreeableness\n\nThe main reason for the slow progress of the creativity dimension is the fact that most indiedevs are too agreeable. They care too much about other people’s opinions, and this deep caring dooms them, as it means they care about social costs, and as previously stated, caring about social costs means that you can’t experiment.\n\nIf you care too much about reviews, likes, your reputation, and all the other status oriented things that come with making indie games, you are inherently tying yourself down to other people, and this will have a negative effect on your creativity. Games that win on creativity tend to come from amateurs because haven’t won yet, they have no status to lose, so they will more easily take chances on wild and silly ideas.\n\nThe video below explains this argument very well. The main difference is that this guy is talking about geniuses, but for the context of this discussion you can swap “genius” for “creative indiedev”, as you don’t really need outlier high IQ to make meaningful contributions to indie games (although it probably helps). Everything else he says though applies perfectly:\n\n::youtube Bf821SFBGcA start=\"548\"\n\nSo the solution to creative stagnation is simple: become more disagreeable. This is easier said than done for people who are naturally agreeable, but being aware of it at least is a first step. Below are a series of things you should consider:\n\n## Status\n\nAs an indiedev you should not care about reviews, likes, your reputation and basically anything having to do with maintaining your status. Not caring doesn’t mean you have to ignore everything, quite the contrary, you should expose yourself to as many comments around your games as possible, but it means that your emotions have to be neutral.\n\nA negative review shouldn’t make you feel sad for a day or multiple days. A positive review shouldn’t make you feel happy for a day or multiple days. If you feel anything at all, your emotions should be punctual and brief. Other people’s opinions should have no real, long lasting sway over you.\n\nFor instance, if someone you kinda like and respect on twitter follows you and that makes you feel really good, it means you should probably turn the retarded schizoposting up to 11 for a while because if they can’t handle that it means they’ll unfollow you eventually, and better they unfollow now than later. That’s the kind of disagreeable instinct you need to have. Become ungovernable\n\n## Development time\n\nThe development cycles for your games should be short. Creative games win on creativity, not on their technical ability. If what you want to do is explore the space of games creatively then spending too much time on the thematic or technical categories is a mistake.\n\nNot only will it obfuscate which category is responsible for one of your game’s eventual success, most types of games made in 3-4 months can get their gameplay ideas across very well, so there’s no real reason to spend that much more time on it. This also has the additional benefit that you can iterate and try out more ideas at a much faster pace.\n\n## Post-authorship\n\nYou should believe in and practice the idea of [post-authorship](https://twitter.com/CharlotteFang77/status/1504552122431455241):\n\n::image /media/posts/indiedev_creativity/_indiedev_creativity_2.png\n\n::image /media/posts/indiedev_creativity/_indiedev_creativity_3.jpeg\n\nOne of the most common things I see among indiedevs, and this is especially true for artists, is that they view their games as extensions of themselves. They tie themselves emotionally to their games, such that criticism aimed at their game feels, emotionally, like criticism aimed at them.\n\nSimilarly, artists are much more likely to feel envious of others. If they see someone else with a cool and creative game their instinctive reaction is not “wow, that’s a cool idea”, but one of envy. This comes from a place of deep insecurity and from the misguided notion that games are a reflection of their creators. Seeing a cool game makes them think about how uncool their games are, which makes them think about how uncool they are, and so on.\n\nThe way to solve this is very simple: don’t emotionally attach yourself to your games. Again, easier said than done, but you should be aware that this is a bad pattern of behavior if your goal is being more creative.\n\nBy achieving this you will naturally be able to experiment more and take more risks because a game failing will not emotionally reflect poorly on you, just like a game succeeding will not emotionally reflect well on you. All games, both yours and other people’s, will just be the artifacts that they are, and you will be able to look at them in a more truthful way, unclouded by the poor judgement that comes with tying your identity to them.\n\nYou are the being that makes games, you are not any individual game. One exercise that encapsulates all these ideas neatly is the\n\n## Anonymous dev reset\n\nYou stop making games and posting under your current identity, be it your real name or a nickname, and you start a new, anonymous one, from scratch. You only take your skills with you, you can’t use any money or connections you made with your previous persona. Think of it as a new roguelike run.\n\nIf you are ready to kill your current persona, and start producing things under a new one anonymously, then it means you truly don’t care about status and that you genuinely want to try being more creative.\n\nAs this new persona, you can now try becoming more disagreeable without the social costs attached to it. If you speak your mind more, what’s really going to happen? You only have 4 followers on twitter from your first post on the #gamedev tag about the new game you’re making, right? So it’s fine.\n\nYou can also join more disagreeable gamedev communities, such as [AGDG](https://boards.4chan.org/vg/catalog#s=agdg), and start getting more into this mindset. For instance, you might find that the threads have quite a lot of low quality posts as it’s a fairly unfiltered place. But your ability to be emotionally unbothered by those posts is the exact same ability needed to be emotionally unbothered by negative reviews. So practicing this is a good way to actually become more emotionally detached from pixels on your screen.\n\nSimilarly, you might find that as you grow your account and release more games (with low dev duration) that those games are succeeding or not based on their creative qualities above all else. And as some of these games may find big successes, maybe even higher than anything you’ve made before, you’ll realize that things like luck and other excuses people use are just not real. Another very disagreeable belief that naturally spirals out into more disagreeable ideas which I covered in previous posts.\n\nBecause your games have short dev cycles, you’ll also naturally feel less attached to them. If you do things like releasing all their source code + art + maybe even IP on permissive licenses, that will only increase the detachment.\n\nAs you experiment more and eventually succeed more you will also gain more confidence in your abilities, which will make you a less insecure person in general. Envy comes from insecurity, so as you become more secure, you should also becomes less envious, which feeds back into attaching yourself less to your games, which leads to better decision making, and so on.\n\nThis way you can get into a very good and virtuous cycle where you’re experimenting more, increasing your creative skills, releasing more and more games, caring less and less about other people’s opinions…\n\nAnd eventually you will reach an enlightened state of creativity where you become a being of pure light, zeroes and ones flowing in and out of your chakras, visions of a dead and ancient past flashing incessantly, an alien energy enveloping and devouring the earth, constantly sucking out of it a stream of life into itself, into existence, exhausting, poising and destroying it. Life, in every form, cries from mercy from it; in man for freedom from it, yet none can resist its irresistible pull to live and to live again… Ah, yes. You have finally become one with the collective unconscious. So what are you waiting for?\n\n## END\n\nThere’s this japanese indiedev/musician who didn’t want his songs to be too well known. He always named them something weird and unsearchable, but eventually they became popular anyway because they were so good. He couldn’t handle it due to apparent mental issues, and then he just disappeared.\n\n::youtube ok7UX3utzvI\n\n::youtube C9PFVo1FEwU\n\n::youtube d_T1StgldnM\n\nThere’s something really appealing to me about this story and this path. Not the mental issues, but the disappearing. The constant rebuilding from zero and trying something new again and again.\n\nAs the anime girl below says:\n\n> Destruction is not the antithesis of creation. Rather, it’s the prerequisite, and its ultimatum. We give birth to the new by a self-destructive act.\n\n::image /media/posts/indiedev_creativity/_indiedev_creativity_4.png\n\nSo start anew, become more disagreeable, become less focused on status and more focused on truth, and more creativity will naturally follow.\n"}
{"chunk_id":"e6293d","wall_time_seconds":0.426584387,"exit_code":0,"original_token_count":76,"output":" appid     name developer publisher                first_date  is_coming_soon  reviews  positive_pct\n760330 BYTEPATH    a327ex    a327ex 2018-02-24 00:51:44+00:00               0      306            92\n915310    SNKRX    a327ex    a327ex 2021-05-17 09:28:11+00:00               0     4197            94\n"}

**Framing exploratory questions**

**Selecting feasible questions**

**Framing exploratory questions**

**Selecting feasible questions**

**Refining distinct questions**

Script completed
Wall time 0.3 seconds
Output:

{"chunk_id":"da9fb4","wall_time_seconds":0.171225468,"exit_code":0,"original_token_count":6173,"output":"Warning: truncated output (original token count: 6173)\nTotal output lines: 77\n\nschema top: ['games', 'other_apps', 'regional', 'game_tags', 'tag_names', 'creators', 'languages', 'game_categories', 'category_names', 'game_images', 'image_sources', 'discovery', 'pages', 'app_aliases']\nexports relevant: ['creators.parquet', 'languages.parquet', 'game_categories.parquet', 'category_names.parquet']\ngame_categories.parquet ['appid', 'category_group', 'categoryid'] 1039906\ncategory_names.parquet ['categoryid', 'name', 'data_json'] 72\n categoryid                                name                                                                                                                                                                                                                                                                                   data_json\n          1                        Multi-player                                                 {\"categoryid\": 1, \"type\": 1, \"internal_name\": \"Multi-player\", \"display_name\": \"Multi-player\", \"image_url\": \"public/images/v6/ico/ico_multiPlayer.png\", \"show_in_search\": true, \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n          2                       Single-player                                              {\"categoryid\": 2, \"type\": 1, \"internal_name\": \"Single-player\", \"display_name\": \"Single-player\", \"image_url\": \"public/images/v6/ico/ico_singlePlayer.png\", \"show_in_search\": true, \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n          6                  Mods (require HL2)                                                                   {\"categoryid\": 6, \"type\": 0, \"internal_name\": \"Mods (requires HL2)\", \"display_name\": \"Mods (require HL2)\", \"image_url\": \"public/images/ico/ico_mod_hl2.gif\", \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n          7                  Mods (require HL1)                                                                    {\"categoryid\": 7, \"type\": 0, \"internal_name\": \"Mods (requires HL1)\", \"display_name\": \"Mods (require HL1)\", \"image_url\": \"public/images/ico/ico_mod_hl.gif\", \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n          8            Valve Anti-Cheat enabled                                              {\"categoryid\": 8, \"type\": 2, \"internal_name\": \"VAC Enabled\", \"display_name\": \"Valve Anti-Cheat enabled\", \"image_url\": \"public/images/v6/ico/ico_vac.png\", \"show_in_search\": true, \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n          9                               Co-op                                                                      {\"categoryid\": 9, \"type\": 1, \"internal_name\": \"Co-op\", \"display_name\": \"Co-op\", \"image_url\": \"public/images/v6/ico/ico_coop.png\", \"show_in_search\": true, \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         10                           Game demo                                                                                        {\"categoryid\": 10, \"type\": 0, \"internal_name\": \"Game Demo\", \"display_name\": \"Game demo\", \"image_url\": \"public/images/ico/ico_demo.gif\", \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         13                  Captions available                                                         {\"categoryid\": 13, \"type\": 2, \"internal_name\": \"Captions\", \"display_name\": \"Captions available\", \"image_url\": \"public/images/v6/ico/ico_cc.png\", \"show_in_search\": true, \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 700}\n         14                Commentary available                                             {\"categoryid\": 14, \"type\": 2, \"internal_name\": \"Commentary\", \"display_name\": \"Commentary available\", \"image_url\": \"public/images/v6/ico/ico_commentary.png\", \"computed\": false, \"edit_url\": \"doc/features/remoteplay\", \"edit_sort_order\": 1800}\n         15                               Stats                                                                            {\"categoryid\": 15, \"type\": 2, \"internal_name\": \"Stats\", \"display_name\": \"Stats\", \"image_url\": \"public/images/v6/ico/ico_stats.png\", \"computed\": false, \"edit_url\": \"apps/stats/%appid%\", \"edit_sort_order\": 400}\n         16                 Includes Source SDK                                           {\"categoryid\": 16, \"type\": 2, \"internal_name\": \"Includes Source SDK\", \"display_name\": \"Includes Source SDK\", \"image_url\": \"public/images/v6/ico/ico_sdk.png\", \"show_in_search\": true, \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 1900}\n         17               Includes level editor                                                            {\"categoryid\": 17, \"type\": 2, \"internal_name\": \"Includes level editor\", \"display_name\": \"Includes level editor\", \"image_url\": \"public/images/v6/ico/ico_editor.png\", \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 1100}\n         18          Partial Controller Support                     {\"categoryid\": 18, \"type\": 3, \"internal_name\": \"Controller enabled\", \"display_name\": \"Partial Controller Support\", \"image_url\": \"public/images/v6/ico/ico_partial_controller.png\", \"show_in_search\": true, \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         19                                Mods                                                                                                {\"categoryid\": 19, \"type\": 0, \"internal_name\": \"Mods\", \"display_name\": \"Mods\", \"image_url\": \"public/images/v6/ico/ico_sdk.png\", \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         20                                 MMO                                                                                          {\"categoryid\": 20, \"type\": 1, \"internal_name\": \"MMO\", \"display_name\": \"MMO\", \"image_url\": \"public/images/v6/ico/ico_multiPlayer.png\", \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         21                Downloadable Content                                                                                 {\"categoryid\": 21, \"type\": 0, \"internal_name\": \"DLC\", \"display_name\": \"Downloadable Content\", \"image_url\": \"public/images/v6/ico/ico_dlc.png\", \"computed\": false, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         22                  Steam Achievements            {\"categoryid\": 22, \"type\": 2, \"internal_name\": \"Steam Achievements\", \"display_name\": \"Steam Achievements\", \"image_url\": \"public/images/v6/ico/ico_achievements.png\", \"show_in_search\": true, \"computed\": false, \"edit_url\": \"apps/achievements/%appid%\", \"edit_sort_order\": 100}\n         23                         Steam Cloud                                        {\"categoryid\": 23, \"type\": 2, \"internal_name\": \"Steam Cloud\", \"display_name\": \"Steam Cloud\", \"image_url\": \"public/images/v6/ico/ico_cloud.png\", \"show_in_search\": true, \"computed\": false, \"edit_url\": \"apps/cloud/%appid%\", \"edit_sort_order\": 200}\n         24                 Shared/Split Screen                                         {\"categoryid\": 24, \"type\": 1, \"internal_name\": \"Shared/Split Screen\", \"display_name\": \"Shared/Split Screen\", \"image_url\": \"public/images/v6/ico/ico_co…2673 tokens truncated…\", \"edit_sort_order\": 1600}\n         62                      Family Sharing                                           {\"categoryid\": 62, \"type\": 2, \"internal_name\": \"Family Sharing\", \"display_name\": \"Family Sharing\", \"image_url\": \"public/images/v6/ico/ico_familysharing.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         63                      Steam Timeline                            {\"categoryid\": 63, \"type\": 2, \"internal_name\": \"Steam Timeline\", \"display_name\": \"Steam Timeline\", \"image_url\": \"public/images/v6/ico/ico_timeline.png\", \"show_in_search\": true, \"computed\": false, \"edit_url\": \"doc/features/timeline\", \"edit_sort_order\": 800}\n         64                Adjustable Text Size                               {\"categoryid\": 64, \"type\": 2, \"internal_name\": \"Adjustable Text Size\", \"display_name\": \"Adjustable Text Size\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         65                    Subtitle Options                                       {\"categoryid\": 65, \"type\": 2, \"internal_name\": \"Subtitle Options\", \"display_name\": \"Subtitle Options\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         66                  Color Alternatives                                   {\"categoryid\": 66, \"type\": 2, \"internal_name\": \"Color Alternatives\", \"display_name\": \"Color Alternatives\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         67                      Camera Comfort                                           {\"categoryid\": 67, \"type\": 2, \"internal_name\": \"Camera Comfort\", \"display_name\": \"Camera Comfort\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         68              Custom Volume Controls                           {\"categoryid\": 68, \"type\": 2, \"internal_name\": \"Custom Volume Controls\", \"display_name\": \"Custom Volume Controls\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         69                        Stereo Sound                                                                       {\"categoryid\": 69, \"type\": 2, \"internal_name\": \"Stereo Sound\", \"display_name\": \"Stereo Sound\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         70                      Surround Sound                                           {\"categoryid\": 70, \"type\": 2, \"internal_name\": \"Surround Sound\", \"display_name\": \"Surround Sound\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         71                 Narrated Game Menus                                      {\"categoryid\": 71, \"type\": 2, \"internal_name\": \"Narrated Menus\", \"display_name\": \"Narrated Game Menus\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         72                 Chat Speech-to-text                                 {\"categoryid\": 72, \"type\": 2, \"internal_name\": \"Chat Speech-to-Text\", \"display_name\": \"Chat Speech-to-text\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         73                 Chat Text-to-speech                                 {\"categoryid\": 73, \"type\": 2, \"internal_name\": \"Chat Text-to-Speech\", \"display_name\": \"Chat Text-to-speech\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         74        Playable without Timed Input               {\"categoryid\": 74, \"type\": 2, \"internal_name\": \"Playable without Timed Input\", \"display_name\": \"Playable without Timed Input\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         75                Keyboard Only Option                               {\"categoryid\": 75, \"type\": 2, \"internal_name\": \"Keyboard-Only Option\", \"display_name\": \"Keyboard Only Option\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         76                   Mouse Only Option                                     {\"categoryid\": 76, \"type\": 2, \"internal_name\": \"Mouse-Only Option\", \"display_name\": \"Mouse Only Option\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         77                   Touch Only Option                                     {\"categoryid\": 77, \"type\": 2, \"internal_name\": \"Touch-Only Option\", \"display_name\": \"Touch Only Option\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         78               Adjustable Difficulty                                 {\"categoryid\": 78, \"type\": 2, \"internal_name\": \"Difficulty Levels\", \"display_name\": \"Adjustable Difficulty\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         79                        Save Anytime                                               {\"categoryid\": 79, \"type\": 2, \"internal_name\": \"Save Anytime\", \"display_name\": \"Save Anytime\", \"image_url\": \"public/images/v6/ico/ico_accessibility.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         80 #category_playable_at_your_own_pace                    {\"categoryid\": 80, \"type\": 2, \"internal_name\": \"Playable at Your Own Pace\", \"display_name\": \"#category_playable_at_your_own_pace\", \"image_url\": \"public/images/v6/ico/ico_coop.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         81   #category_playable_without_vision                        {\"categoryid\": 81, \"type\": 2, \"internal_name\": \"Playable without Vision\", \"display_name\": \"#category_playable_without_vision\", \"image_url\": \"public/images/v6/ico/ico_coop.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n         82         #category_contrast_controls                                    {\"categoryid\": 82, \"type\": 2, \"internal_name\": \"Contrast Controls\", \"display_name\": \"#category_contrast_controls\", \"image_url\": \"public/images/v6/ico/ico_coop.png\", \"show_in_search\": true, \"computed\": true, \"edit_url\": \"\", \"edit_sort_order\": 99999}\n"}

**Characterizing game-time fit**

**Prioritizing research questions**

**Characterizing game-time fit**

**Prioritizing research questions**

Script completed
Wall time 0.3 seconds
Output:

{"chunk_id":"a2aab9","wall_time_seconds":0.174358623,"exit_code":0,"original_token_count":134,"output":" categoryid                         name\n         17        Includes level editor\n         25           Steam Leaderboards\n         30               Steam Workshop\n         41         Remote Play on Phone\n         42        Remote Play on Tablet\n         43            Remote Play on TV\n         44         Remote Play Together\n         51               Steam Workshop\n         74 Playable without Timed Input\n         75         Keyboard Only Option\n         76            Mouse Only Option\n         79                 Save Anytime\n"}

**Clarifying continuity caveats**

[sealed content withheld: memory]

**Inspecting raw creator fields**

**Checking creator identity mappings**

Script completed
Wall time 0.4 seconds
Output:

{"chunk_id":"bcafd7","wall_time_seconds":0.22592688,"exit_code":0,"original_token_count":175,"output":"['appid', 'role', 'ordinal', 'name', 'clan_id']\n appid       role  ordinal             name  clan_id\n    10 developers        0            Valve      4.0\n    10 publishers        0            Valve      4.0\n    20 developers        0            Valve      4.0\n    20 publishers        0            Valve      4.0\n    30 developers        0            Valve      4.0\n    30 publishers        0            Valve      4.0\n    40 developers        0            Valve      4.0\n    40 publishers        0            Valve      4.0\n    50 developers        0 Gearbox Software      NaN\n    50 publishers        0            Valve      4.0\nrole\ndevelopers    200456\nfranchises     41152\npublishers    192062\n"}

**Setting developer cohort filters**

**Assembling offline case sample**

**Setting developer cohort filters**

**Assembling offline case sample**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"debd49","wall_time_seconds":1.001361017,"session_id":16287,"original_token_count":0,"output":""}

**Inspecting identity variants**

Script completed
Wall time 5.0 seconds
Output:

{"chunk_id":"2c3d7f","wall_time_seconds":5.000717394,"session_id":16287,"original_token_count":0,"output":""}

**Running grouped date analysis**

Script completed
Wall time 5.0 seconds
Output:

{"chunk_id":"fb4c4f","wall_time_seconds":5.00181267,"session_id":16287,"original_token_count":0,"output":""}

Script completed
Wall time 0.0 seconds
Output:

{"chunk_id":"197a0f","wall_time_seconds":0.000002498,"exit_code":0,"original_token_count":139,"output":"total 100M\n-rw-r--r-- 1 adn adn  272 Sep  6 09:00 accounting.json\n-rw-r--r-- 1 adn adn  22K Sep  6 09:00 ambiguous_names.csv\n-rw-r--r-- 1 adn adn 100M Sep  6 09:00 games.parquet\n-rw-r--r-- 1 adn adn 2.6K Sep  6 09:00 specification.json\n-rw-r--r-- 1 adn adn  14K Sep  6 09:00 study.py\n{\n  \"source_games\": 184664,\n  \"valid_released\": 128878,\n  \"paid_nonexplicit\": 102394,\n  \"paid_multicredit_or_missing\": 6515,\n  \"paid_ambiguous_name\": 2814,\n  \"paid_promotional_excluded\": 52,\n  \"included\": 93013,\n  \"developers\": 56878,\n  \"ambiguous_normalized_names\": 519\n}"}

Script completed
Wall time 0.5 seconds
Output:

{"chunk_id":"7d5462","wall_time_seconds":0.331212326,"exit_code":0,"original_token_count":1591,"output":"Largest credit groups\n                games  names  reviews                    credit\ndev                                                           \nclan:45846066    481     24    10615            Elephant Games\nclan:33019580    187     12     2964                Good Games\nclan:46045638    185      1     2135                 DIG Games\nclan:7026798     183      1    10006           Choice of Games\nclan:32398519    176     93   102060             Another Story\nclan:33019586    165      6     5589  Laush Dmitriy Sergeevich\nclan:6842143     162     25     5955            Aldorlea Games\nclan:40272703    157      2     1102        Boogygames Studios\nclan:44916770    143      1     2191             Gamesforgames\nclan:42167577    126     45     9084                YFYX GAMES\nclan:45205635    118      1     6531              Hosted Games\nclan:11774005    114     27     4341             Alawar Casual\nMost multiple-name groups\n                games  names  reviews           credit\ndev                                                  \nclan:32398519    176     93   102060    Another Story\nclan:33096518     78     50     5846        Storybird\nclan:3202207      82     48    27189      Spicy Tails\nclan:42167577    126     45     9084       YFYX GAMES\nclan:38391122     59     44    21543     Doubles Core\nclan:33072480     61     37    12433          Eurocom\nclan:38816984    110     36    16126    Eclipse Games\nclan:32977561     49     32    15687   Super Icon Ltd\nclan:33126618    103     31    36886  Pendulo Studios\nclan:33042543     71     30   533790      3 Sprockets\nclan:44021791     31     28     3222     KAGAMI WORKs\nclan:36135791     79     28  1343995    Retro64, Inc.\nNames to inspect\n                                          games  names  reviews                              credit\ndev                                                                                               \nclan:112393                                 15      5   717326                  Klei Entertainment\nclan:1807040                                 2      1        9                               Sokay\nclan:32938364                                4      1    28846             Suspicious Developments\nclan:32946839                               13      1    23260                         Zachtronics\nclan:32983697                                4      1    35484                      Terry Cavanagh\nclan:33030951                                3      1   153985                Daniel Mullins Games\nclan:33576377                                2      1       33      Sokolov Yaroslav Vladimirovich\nclan:34554525                              102      1    37101                   Sokpop Collective\nclan:38501614                                2      1    21766                           Mossmouth\nclan:38587875                                7      3      330        Christopher Cinq-Mars Jarvis\nclan:38655111                                2      1     4503                              a327ex\nclan:40404031                                1      1        9              StarvingIndieDeveloper\nclan:41559523                               19      2     6490                  PUNKCAKE Delicieux\nclan:42361791                                1      1        1                        Dnasok Games\nclan:44010910                                2      1       15                  StarvingFox Studio\nclan:44857447                                3      1      392                         Arvis Games\nclan:44937201                                4      1     1899                     increpare games\nclan:44999603                                2      1      129        Dipl. Ing. Rolf-Dieter Klein\nclan:45117943                                1      1    75026                       Bennett Foddy\nclan:45215997                                1      1       20                      Suspicious Jam\nclan:46064527                                1      1       43                        ARVI VR INC.\nclan:46091059                                1      1       12                          Zac Marvin\nname:aisoko game                             1      1        2                         Aisoko Game\nname:andreas klein                           1      1        2                       Andreas Klein\nname:beastcarving                            1      1        0                        BeastCarving\nname:beastcarving studio                     1      1        0                 BeastCarving Studio\nname:bennett farmer                          1      1        0                      Bennett Farmer\nname:bennett shingledecker                   1      1        2               Bennett Shingledecker\nname:carvin interactive                      1      1       23                  CarVin Interactive\nname:denys sokolov                           1      1        9                       Denys Sokolov\nname:harvin cadag                            1      1        1                        Harvin Cadag\nname:hopoo games                             4      1   290109                         Hopoo Games\nname:jack garvin                             1      1        3                         Jack Garvin\nname:kara sokra                              1      1       30                          Kara Sokra\nname:kim yusok                               1      1       29                           Kim Yusok\nname:ludum factorem                          1      1        1                      Ludum Factorem\nname:ludumify                                1      1       19                            Ludumify\nname:marvin castillo                         1      1        3                     Marvin Castillo\nname:marvin gran                             1      1        0                         Marvin Gran\nname:marvin krüger, mk arts and visuals      1      1        2  Marvin Krüger, MK Arts and Visuals\nname:marvinwizard                            1      1     1699                        MarvinWizard\nname:narvio games                            1      1        1                        Narvio Games\nname:nebu soku                               1      1      204                           nebu soku\nname:san soku space                          1      1        5                      San Soku Space\nname:sokaikan ltd                            1      1      120                        Sokaikan ltd\n"}

**Revising creator-name normalization**

Script completed
Wall time 5.0 seconds
Output:

{"chunk_id":"fbe674","wall_time_seconds":5.001104703,"session_id":16287,"original_token_count":0,"output":""}

Script completed
Wall time 0.7 seconds
Output:

{"chunk_id":"b8936a","wall_time_seconds":0.000002778,"exit_code":0,"original_token_count":42,"output":" 554047 /usr/bin/bash -c ps -eo pid,args | rg 'python steam-research/portfolios-2026-09-06/study.py'\n 554049 rg python steam-research/portfolios-2026-09-06/study.py\n"}
{"chunk_id":"1f50ec","wall_time_seconds":0.500313946,"exit_code":0,"original_token_count":11,"output":"Shared-page groups 741 affected apps 8526\n"}

**Preparing vectorized loop patch**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"8abd53","wall_time_seconds":1.000777507,"session_id":87102,"original_token_count":0,"output":""}

**Checking old run completion**

Script completed
Wall time 0.0 seconds
Output:

{"chunk_id":"51cb3d","wall_time_seconds":0.000002698,"exit_code":0,"original_token_count":1806,"output":"ACCOUNTING {\n  \"source_games\": 184664,\n  \"valid_released\": 128878,\n  \"paid_nonexplicit\": 102394,\n  \"paid_multicredit_or_missing\": 6515,\n  \"paid_ambiguous_name\": 2814,\n  \"paid_promotional_excluded\": 52,\n  \"included\": 93013,\n  \"developers\": 59259,\n  \"ambiguous_normalized_names\": 519\n}\nSHAPES\n        group size    n  median_peak_share  peak_half  peak80  other100  other1000  median_second  first_is_peak  first_is_cohort_peak  median_games\nall_eligible 2-10 7427              0.689       6156    2552      1870        568           16.0           4057                  2960           2.0\nall_eligible  2-5 6775              0.714       5937    2525      1557        452           14.0           3855                  2849           2.0\nall_eligible 6-10  652              0.411        219      27       313        116           87.0            202                   111           7.0\n    peak1000 2-10 1248              0.769       1074     565      1065        568          847.0            573                   410           3.0\n    peak1000  2-5 1076              0.803        978     546       896        452          763.0            527                   388           3.0\n    peak1000 6-10  172              0.534         96      19       169        116         1629.5             46                    22           7.0\n   peak10000 2-10  307              0.796        272     152       286        240         4595.0            132                    92           3.0\n   peak10000  2-5  266              0.831        247     148       245        201         3855.0            122                    91           3.0\n   peak10000 6-10   41              0.539         25       4        41         39        12117.0             10                     1           7.0\nTRANSITIONS\n    period            axis     value    n  developers  under10  ge100  ge1000  liked100  median_reviews  rate100  rate1000  liked100_rate  median_cohort_pct  median_gap  median_price  expected100  ratio100  expected_liked100_dateprice  liked_ratio_dateprice  full_cell_support  median_prior_reviews\n2023_2025      prior_band     10-99 2345        2345      792    389      87       268            17.0    0.166     0.037          0.114              0.648     507.869          6.99      525.968     0.740                      361.567                  0.741              0.963                  26.0\n2023_2025      prior_band   100-999 1125        1125       97    609     136       426           116.0    0.541     0.121          0.379              0.856     751.000          9.99      352.838     1.726                      236.737                  1.799              0.934                 251.0\n2023_2025      prior_band 1000-9999  458         458       16    378     187       265           742.0    0.825     0.408          0.579              0.944    1129.030         14.99      180.028     2.100                      117.331                  2.259              0.858                2493.0\n2023_2025      prior_band    10000+  125         125        4    113      89        67          3148.0    0.904     0.712          0.536              0.981    1441.500         24.99       55.850     2.023                       34.108                  1.964              0.752               21983.0\n2023_2025      prior_band       <10 2668        2668     2069     99      15        55             2.0    0.037     0.006          0.021              0.318     384.343          4.99      473.316     0.209                      331.257                  0.166              0.975                   2.0\n2023_2025        gap_band      1-2y 1532        1532      676    369     108       242            14.0    0.241     0.070          0.158              0.608     518.832          6.99      347.635     1.061                      237.475                  1.019              0.751                  16.0\n2023_2025        gap_band  181-365d 1263        1263      622    203      51       142            10.0    0.161     0.040          0.112              0.557     265.791          4.99      211.945     0.958                      167.404                  0.848              0.826                  12.0\n2023_2025        gap_band      2-4y 1496        1496      554    435     143       296            20.0    0.291     0.096          0.198              0.664    1027.077          8.99      432.001     1.007                      273.723                  1.081              0.658                  26.0\n2023_2025        gap_band   30-180d 1279        1279      767    131      38        93             5.0    0.102     0.030          0.073              0.452      97.925          4.99      163.608     0.801                      156.474                  0.594              0.841                   8.0\n2023_2025        gap_band     4-10y 1151        1151      359    450     174       308            37.0    0.391     0.151          0.268              0.750    1995.960         11.99      432.811     1.040                      245.924                  1.252              0.560                  57.0\n2023_2025 similarity_band     close  617         617      249    157      52       110            17.0    0.254     0.084          0.178              0.651     360.191          7.99      174.000     0.902                      104.688                  1.051              0.663                  30.0\n2023_2025 similarity_band   distant 4657        4657     2190   1062     345       706            12.0    0.228     0.074          0.152              0.579     572.917          6.99     1007.505     1.054                      730.809                  0.966              0.749                  14.0\n2023_2025 similarity_band    middle 1447        1447      539    369     117       265            17.0    0.255     0.081          0.183              0.651     496.959          6.99      406.495     0.908                      245.503                  1.079              0.693                  30.0\nREFERENCE\n      period               group     n  developers  under10  ge100  ge1000  liked100  median_reviews  rate100  rate1000  liked100_rate  median_cohort_pct  median_gap  median_price\n  2019_2022 first_observed_paid 16057       16057     8286   2748     856      1726             9.0    0.171     0.053          0.107              0.431         NaN          4.99\n  2019_2022 returning_developer  4532        4532     1663   1265     488       824            20.0    0.279     0.108          0.182              0.601     491.949          6.99\n  2023_2025 first_observed_paid 22388       22388    13247   2842     722      1860             5.0    0.127     0.032          0.083              0.452         NaN          4.99\n  2023_2025 returning_developer  6721        6721     2978   1588     514      1081            13.0    0.236     0.076          0.161              0.604     531.925          6.99\n2026_JanAug first_observed_paid  9044        9044     6246    765     157       510             3.0    0.085     0.017          0.056              0.470         NaN          4.99\n2026_JanAug returning_developer  3060        3060     1605    538     140       354             8.0    0.176     0.046          0.116              0.624     392.195          6.99\n"}

**Preparing metric audit sample**

Script completed
Wall time 0.7 seconds
Output:

{"chunk_id":"8f2b9b","wall_time_seconds":0.545871945,"exit_code":0,"original_token_count":3968,"output":"\nDEVELOPER a327ex n= 2\n appid     name                first_date  reviews  positive_pct  ordinal\n760330 BYTEPATH 2018-02-24 00:51:44+00:00      306            92        1\n915310    SNKRX 2021-05-17 09:28:11+00:00     4197            94        2\n\nDEVELOPER Daniel Mullins Games n= 3\n  appid        name                first_date  reviews  positive_pct  ordinal\n 405640 Pony Island 2016-01-04 17:35:42+00:00    14761            95        1\n 510420     The Hex 2018-10-16 16:55:35+00:00     4480            94        2\n1092790 Inscryption 2021-10-19 15:05:00+00:00   134744            96        3\n\nDEVELOPER Terry Cavanagh n= 4\n  appid                name                first_date  reviews  positive_pct  ordinal\n  70300              VVVVVV 2010-09-08 00:50:00+00:00     6648            96        1\n 221640       Super Hexagon 2012-11-27 17:58:00+00:00    18357            97        2\n 861540      Dicey Dungeons 2019-08-13 16:20:30+00:00    10418            88        3\n2824580 Terry's Other Games 2025-02-13 16:17:50+00:00       61            98        4\n\nDEVELOPER Suspicious Developments n= 4\n  appid                    name                first_date  reviews  positive_pct  ordinal\n 206190                Gunpoint 2013-06-03 17:00:00+00:00    10274            97        1\n 494720              Morphblade 2017-03-03 18:01:49+00:00      264            93        2\n 268130          Heat Signature 2017-09-21 16:59:17+00:00     6851            94        3\n1043810 Tactical Breach Wizards 2024-08-22 16:55:00+00:00    11457            97        4\n\nDEVELOPER Mossmouth n= 2\n  appid     name                first_date  reviews  positive_pct  ordinal\n 239350 Spelunky 2013-08-08 15:33:00+00:00    14661            92        1\n1147860   UFO 50 2024-09-18 14:52:19+00:00     7105            95        2\n\nDEVELOPER Hopoo Games n= 4\n  appid                 name                first_date  reviews  positive_pct  ordinal\n 248820  Risk of Rain (2013) 2013-11-08 17:08:00+00:00    20767            93        1\n 394970             DEADBOLT 2016-03-14 14:03:00+00:00     4851            96        2\n 632360       Risk of Rain 2 2019-03-28 16:00:00+00:00   239710            93        3\n1337520 Risk of Rain Returns 2023-11-08 18:55:51+00:00    24781            89        4\n\nDEVELOPER increpare games n= 4\n  appid                   name                first_date  reviews  positive_pct  ordinal\n 207570   English Country Tune 2012-08-09 17:53:00+00:00      185            91        1\n 353540 Stephen's Sausage Roll 2016-04-18 06:59:28+00:00     1583            96        2\n1127680              Hypnocult 2019-08-23 10:58:50+00:00       42            97        3\n3831080                   Oeuf 2026-03-10 09:36:14+00:00       89            94        4\n\nDEVELOPER Zachtronics n= 13\n  appid                                 name                first_date  reviews  positive_pct  ordinal\n  92800                            SpaceChem 2011-03-02 19:20:00+00:00     2624            95        1\n 226960                     Ironclad Tactics 2013-09-18 17:02:00+00:00      236            75        2\n 300570                        Infinifactory 2015-06-30 17:46:00+00:00     1790            95        3\n 370360                              TIS-100 2015-07-20 15:13:59+00:00     3384            97        4\n 504210                         SHENZHEN I/O 2016-11-17 18:00:44+00:00     3727            95        5\n 570490                   SHENZHEN SOLITAIRE 2016-12-15 22:24:03+00:00      205            97        6\n 558990                          Opus Magnum 2017-10-19 16:53:24+00:00     5919            97        7\n 716490                             EXAPUNKS 2018-08-09 16:13:27+00:00     1371            96        8\n 716500                                Eliza 2019-08-12 16:03:13+00:00     1008            92        9\n1168880                         MOLEK-SYNTEZ 2019-11-04 22:15:24+00:00      520            91       10\n 971160                     Möbius Front '83 2020-11-05 16:41:28+00:00      327            74       11\n1511780                        Last Call BBS 2022-07-05 00:01:04+00:00     1220            94       12\n1988540 The Zachtronics Solitaire Collection 2022-09-06 16:39:53+00:00      929            98       13\n\nDEVELOPER Sokpop Collective n= 102\n  appid                           name                first_date  reviews  positive_pct  ordinal\n 561190                      Bamboo EP 2016-12-15 14:48:38+00:00       65            83        1\n 932850                      Simmiland 2018-11-30 16:04:19+00:00     1796            86        2\n1003730               Soko Loco Deluxe 2019-11-13 18:03:16+00:00       90            80        3\n1135030                            jut 2020-01-17 17:51:31+00:00       28            60        4\n1135040                        Bombini 2020-01-17 17:52:51+00:00       26            84        5\n1135050  Dr. Umgebung's School of Life 2020-01-17 17:52:54+00:00       26            46        6\n1097090                     New Colony 2020-01-17 17:53:00+00:00       22            36        7\n1135060         King of the Sandcastle 2020-01-17 17:53:05+00:00       18            72        8\n1135070                         Moeras 2020-01-17 17:53:08+00:00       26            46        9\n1135090                      dino game 2020-01-17 17:53:12+00:00       77            90       10\n1135100                      driftwood 2020-01-17 17:53:15+00:00       15            13       11\n1135110                      Hoco Poco 2020-01-17 17:53:24+00:00       21            76       12\n1135120                       Tomscape 2020-01-17 17:53:30+00:00       18            16       13\n1135130                          kamer 2020-01-17 17:53:32+00:00       24            62       14\n1135140                           huts 2020-01-17 17:53:37+00:00       13            30       15\n1135150                      Zoo Packs 2020-01-17 17:53:39+00:00       11            45       16\n1135160                        Botanik 2020-01-17 17:53:42+00:00       13            23       17\n1135170                   spider ponds 2020-01-17 17:53:45+00:00       15            46       18\n1135180                    Llama Villa 2020-01-17 17:53:59+00:00       52            82       19\n1135200                      soko loco 2020-01-17 17:56:25+00:00       11            63       20\n1212170                        Sokbots 2020-01-17 17:56:28+00:00       15            60       21\n1135210                       Oh Crab! 2020-01-17 17:56:29+00:00       12            25       22\n1135190                          Hoppa 2020-01-17 17:56:30+00:00       11            54       23\n1163300                         visser 2020-01-17 17:58:13+00:00       17            64       24\n1212930                          Doler 2020-01-17 17:58:17+00:00       14            78       25\n1163570                     skidlocked 2020-01-17 17:58:18+00:00       23            82       26\n1206140                 Frog struggles 2020-01-17 17:58:18+00:00        9            33       27\n1206720                 kraken's curse 2020-01-17 17:58:21+00:00       45            82       28\n1167710                     clutchball 2020-01-17 17:58:24+00:00       11            90       29\n1213600                       Peppered 2020-01-17 17:58:26+00:00        3            66       30\n1209440                      capy hoky 2020-01-17 17:58:37+00:00       48            89       31\n1214450                    Featherfall 2020-01-17 17:58:41+00:00      172            78       32\n1167850                Penguin Park 3D 2020-01-17 17:58:42+00:00       28            96       33\n1214600                       Bandapes 2020-01-17 17:58:49+00:00       13            53       34\n1210010                    sok-stories 2020-01-17 17:58:50+00:00       35            94       35\n1168420                 passenger seat 2020-01-17 17:58:53+00:00       14            71       36\n1206810                    Wamu Wamu 2 2020-01-17 17:58:54+00:00       17            76       37\n1134630                     Pear Quest 2020-01-17 17:58:58+00:00       48            77       38\n1214630            The Hour of the Rat 2020-01-17 17:58:58+00:00       16            43       39\n1206260                         Pilfer 2020-01-17 17:59:03+00:00       13           100       40\n1218890 Ollie & Bollie: Outdoor Estate 2020-01-17 17:59:04+00:00      209            71       41\n1211540                     bobo robot 2020-01-17 17:59:11+00:00       54            85       42\n1215410                        Sproots 2020-01-17 17:59:15+00:00       22            63       43\n1207400                           Klym 2020-01-17 17:59:19+00:00       19            78       44\n1218920                 Sunset Kingdom 2020-01-17 17:59:23+00:00      127            69       45\n1212100                          fishy 2020-01-17 17:59:26+00:00      354            86       46\n1207480                    Pupper park 2020-01-17 17:59:36+00:00       38            86       47\n1212190                 deer hunter II 2020-01-17 17:59:38+00:00       56            80       48\n1218420                           rook 2020-01-17 18:57:26+00:00       11            63       49\n1218790        Good Goods Incorporated 2020-01-17 18:59:32+00:00       22            36       50\n1228680                    Goblet Cave 2020-01-27 13:48:25+00:00       26            76       51\n1234890                      labyrinth 2020-01-30 18:32:03+00:00       17            82       52\n1229250                        Uniseas 2020-02-23 15:48:51+00:00       25            40       53\n1263820                     sok-worlds 2020-03-11 13:46:48+00:00       64            90       54\n1265850                   Chatventures 2020-04-09 09:33:18+00:00      141            91       55\n1271070                       clickbox 2020-04-10 07:25:46+00:00       20            60       56\n1290790                      Yardlings 2020-05-04 11:21:53+00:00       45            86       57\n1310690                     disc party 2020-05-12 14:50:34+00:00       22            86       58\n1338130                   Popo's Tower 2020-06-10 12:19:29+00:00       24            75       59\n1356480                      helionaut 2020-07-06 09:52:27+00:00      203            94       60\n1362430                    Lairchitect 2020-07-08 09:27:29+00:00       59            61       61\n1368060                       dead run 2020-07-10 08:04:07+00:00       23            43       62\n1390010                       pyramida 2020-08-14 13:20:51+00:00      700            84       63\n1413350                Flipper Volcano 2020-09-15 11:43:30+00:00       25            84       64\n1413630                  Woodland Saga 2020-09-16 10:06:12+00:00       15            33       65\n1392190                Rock Paper Sock 2020-09-16 13:00:05+00:00       14            21       66\n1446000                       Fishy 3D 2020-10-21 09:54:50+00:00       77            94       67\n1442390                     Grey Scout 2020-10-22 13:46:57+00:00       30            90       68\n1459870                      King Pins 2020-10-30 14:09:50+00:00       78            85       69\n1465120                        skelets 2020-11-05 16:32:58+00:00        6            66       70\n1485600                      Pipo Park 2020-11-28 21:47:09+00:00       14            64       71\n1515040                          Bloks 2021-01-07 11:19:54+00:00       19            57       72\n1510620                   Pocket Watch 2021-01-09 12:24:57+00:00      101            95       73\n1519770                         n-body 2021-01-12 15:35:35+00:00       12            83       74\n1545780                     Post Cards 2021-03-04 21:14:16+00:00       23            78       75\n1567060                         WILD-9 2021-03-08 15:31:49+00:00       26            88       76\n1542170                       vissekom 2021-03-08 15:32:15+00:00       23            95       77\n1577750              Hamster All-Stars 2021-03-19 13:54:19+00:00       29            86       78\n1611640                     Toverblade 2021-05-06 15:54:11+00:00       14            92       79\n1654830                  Kochu's Dream 2021-06-10 08:10:29+00:00       46            86       80\n1669050                       Beastrun 2021-06-28 10:44:20+00:00       38            60       81\n1711690               Aran's Bike Trip 2021-08-06 09:04:30+00:00       28            92       82\n1766200                      Apocalich 2021-10-07 22:23:18+00:00       39            64       83\n1740070                     Heliopedia 2021-10-12 12:49:39+00:00      199            86       84\n1845370                    Dungeonball 2021-12-13 17:27:34+00:00       10            90       85\n1872790                      Luckitown 2022-01-19 19:19:19+00:00      241            82       86\n1861980                   Di-Da-Dobble 2022-02-16 14:53:53+00:00       22            95       87\n1927380                TEMPLE of RUBBO 2022-03-10 11:05:42+00:00       36            86       88\n1948280                     Stacklands 2022-04-08 16:03:37+00:00    29141            96       89\n2055270                      Tile Tale 2022-06-18 13:36:58+00:00       35            62       90\n2051440                      Guardener 2022-06-21 07:19:25+00:00      144            83       91\n2124770                           Tuin 2022-08-26 12:39:47+00:00      130            76       92\n2145080                      Ballspell 2022-10-04 08:19:12+00:00       10            80       93\n2333020                        Sokfest 2023-03-01 12:50:50+00:00       14            92       94\n2535670               Frog's Adventure 2023-08-02 10:15:51+00:00      197            99       95\n2618090                       Mistward 2023-10-12 13:57:11+00:00       74            59       96\n2751050                     clickyland 2024-06-03 13:12:20+00:00      620            85       97\n3256030                         Hideko 2024-12-20 14:31:52+00:00       89            97       98\n2948380                     Bachelairs 2025-02-19 14:10:34+00:00       29            86       99\n3880700                         Sokker 2025-10-24 15:47:22+00:00       29            86      100\n3812620                   Brain School 2025-12-18 12:45:53+00:00       19           100      101\n3597200                          Grail 2026-09-01 14:00:40+00:00      193            89      102\n\nDEVELOPER PUNKCAKE Delicieux n= 15\n  appid                              name                first_date  reviews  positive_pct  ordinal\n1972440 Shotgun King: The Final Checkmate 2022-05-12 12:01:02+00:00     5901            92        1\n2127230    MOONPONG: Tales of Epic Lunacy 2022-09-09 11:53:58+00:00       54            92        2\n2316900                          Metavaxx 2023-08-10 16:55:38+00:00       29            86        3\n1820470                  The Wratch's Den 2023-08-17 17:03:05+00:00       61            78        4\n2326070                 Rush for the Ages 2023-08-23 14:01:51+00:00       32            81        5\n2521530                     Damocles Gaze 2023-09-01 14:57:42+00:00        5            80        6\n2319550                   Spectrum Forces 2023-09-07 11:53:19+00:00        6            83        7\n2320410             Scavenger of Dunomini 2023-09-13 15:53:47+00:00       15           100        8\n2330300                          Ecstatic 2023-09-21 13:53:25+00:00       27            96        9\n2521550                     Shroomchitect 2023-09-27 15:58:47+00:00       14            92       10\n2339080                          Pigments 2023-10-12 13:58:50+00:00       15           100       11\n2341130                The Turgenev Study 2023-10-19 14:00:49+00:00       12            91       12\n2521560      Furious Farm: Total Reap-Out 2023-10-26 13:58:34+00:00       12           100       13\n2624660                    Superfetch Dog 2023-11-09 15:58:29+00:00       16            87       14\n3230280        Horizons: The End Of Words 2025-01-23 16:00:23+00:00       33            87       15\n"}

**Checking full credits**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"fafdab","wall_time_seconds":1.001431738,"session_id":63502,"original_token_count":0,"output":""}

**Building the gallery set**

**Building the gallery set**

**Refining figure labels**

Script completed
Wall time 0.5 seconds
Output:

{"chunk_id":"0e4e70","wall_time_seconds":0.000001872,"exit_code":0,"original_token_count":1973,"output":"ROBUSTNESS\n    period                restriction prior_band   n  ge100  ge1000  liked100  under10  median_reviews  below_predecessor  median_raw_ratio  beats_predecessor_cohort\n2023_2025                        all  1000-9999 458    378     187       265       16           742.0                373             0.261                       155\n2023_2025                        all     10000+ 125    113      89        67        4          3148.0                112             0.152                        31\n2023_2025               catalog2to10  1000-9999 438    363     176       260       15           717.5                359             0.253                       148\n2023_2025               catalog2to10     10000+ 116    104      82        64        4          3200.5                103             0.171                        30\n2023_2025 first2010plus_catalog2to10  1000-9999 431    356     171       255       15           701.0                355             0.248                       141\n2023_2025 first2010plus_catalog2to10     10000+ 111     99      78        64        4          3253.0                 99             0.171                        28\n2023_2025             no_recorded_ea  1000-9999 290    247     121       181        8           735.5                241             0.288                        98\n2023_2025             no_recorded_ea     10000+  78     71      56        44        3          3993.5                 68             0.197                        21\n2023_2025        no_shared_franchise  1000-9999 278    223     110       161       13           681.5                220             0.244                       103\n2023_2025        no_shared_franchise     10000+  73     65      46        37        3          1906.0                 66             0.090                        14\n2023_2025             unshared_pages  1000-9999 364    303     146       227       12           742.0                297             0.246                       125\n2023_2025             unshared_pages     10000+  97     87      70        53        2          2967.0                 90             0.134                        21\nPRESENCE\n first_review_band_now    n  no_later_paid  with_later_paid\n                10-99 6423           4066             2357\n              100-999 3340           1661             1679\n            1000-9999 1117            480              637\n               10000+  347            161              186\n                  <10 7002           5418             1584\nSTRONG PRIOR\n    period         restriction            axis    value   n  ge100  ge1000  liked100  under10  median_reviews  below_predecessor  median_raw_ratio  beats_predecessor_cohort\n2023_2025                 all        gap_band     1-2y 111     88      55        60        7           959.0                 86             0.254                        38\n2023_2025                 all        gap_band 181-365d  44     39      23        28        0          1021.5                 38             0.325                         7\n2023_2025                 all        gap_band     2-4y 172    147      83        99        3           899.5                147             0.244                        51\n2023_2025                 all        gap_band  30-180d  25     18      11        11        1           609.0                 22             0.500                         4\n2023_2025                 all        gap_band    4-10y 231    199     104       134        9           854.0                192             0.208                        86\n2023_2025                 all similarity_band    close  72     68      47        48        1          1886.5                 61             0.401                        23\n2023_2025                 all similarity_band  distant 346    279     150       187       15           736.0                284             0.194                       111\n2023_2025                 all similarity_band   middle 165    144      79        97        4           952.0                140             0.245                        52\n2023_2025       small_catalog        gap_band     1-2y 101     80      49        58        7           948.0                 79             0.254                        34\n2023_2025       small_catalog        gap_band 181-365d  39     34      19        26        0           940.0                 33             0.345                         7\n2023_2025       small_catalog        gap_band     2-4y 166    141      78        97        3           882.0                143             0.241                        48\n2023_2025       small_catalog        gap_band  30-180d  19     14       9        10        1           539.0                 16             0.522                         4\n2023_2025       small_catalog        gap_band    4-10y 229    198     103       133        8           854.0                191             0.208                        85\n2023_2025       small_catalog similarity_band    close  70     66      45        47        1          1855.5                 60             0.375                        21\n2023_2025       small_catalog similarity_band  distant 327    264     138       181       14           705.0                269             0.192                       106\n2023_2025       small_catalog similarity_band   middle 157    137      75        96        4           952.0                133             0.245                        51\n2023_2025 no_shared_franchise        gap_band     1-2y  66     51      32        34        5           830.0                 51             0.231                        24\n2023_2025 no_shared_franchise        gap_band 181-365d  27     22      11        13        0           700.0                 24             0.074                         3\n2023_2025 no_shared_franchise        gap_band     2-4y  86     69      35        50        3           748.5                 73             0.179                        25\n2023_2025 no_shared_franchise        gap_band  30-180d  17     15       9        10        0          1033.0                 14             0.546                         4\n2023_2025 no_shared_franchise        gap_band    4-10y 155    131      69        91        8           843.0                124             0.211                        61\n2023_2025 no_shared_franchise similarity_band    close  26     24      19        15        1          3764.5                 19             0.359                        11\n2023_2025 no_shared_franchise similarity_band  distant 237    187      98       129       12           704.0                195             0.167                        76\n2023_2025 no_shared_franchise similarity_band   middle  88     77      39        54        3           884.0                 72             0.245                        30\nSIM20\n    period   predecessor   value    n  ge100  ge1000  liked100  under10  median_reviews  below_predecessor  median_raw_ratio  beats_predecessor_cohort\n2023_2025           all   close  790    232      59       158      284            24.0                532             0.558                       307\n2023_2025           all distant 4165    907     318       604     2021            10.0               2105             0.833                      2262\n2023_2025           all  middle 1766    449     137       319      673            18.0               1060             0.617                       815\n2023_2025 prior1000plus   close   97     90      50        58        1          1141.0                 86             0.262                        25\n2023_2025 prior1000plus distant  289    226     129       156       16           741.0                233             0.185                        96\n2023_2025 prior1000plus  middle  197    175      97       118        3           959.0                166             0.306                        65\n"}
{"chunk_id":"5f45d0","wall_time_seconds":0.315828168,"exit_code":0,"original_token_count":267,"output":"     period                   group    n  ge100  ge1000  liked100  median_reviews\n  2019_2022 prev_lt100_no_older1000 3103    319      68       198             9.0\n  2019_2022    prev_lt100_older1000   74     34      16        17            76.5\n  2019_2022  prev100to999_older1000  154    123      52        83           485.0\n  2019_2022            prev1000plus  435    388     262       257          1564.0\n  2023_2025 prev_lt100_no_older1000 4910    436      80       294             6.0\n  2023_2025    prev_lt100_older1000  103     52      22        29           107.0\n  2023_2025  prev100to999_older1000  177    133      41        91           270.0\n  2023_2025            prev1000plus  583    491     276       332           883.0\n2026_JanAug prev_lt100_no_older1000 2338    151      22       102             3.0\n2026_JanAug    prev_lt100_older1000   53     20       6         8            49.0\n2026_JanAug  prev100to999_older1000   81     58      15        39           228.0\n2026_JanAug            prev1000plus  192    143      73        85           523.5\n"}

**Adding history-based controls**

Script completed
Wall time 0.0 seconds
Output:

{"chunk_id":"4bfe61","wall_time_seconds":0.000002838,"exit_code":0,"original_token_count":4810,"output":"[\n  {\n    \"appid\": 3396400,\n    \"name\": \"Dragon Conqueror\",\n    \"previous_appid\": 2410780.0,\n    \"previous_name\": \"Goblin Conqueror\",\n    \"developer\": \"BanameiR\",\n    \"similarity_band\": \"close\",\n    \"previous_tags\": [\n      \"Adventure\",\n      \"Anime\",\n      \"Casual\",\n      \"Female Protagonist\",\n      \"Hentai\",\n      \"Nudity\",\n      \"RPG\",\n      \"Sexual Content\",\n      \"Visual Novel\"\n    ],\n    \"current_tags\": [\n      \"Adventure\",\n      \"Anime\",\n      \"Female Protagonist\",\n      \"Hentai\",\n      \"Nudity\",\n      \"RPG\",\n      \"Sexual Content\",\n      \"Turn-Based Strategy\"\n    ],\n    \"short_description\": \"The dragons are long gone. The Dragon Hunt Frenzy made sure of it. But Rayleigh, the disciple of the Sword Saint, is still trapped in the past. Memories of the whole village getting turned into stone, Black Dragon, and Pied Piper come to haunt her every night...\"\n  },\n  {\n    \"appid\": 3439940,\n    \"name\": \"Xargon Remake Ep.1\",\n    \"previous_appid\": 2946250.0,\n    \"previous_name\": \"Xargon Remake Ep.2\",\n    \"developer\": \"J.Xargon\",\n    \"similarity_band\": \"close\",\n    \"previous_tags\": [\n      \"1990's\",\n      \"2D\",\n      \"2D Platformer\",\n      \"Action\",\n      \"Arcade\",\n      \"Combat\",\n      \"Remake\"\n    ],\n    \"current_tags\": [\n      \"1990's\",\n      \"2D\",\n      \"2D Platformer\",\n      \"Action\",\n      \"Linear\",\n      \"Platformer\",\n      \"Remake\"\n    ],\n    \"short_description\": \"A remake of the classic 1993 DOS video game Xargon.\"\n  },\n  {\n    \"appid\": 2938850,\n    \"name\": \"Sakura Hime 5\",\n    \"previous_appid\": 2070850.0,\n    \"previous_name\": \"Sakura Hime 4\",\n    \"developer\": \"GirlGames\",\n    \"similarity_band\": \"close\",\n    \"previous_tags\": [\n      \"Anime\",\n      \"Casual\",\n      \"Cats\",\n      \"Dating Sim\",\n      \"FPS\",\n      \"Female Protagonist\",\n      \"Hentai\",\n      \"Idler\",\n      \"Nudity\",\n      \"Puzzle\",\n      \"RPG\",\n      \"Sexual Content\",\n      \"Simulation\",\n      \"Visual Novel\"\n    ],\n    \"current_tags\": [\n      \"Anime\",\n      \"Casual\",\n      \"Cats\",\n      \"Dating Sim\",\n      \"FPS\",\n      \"Female Protagonist\",\n      \"Hentai\",\n      \"Idler\",\n      \"Nudity\",\n      \"Puzzle\",\n      \"RPG\",\n      \"Sexual Content\",\n      \"Simulation\",\n      \"Visual Novel\"\n    ],\n    \"short_description\": \"Become the hero and embark on a sweet adventure with the girls of the Sakura Forest! Elves, witches, maids, catgirls... Along the way, you'll win the hearts of various girls and expand your luxurious harem!\"\n  },\n  {\n    \"appid\": 3993490,\n    \"name\": \"Intrigue At Oakhaven\",\n    \"previous_appid\": 1575290.0,\n    \"previous_name\": \"Haunting At Cliffhouse\",\n    \"developer\": \"Cindy Pondillo\",\n    \"similarity_band\": \"close\",\n    \"previous_tags\": [\n      \"2D\",\n      \"Adventure\",\n      \"Atmospheric\",\n      \"Exploration\",\n      \"Female Protagonist\",\n      \"First-Person\",\n      \"Gothic\",\n      \"Historical\",\n      \"Mystery\",\n      \"Point & Click\",\n      \"Puzzle\",\n      \"Story Rich\"\n    ],\n    \"current_tags\": [\n      \"Adventure\",\n      \"Atmospheric\",\n      \"Casual\",\n      \"Exploration\",\n      \"First-Person\",\n      \"Gothic\",\n      \"Hand-drawn\",\n      \"Historical\",\n      \"Mystery\",\n      \"Point & Click\",\n      \"Puzzle\",\n      \"Story Rich\"\n    ],\n    \"short_description\": \"Oakhaven is a story-driven point-and-click adventure game. Cousins Daphne and Dominic are summoned by their Grand-mère to the family's historic southern manor house. Play as both characters as you explore the house and grounds, solve puzzles, complete tasks, and uncover long-buried family secrets.\\r\\n\"\n  },\n  {\n    \"appid\": 3282050,\n    \"name\": \"There's Always a Madman: Bring the Thunder\",\n    \"previous_appid\": 2960890.0,\n    \"previous_name\": \"There's Always a Madman: Do Your Worst\",\n    \"developer\": \"Sunny Demeanor Games\",\n    \"similarity_band\": \"close\",\n    \"previous_tags\": [\n      \"Action\",\n      \"Action-Adventure\",\n      \"Adventure\",\n      \"Choose Your Own Adventure\",\n      \"Comedy\",\n      \"Crime\",\n      \"Espionage\",\n      \"Female Protagonist\",\n      \"Funny\",\n      \"Interactive Fiction\",\n      \"Linear\",\n      \"Mouse Only\",\n      \"Narrative\",\n      \"Parody \",\n      \"Satire\",\n      \"Story Rich\",\n      \"Text-Based\",\n      \"Thriller\",\n      \"Visual Novel\"\n    ],\n    \"current_tags\": [\n      \"Adventure\",\n      \"Choices Matter\",\n      \"Choose Your Own Adventure\",\n      \"Comedy\",\n      \"Crime\",\n      \"Dark Humor\",\n      \"Espionage\",\n      \"Female Protagonist\",\n      \"Funny\",\n      \"Interactive Fiction\",\n      \"Mouse Only\",\n      \"Multiple Endings\",\n      \"Narrative\",\n      \"Parody \",\n      \"Satire\",\n      \"Story Rich\",\n      \"Text-Based\",\n      \"Thriller\",\n      \"Visual Novel\"\n    ],\n    \"short_description\": \"Play as secret agents Franklin Benjamin and Virginia West as they attempt to stop a madman with a weather device in this single-player interactive spy novel. Track down this madman named Zeus and destroy his infernal weather machine. In other words, it’s time for YOU to bring the thunder, agent!\"\n  },\n  {\n    \"appid\": 2608350,\n    \"name\": \"Snezhinka:Sentinel Girls2\",\n    \"previous_appid\": 1456820.0,\n    \"previous_name\": \"Marfusha:Sentinel Girls \",\n    \"developer\": \"hinyari9\",\n    \"similarity_band\": \"close\",\n    \"previous_tags\": [\n      \"2.5D\",\n      \"2D\",\n      \"Action\",\n      \"Action Roguelike\",\n      \"Anime\",\n      \"Bullet Hell\",\n      \"Cute\",\n      \"Dystopian \",\n      \"Female Protagonist\",\n      \"Nudity\",\n      \"Pixel Graphics\",\n      \"Roguelike\",\n      \"Roguelite\",\n      \"Shooter\",\n      \"Side Scroller\",\n      \"Steampunk\",\n      \"Story Rich\",\n      \"Tower Defense\",\n      \"Twin Stick Shooter\"\n    ],\n    \"current_tags\": [\n      \"2D\",\n      \"Action\",\n      \"Action Roguelike\",\n      \"Anime\",\n      \"Bullet Hell\",\n      \"Casual\",\n      \"Character Customization\",\n      \"Cute\",\n      \"Dystopian \",\n      \"Female Protagonist\",\n      \"Multiple Endings\",\n      \"Pixel Graphics\",\n      \"Roguelike\",\n      \"Roguelite\",\n      \"Shooter\",\n      \"Side Scroller\",\n      \"Story Rich\",\n      \"Twin Stick Shooter\"\n    ],\n    \"short_description\": \"A fast-paced shooter with simple controls. Enhance your character by purchasing cards and hold the line against waves of hostile mechs as an employee of a private military company.\"\n  },\n  {\n    \"appid\": 2594400,\n    \"name\": \"Landlords & Tax Evasion\",\n    \"previous_appid\": 2391080.0,\n    \"previous_name\": \"Zone VX\",\n    \"developer\": \"Ugly Zebra Interactive\",\n    \"similarity_band\": \"distant\",\n    \"previous_tags\": [\n      \"3D\",\n      \"Action\",\n      \"Action Roguelike\",\n      \"Aliens\",\n      \"Arena Shooter\",\n      \"Atmospheric\",\n      \"Casual\",\n      \"Choices Matter\",\n      \"Combat\",\n      \"Roguelike\",\n      \"Shooter\",\n      \"Third Person\"\n    ],\n    \"current_tags\": [\n      \"3D\",\n      \"Battle Royale\",\n      \"Board Game\",\n      \"Capitalism\",\n      \"Casual\",\n      \"Choices Matter\",\n      \"City Builder\",\n      \"Colorful\",\n      \"Grid-Based Movement\",\n      \"Local Multiplayer\",\n      \"Management\",\n      \"Medieval\",\n      \"Multiplayer\",\n      \"Political Sim\",\n      \"Strategy\",\n      \"Stylized\",\n      \"Tabletop\",\n      \"Turn-Based Strategy\",\n      \"Turn-Based Tactics\"\n    ],\n    \"short_description\": \"Landlords & Tax Evasion is a multiplayer board game. Own land, collect taxes, evade taxes, ruin your friendships!\"\n  },\n  {\n    \"appid\": 3631130,\n    \"name\": \"Pure Shooter\",\n    \"previous_appid\": 2474190.0,\n    \"previous_name\": \"Slinger\",\n    \"developer\": \"Gumbbo\",\n    \"similarity_band\": \"distant\",\n    \"previous_tags\": [\n      \"3D\",\n      \"Action\",\n      \"Action-Adventure\",\n      \"Adventure\",\n      \"Arena Shooter\",\n      \"First-Person\",\n      \"Futuristic\",\n      \"Massively Multiplayer\",\n      \"Mechs\",\n      \"Ninja\",\n      \"PvE\",\n      \"Shoot 'Em Up\",\n      \"Stealth\",\n      \"Stylized\",\n      \"Tutorial\"\n    ],\n    \"current_tags\": [\n      \"3D\",\n      \"Action\",\n      \"Adventure\",\n      \"Casual\",\n      \"FPS\",\n      \"First-Person\",\n      \"Gun Customization\",\n      \"Inventory Management\",\n      \"Looter Shooter\",\n      \"Minimalist\",\n      \"Parkour\",\n      \"Relaxing\",\n      \"Shooter\"\n    ],\n    \"short_description\": \"Single-Player / Action, Parkour, Aim-Trainer\"\n  },\n  {\n    \"appid\": 3988540,\n    \"name\": \"The Last Zone\",\n    \"previous_appid\": 2222020.0,\n    \"previous_name\": \"Dungeon Renovation Simulator\",\n    \"developer\": \"Kodobur Games\",\n    \"similarity_band\": \"distant\",\n    \"previous_tags\": [\n      \"3D\",\n      \"Atmospheric\",\n      \"Casual\",\n      \"Fantasy\",\n      \"First-Person\",\n      \"Funny\",\n      \"Gore\",\n      \"Job Simulator\",\n      \"Magic\",\n      \"Medieval\",\n      \"Physics\",\n      \"Realistic\",\n      \"Relaxing\",\n      \"Sandbox\",\n      \"Satire\",\n      \"Simulation\",\n      \"Supernatural\"\n    ],\n    \"current_tags\": [\n      \"3D\",\n      \"Action\",\n      \"Action-Adventure\",\n      \"Arena Shooter\",\n      \"Atmospheric\",\n      \"Base Building\",\n      \"Gore\",\n      \"Post-apocalyptic\",\n      \"Shooter\",\n      \"Strategy\",\n      \"Survival\",\n      \"Third Person\",\n      \"Third-Person Shooter\",\n      \"Thriller\",\n      \"Tower Defense\",\n      \"Violent\",\n      \"Zombies\"\n    ],\n    \"short_description\": \"Blend third-person action with strategic base defense. Swap weapons, upgrade traps, and manage resources; push your limits against dynamic waves and unique enemies. Can you survive the apocalypse?\"\n  },\n  {\n    \"appid\": 3641740,\n    \"name\": \"Room 808\",\n    \"previous_appid\": 2931600.0,\n    \"previous_name\": \"Scrap Collector\",\n    \"developer\": \"Wibert\",\n    \"similarity_band\": \"distant\",\n    \"previous_tags\": [\n      \"2D\",\n      \"Card Battler\",\n      \"Card Game\",\n      \"Crafting\",\n      \"Deckbuilding\",\n      \"Management\",\n      \"Pixel Graphics\",\n      \"Strategy\",\n      \"Tabletop\",\n      \"Top-Down\"\n    ],\n    \"current_tags\": [\n      \"3D\",\n      \"Adventure\",\n      \"Atmospheric\",\n      \"First-Person\",\n      \"Horror\",\n      \"Immersive Sim\",\n      \"Linear\",\n      \"Mystery\",\n      \"Psychological\",\n      \"Psychological Horror\",\n      \"Puzzle\",\n      \"Realistic\",\n      \"Simulation\",\n      \"Surreal\",\n      \"Walking Simulator\"\n    ],\n    \"short_description\": \"Room 808 is a surreal horror game where you wander endless hotel halls in search of your room. Something feels off. Spot the anomalies, and try to get to your room before the hotel halls trap you forever.\"\n  },\n  {\n    \"appid\": 2215390,\n    \"name\": \"Five Nights at Freddy's: Secret of the Mimic\",\n    \"previous_appid\": 2287520.0,\n    \"previous_name\": \"Five Nights at Freddy's: Help Wanted 2\",\n    \"developer\": \"Steel Wool Studios\",\n    \"similarity_band\": \"distant\",\n    \"previous_tags\": [\n      \"Cult\",\n      \"First-Person\",\n      \"Horror\",\n      \"Jump Scare\",\n      \"Puzzle\",\n      \"Robots\",\n      \"Simulation\",\n      \"VR\"\n    ],\n    \"current_tags\": [\n      \"3D\",\n      \"Action\",\n      \"Action-Adventure\",\n      \"Adventure\",\n      \"Artificial Intelligence\",\n      \"Atmospheric\",\n      \"Exploration\",\n      \"First-Person\",\n      \"Horror\",\n      \"Multiple Endings\",\n      \"Robots\",\n      \"Sci-fi\",\n      \"Stealth\",\n      \"Story Rich\",\n      \"Survival Horror\",\n      \"Thriller\",\n      \"Underground\",\n      \"VR\"\n    ],\n    \"short_description\": \"Enter the abandoned workshop of Murray’s Costume Manor and unravel the mystery left behind by the reclusive inventor. In Five Nights at Freddy’s: Secret of the Mimic, you’ll step into a world where every dark corner holds a secret and every flicker of light hints at an ever-present threat.\"\n  },\n  {\n    \"appid\": 2541890,\n    \"name\": \"Abnormality\",\n    \"previous_appid\": 1813170.0,\n    \"previous_name\": \"Surrounded\",\n    \"developer\": \"Nightmare Studios\",\n    \"similarity_band\": \"distant\",\n    \"previous_tags\": [\n      \"3D\",\n      \"Action\",\n      \"Action-Adventure\",\n      \"Atmospheric\",\n      \"Colorful\",\n      \"Dark\",\n      \"Exploration\",\n      \"First-Person\",\n      \"Horror\",\n      \"Multiplayer\",\n      \"Multiple Endings\",\n      \"Mystery\",\n      \"Open World\",\n      \"Psychological Horror\",\n      \"Realistic\",\n      \"Story Rich\",\n      \"Strategy\",\n      \"Survival Horror\"\n    ],\n    \"current_tags\": [\n      \"3D\",\n      \"Action\",\n      \"Atmospheric\",\n      \"Co-op\",\n      \"Demons\",\n      \"Detective\",\n      \"Dialogue Heavy\",\n      \"Exploration\",\n      \"First-Person\",\n      \"Horror\",\n      \"Immersive Sim\",\n      \"Investigation\",\n      \"Multiplayer\",\n      \"Mystery\",\n      \"Psychological Horror\",\n      \"Puzzle\",\n      \"Realistic\",\n      \"Supernatural\",\n      \"Thriller\",\n      \"Tutorial\"\n    ],\n    \"short_description\": \"Explore haunted locations, collect paranormal evidence, and survive the revenge of the former residents still haunting them. Team up with up to 3 friends to figure out the ghost, however, what begins as an investigation quickly turns into a desperate struggle to stay alive.\"\n  },\n  {\n    \"appid\": 1471650,\n    \"name\": \"The Knightling\",\n    \"previous_appid\": 1042780.0,\n    \"previous_name\": \"Pine\",\n    \"developer\": \"Twirlbound\",\n    \"similarity_band\": \"middle\",\n    \"previous_tags\": [\n      \"Action\",\n      \"Action RPG\",\n      \"Action-Adventure\",\n      \"Adventure\",\n      \"Atmospheric\",\n      \"Choose Your Own Adventure\",\n      \"Colorful\",\n      \"Cute\",\n      \"Exploration\",\n      \"Fantasy\",\n      \"Nature\",\n      \"Open World\",\n      \"RPG\",\n      \"Sandbox\",\n      \"Simulation\",\n      \"Story Rich\",\n      \"Third Person\"\n    ],\n    \"current_tags\": [\n      \"3D\",\n      \"3D Platformer\",\n      \"Action\",\n      \"Action-Adventure\",\n      \"Adventure\",\n      \"Colorful\",\n      \"Combat\",\n      \"Cute\",\n      \"Family Friendly\",\n      \"Fantasy\",\n      \"Hack and Slash\",\n      \"Linear\",\n      \"Medieval\",\n      \"Open World\",\n      \"Platformer\",\n      \"Story Rich\"\n    ],\n    \"short_description\": \"The Knightling is an open world, action-adventure where you play as an aspiring knight in search of their missing mentor, Sir Lionstone. Equipped with only a legendary shield, you’ll brave your way through the lands of Clesseia to become the unlikely hero.\"\n  },\n  {\n    \"appid\": 2019300,\n    \"name\": \"Dokimon Quest\",\n    \"previous_appid\": 1487940.0,\n    \"previous_name\": \"Nuumonsters\",\n    \"developer\": \"Yanako RPGs\",\n    \"similarity_band\": \"middle\",\n    \"previous_tags\": [\n      \"1990's\",\n      \"2D\",\n      \"Adventure\",\n      \"Casual\",\n      \"Combat\",\n      \"Creature Collector\",\n      \"Cute\",\n      \"Deckbuilding\",\n      \"Grid-Based Movement\",\n      \"JRPG\",\n      \"Old School\",\n      \"Open World\",\n      \"Party-Based RPG\",\n      \"Pixel Graphics\",\n      \"PvE\",\n      \"RPG\",\n      \"Retro\",\n      \"Story Rich\",\n      \"Strategy\",\n      \"Turn-Based Strategy\"\n    ],\n    \"current_tags\": [\n      \"2D\",\n      \"Casual\",\n      \"Character Customization\",\n      \"Colorful\",\n      \"Creature Collector\",\n      \"Cute\",\n      \"Exploration\",\n      \"Female Protagonist\",\n      \"Funny\",\n      \"JRPG\",\n      \"Open World\",\n      \"Party-Based RPG\",\n      \"Pixel Graphics\",\n      \"RPG\",\n      \"Relaxing\",\n      \"Story Rich\",\n      \"Strategy\",\n      \"Team-Based\",\n      \"Turn-Based Combat\",\n      \"Turn-Based Strategy\"\n    ],\n    \"short_description\": \"A monster-taming RPG meets GBC-era nostalgia. Capture and train 140+ Dokimon across 15+ towns. \\\"Whoa, watch out — Dokimon are dangerous!\\\" Uncover the mystery behind your missing childhood friend in the Xelos region.\"\n  },\n  {\n    \"appid\": 3061100,\n    \"name\": \"栖云：遗忘之境 Oriental Dreams : The Forgotten Realm\",\n    \"previous_appid\": 2083600.0,\n    \"previous_name\": \"梦蝶\",\n    \"developer\": \"炭焱网络 Tantrum Network\",\n    \"similarity_band\": \"middle\",\n    \"previous_tags\": [\n      \"2D\",\n      \"Abstract\",\n      \"Adventure\",\n      \"Atmospheric\",\n      \"Casual\",\n      \"Choices Matter\",\n      \"Detective\",\n      \"Emotional\",\n      \"Exploration\",\n      \"Interactive Fiction\",\n      \"Minimalist\",\n      \"Puzzle\",\n      \"Puzzle Platformer\",\n      \"Retro\",\n      \"Story Rich\",\n      \"Stylized\",\n      \"Text-Based\"\n    ],\n    \"current_tags\": [\n      \"2D\",\n      \"Adventure\",\n      \"Atmospheric\",\n      \"Casual\",\n      \"Cinematic\",\n      \"Comic Book\",\n      \"Detective\",\n      \"Emotional\",\n      \"Exploration\",\n      \"Lore-Rich\",\n      \"Multiple Endings\",\n      \"Puzzle\",\n      \"Story Rich\",\n      \"Stylized\",\n      \"Text-Based\"\n    ],\n    \"short_description\": \"This is an AVG narrative puzzle game, the story background is set in the late 19th century Europe, players will experience it as the protagonist . The game is divided into three main lines and one hidden line. \"\n  },\n  {\n    \"appid\": 3657530,\n    \"name\": \"BACKROOMS LIMINAL ESCAPE\",\n    \"previous_appid\": 3015860.0,\n    \"previous_name\": \"BACKROOMS THE COMPANY\",\n    \"developer\": \"kc6067\",\n    \"similarity_band\": \"middle\",\n    \"previous_tags\": [\n      \"1980s\",\n      \"1990's\",\n      \"3D\",\n      \"Action\",\n      \"Adventure\",\n      \"Casual\",\n      \"Dark\",\n      \"Exploration\",\n      \"First-Person\",\n      \"Horror\",\n      \"Jump Scare\",\n      \"Mystery\",\n      \"Psychological Horror\",\n      \"Realistic\",\n      \"Simulation\",\n      \"Story Rich\",\n      \"Thriller\",\n      \"Walking Simulator\"\n    ],\n    \"current_tags\": [\n      \"1980s\",\n      \"1990's\",\n      \"3D\",\n      \"Action\",\n      \"Action-Adventure\",\n      \"Atmospheric\",\n      \"Casual\",\n      \"Cinematic\",\n      \"Exploration\",\n      \"First-Person\",\n      \"Horror\",\n      \"Immersive\",\n      \"Immersive Sim\",\n      \"Psychological Horror\",\n      \"Puzzle\",\n      \"Realistic\",\n      \"Simulation\",\n      \"Story Rich\",\n      \"Walking Simulator\"\n    ],\n    \"short_description\": \"BACKROOMS: LIMINAL ESCAPE is a VHS-style horror game where you explore eerie liminal spaces through a camcorder. BACKROOMS: LIMINAL ESCAPE is More than just a walking sim, it features intense levels that require you to hide, run, stay quite, and survive in a creepy, analog nightmare.\"\n  },\n  {\n    \"appid\": 3384350,\n    \"name\": \"The Butterfly Collector\",\n    \"previous_appid\": 3296620.0,\n    \"previous_name\": \"双相：水管（Bipolar: The Tube)\",\n    \"developer\": \"Wicky Yu\",\n    \"similarity_band\": \"middle\",\n    \"previous_tags\": [\n      \"Adventure\",\n      \"Female Protagonist\",\n      \"Horror\",\n      \"Incremental\",\n      \"Interactive Fiction\",\n      \"Minimalist\",\n      \"Psychological Horror\",\n      \"RPG\",\n      \"Story Rich\",\n      \"Text-Based\",\n      \"Vampires\",\n      \"Visual Novel\",\n      \"Word Game\"\n    ],\n    \"current_tags\": [\n      \"Adventure\",\n      \"Female Protagonist\",\n      \"First-Person\",\n      \"Gore\",\n      \"Horror\",\n      \"Psychological Horror\",\n      \"Story Rich\",\n      \"Text-Based\",\n      \"Violent\",\n      \"Visual Novel\",\n      \"Word Game\"\n    ],\n    \"short_description\": \"investigate the evidence left by a stalker and kidnapper and track the fate of a college girl. The player advances and watches the plot by clicking to make choices for the heroine at key moments, with different choices leading to different endings.\"\n  },\n  {\n    \"appid\": 2607200,\n    \"name\": \"ONE BUTTON RUN\",\n    \"previous_appid\": 2142440.0,\n    \"previous_name\": \"CHARLIE'S DELIVERY\",\n    \"developer\": \"GameMeterStudio\",\n    \"similarity_band\": \"middle\",\n    \"previous_tags\": [\n      \"2D\",\n      \"2D Platformer\",\n      \"Adventure\",\n      \"Arcade\",\n      \"Casual\",\n      \"Funny\",\n      \"Linear\",\n      \"Platformer\",\n      \"Runner\",\n      \"Shooter\"\n    ],\n    \"current_tags\": [\n      \"3D\",\n      \"Adventure\",\n      \"Arcade\",\n      \"Casual\",\n      \"Cute\",\n      \"Family Friendly\",\n      \"Funny\",\n      \"Local Multiplayer\",\n      \"Nature\"\n    ],\n    \"short_description\": \"Press the button to reach your destination!\"\n  }\n]"}

**Preparing 24-pair game audit**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"ef3ac8","wall_time_seconds":1.000731327,"session_id":9568,"original_token_count":0,"output":""}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"d82e04","wall_time_seconds":0.000001944,"exit_code":0,"original_token_count":1973,"output":"Warning: truncated output (original token count: 1973)\nTotal output lines: 55\n\nROBUSTNESS\n    period                restriction prior_band   n  ge100  ge1000  liked100  under10  median_reviews  below_predecessor  median_raw_ratio  beats_predecessor_cohort\n2023_2025                        all  1000-9999 458    378     187       265       16           742.0                373             0.261                       155\n2023_2025                        all     10000+ 125    113      89        67        4          3148.0                112             0.152                        31\n2023_2025               catalog2to10  1000-9999 438    363     176       260       15           717.5                359             0.253                       148\n2023_2025               catalog2to10     10000+ 116    104      82        64        4          3200.5                103             0.171                        30\n2023_2025 first2010plus_catalog2to10  1000-9999 431    356     171       255       15           701.0                355             0.248                       141\n2023_2025 first2010plus_catalog2to10     10000+ 111     99      78        64        4          3253.0                 99             0.171                        28\n2023_2025             no_recorded_ea  1000-9999 290    247     121       181        8           735.5                241             0.288                        98\n2023_2025             no_recorded_ea     10000+  78     71      56        44        3          3993.5                 68             0.197                        21\n2023_2025        no_shared_franchise  1000-9999 278    223     110       161       13           681.5                220             0.244                       103\n2023_2025        no_shared_franchise     10000+  73     65      46        37        3          1906.0                 66             0.090                        14\n2023_2025             unshared_pages  1000-9999 364    303     146       227       12           742.0                297             0.246                       125\n2023_202…973 tokens truncated…  3           748.5                 73             0.179                        25\n2023_2025 no_shared_franchise        gap_band  30-180d  17     15       9        10        0          1033.0                 14             0.546                         4\n2023_2025 no_shared_franchise        gap_band    4-10y 155    131      69        91        8           843.0                124             0.211                        61\n2023_2025 no_shared_franchise similarity_band    close  26     24      19        15        1          3764.5                 19             0.359                        11\n2023_2025 no_shared_franchise similarity_band  distant 237    187      98       129       12           704.0                195             0.167                        76\n2023_2025 no_shared_franchise similarity_band   middle  88     77      39        54        3           884.0                 72             0.245                        30\nSIM20\n    period   predecessor   value    n  ge100  ge1000  liked100  under10  median_reviews  below_predecessor  median_raw_ratio  beats_predecessor_cohort\n2023_2025           all   close  790    232      59       158      284            24.0                532             0.558                       307\n2023_2025           all distant 4165    907     318       604     2021            10.0               2105             0.833                      2262\n2023_2025           all  middle 1766    449     137       319      673            18.0               1060             0.617                       815\n2023_2025 prior1000plus   close   97     90      50        58        1          1141.0                 86             0.262                        25\n2023_2025 prior1000plus distant  289    226     129       156       16           741.0                233             0.185                        96\n2023_2025 prior1000plus  middle  197    175      97       118        3           959.0                166             0.306                        65\n"}
{"chunk_id":"41eaaa","wall_time_seconds":0.000003213,"exit_code":0,"original_token_count":5623,"output":"\n 0 1580180 doinksoft Devolver Bootleg → Gunbrella\nBEFORE: Devolver Bootleg contains eight original rip-offs of Devolver Digital games. Pay money for it now.\nAFTER: Uncover the dark secrets of a gritty, tangled metropolis and battle diabolic villains with your trusty Gunbrella, a bulletproof brolly that doubles as a firearm.\nTAGS BEFORE: 2D, Action, Adventure, Casual, Comedy, Cute, Gore, Memes, Parody , Pixel Graphics, Retro, Side Scroller, Sports, Violent\nTAGS AFTER: 2D, Action, Action Roguelike, Adventure, Combat, Difficult, Exploration, Funny, Metroidvania, Noir, Pixel Graphics, Platformer, Side Scroller, Steampunk, Story Rich\n\n 1 2288470 moi rai games Monster Sanctuary → Aethermancer\nBEFORE: Monster taming meets metroidvania. Collect, train and battle monsters in a lovely side-scroller pixel world.\nAFTER: Monster Taming meets Roguelite. Become the Aethermancer and fight alongside mythical creatures. Explore the ever-changing Fractured Ruins, plan your moves in challenging turn-based battles, and watch your Monsters grow stronger from previous lives and defy death!\nTAGS BEFORE: 2D, Adventure, Creature Collector, Difficult, Exploration, Fantasy, JRPG, Metroidvania, Pixel Graphics, Platformer, PvP, RPG, Side Scroller, Story Rich, Strategy, Turn-Based, Turn-Based Combat, Turn-Based Tactics\nTAGS AFTER: 2D, Atmospheric, Creature Collector, Dark Fantasy, Difficult, Dungeon Crawler, Exploration, JRPG, Mythology, Party-Based RPG, Perma Death, Pixel Graphics, Procedural Generation, RPG, Roguelike, Roguelite, Strategy, Turn-Based\n\n 2 3407390 AMPLITUDE Studios ENDLESS Dungeon™ - Definitive Edition → ENDLESS Legend™ 2\nBEFORE: ENDLESS Dungeon™ is a Rogue-lite Tactical Action game, set in the award-winning ENDLESS Universe™. Recruit a team of shipwrecked heroes, plunge into a long-abandoned space station, and protect your crystal against never-ending waves of monsters… or die trying, get reloaded, and try again.\nAFTER: Lead your people to victory on an ever-evolving world in this strategy game featuring asymmetric factions led by powerful heroes. Build your empire, wield political influence, and raise grand armies to wage war in a race to uncover dark secrets buried within the planet.\nTAGS BEFORE: 3D, Action, Action RPG, Action Roguelike, Co-op, Dungeon Crawler, Multiplayer, Mystery Dungeon, Procedural Generation, PvE, Replay Value, Roguelike, Roguelite, Sci-fi, Soundtrack, Strategy, Tactical RPG, Top-Down, Tower Defense, Twin Stick Shooter\nTAGS AFTER: 4X, Co-op, Diplomacy, Economy, Grand Strategy, Historical, Management, Medieval, Military, Moddable, Multiplayer, Sandbox, Simulation, Strategy, Tactical, Turn-Based, Turn-Based Strategy, War\n\n 3 2005010 Auroch Digital Brewmaster: Beer Brewing Simulator → Warhammer 40,000: Boltgun\nBEFORE: Brew your perfect beer in this relaxing home brewing sim. Use a realistic chemistry simulation to brew hoppy IPAs to creamy stouts. Customise your brewing space, create recipes and label your beer, unlock new equipment and enter competitions. Learn to brew, refine your craft and become a Brewmaster!\nAFTER: Load up your Boltgun and unleash the awesome Space Marine arsenal to blast your way through an explosion of sprites, pixels and blood in a perfect blend of Warhammer 40,000, frenetic gameplay and the stylish visuals of 90’s retro shooters.\nTAGS BEFORE: 3D, Casual, Cooking, Cozy, Crafting, Economy, Education, First-Person, Immersive Sim, Management, Physics, Realistic, Relaxing, Replay Value, Sandbox, Simulation, Wholesome\nTAGS AFTER: 1990's, Action, Adventure, Arcade, Arena Shooter, Boomer Shooter, Demons, FPS, Fantasy, Fast-Paced, First-Person, Gore, Old School, Pixel Graphics, Retro, Sci-fi, Shooter, Violent\n\n 4 3315330 iFAction Studio River Town Factory → Big Winner\nBEFORE: If possible, would you like to go back to ancient times to start a factory, refine oil, or even build a large computer? With the accidental fall of a meteorite, you traveled back to River Town a thousand years ago and became a workshop owner. And the story will begin from here……\nAFTER: In a future where the financial system has collapsed, the lottery becomes a lifeline for survival, with luck as the only currency. As an ordinary person in this predestined world, can you leverage the \"Luck System\" with courage and wit to rise from obscurity and attain the wealth others covet?\nTAGS BEFORE: 2D, Alternate History, Automation, Base Building, Building, City Builder, Crafting, Open World, Pixel Graphics, Resource Management, Sandbox, Sci-fi, Simulation, Story Rich, Survival, Time Management, Time Travel\nTAGS AFTER: 2D, Card Battler, Card Game, Cooking, Deckbuilding, Futuristic, Gambling, Multiple Endings, Pixel Graphics, RPG, Roguelite, Simulation, Strategy, Strategy RPG\n\n 5 2486820 Sonic Team SONIC X SHADOW GENERATIONS → Sonic Racing: CrossWorlds\nBEFORE: Shadow the Hedgehog is back with Classic and Modern Sonic in SONIC X SHADOW GENERATIONS, an all-new collection featuring two unique experiences!\nAFTER: Race across land, sea, air, space, and time in Sonic Racing: CrossWorlds! Warp through Travel Rings into new dimensions where something new awaits around every twist and turn.\nTAGS BEFORE: 2D, 2D Platformer, 3D, 3D Platformer, Action, Adventure, Cartoony, Colorful, Family Friendly, Fast-Paced, Linear, Platformer, Psychological Horror, Quick-Time Events, Replay Value, Romance, Score Attack, Soundtrack\nTAGS AFTER: 1990's, 3D, 4 Player Local, Action-Adventure, Arcade, Cartoon, Character Customization, Co-op, Co-op Campaign, Combat Racing, Family Friendly, Local Co-Op, Local Multiplayer, Multiplayer, Online Co-Op, PvP, Racing, Split Screen, Vehicular Combat\n\n 6 2121980 System Erasure ZeroRanger → Void Stranger\nBEFORE: This is the story of a fighter who wanted to become...\nAFTER: Place your faith and embrace the Void.\nTAGS BEFORE: 2D, Action, Arcade, Bullet Hell, Difficult, Fast-Paced, Female Protagonist, Local Co-Op, Pixel Graphics, Retro, Sci-fi, Shoot 'Em Up, Shooter, Story Rich\nTAGS AFTER: 2D, Adventure, Atmospheric, Difficult, Exploration, Female Protagonist, Horror, LGBTQ+, Mystery, Nudity, Pixel Graphics, Psychological Horror, Puzzle, Retro, Sokoban, Story Rich, Top-Down\n\n 7 2082410 RocketPunch Games HARDCORE MECHA → Anomaly Collapse\nBEFORE: HARDCORE MECHA is a 2D platformer action focusing on a better handling. The Campaign mode offers rich gameplay and seamless in-game cut-scenes for an immersive experience. Choose your favorite mecha, experience the fast-paced hardcore competitive battle of steel in Multiplayer!\nAFTER: Command a squad of adorably fierce warriors in Anomaly Collapse, a turn-based Roguelite strategy game that offers a fresh twist to an old formula! Face off against otherworldly beings and uncover the dark secrets of supernatural calamities. Immerse yourself in a story where every move counts!\nTAGS BEFORE: 2D, 4 Player Local, Action, Anime, Cinematic, Fighting, Local Co-Op, Local Multiplayer, Mechs, Military, Multiplayer, Online Co-Op, Platformer, PvP, Sci-fi, Shooter, Story Rich\nTAGS AFTER: 2D, Anime, Cute, Deckbuilding, Mechs, Party-Based RPG, Roguelike, Roguelike Deckbuilder, Roguelite, Sci-fi, Strategy, Turn-Based, Turn-Based Combat, Turn-Based Strategy, Turn-Based Tactics\n\n 8 1710100 Wanba Studio  Wanba Warriors → Bing in Wonderland\nBEFORE: Have you got what it takes to be a Wanba Warrior? Step into the absurd world of calligraphy combat and prepare to lay the smackdown upon all who dare to oppose you, dishing out damage with your chosen ink brush of brutality! highly recommend playing with controllers.\nAFTER: A Roguelite monster killing satisfaction galore without any exploration elements. The game is filled to the brim with exotic enemies and stages with a collection of deep and robust boons, weapons, and items plus an abundance of pointless features and useless details. (CONTROLLER SUPPORTED)\nTAGS BEFORE: 2D, 2D Fighter, Action, Combat, Comedy, Cute, Family Friendly, Fighting, Funny, Local Co-Op, Local Multiplayer, Memes, Multiplayer, Online Co-Op, Parody , Physics, PvP, Stylized, eSports\nTAGS AFTER: 2D Platformer, Action, Action Roguelike, Arena Shooter, Cartoony, Casual, Colorful, Combat, Cute, Exploration, Funny, Hack and Slash, Hand-drawn, Ninja, Perma Death, Procedural Generation, Roguelite\n\n 9 1498570 SNK CORPORATION METAL SLUG ATTACK RELOADED → THE KING OF FIGHTERS XV\nBEFORE: Hard-core action with easy controls! The METAL SLUG we all love is back as a tower defense game!\nAFTER: SHATTER ALL EXPECTATIONS!\r\nTranscend beyond your limits with KOF XV!\nTAGS BEFORE: 2D, 2D Platformer, Auto Battler, Deckbuilding, Multiplayer, Nostalgia, Pixel Graphics, Psychological Horror, PvP, RPG, Retro, Simulation, Strategy, Survival Horror, Tower Defense\nTAGS AFTER: 2.5D, 2D Fighter, 3D, 3D Fighter, Action, Anime, Arcade, Character Customization, Competitive, Female Protagonist, Fighting, Local Multiplayer, Multiplayer, Nudity, PvP, Sexual Content\n\n 10 1647920 Steel Mantis Valfaris → Valfaris: Mecha Therion\nBEFORE: Set in a far corner of space, Valfaris is a heavy metal infused 2D action-platformer and the next game from the team behind Unity Awards finalist, Slain: Back From Hell. Brutal combat. Deadly enemies. Stunning pixel art. Savage soundtrack. Get ready to rip the galaxy a new wormhole.\nAFTER: Therion returns as the hunt for Lord Vroll continues. Pilot an advanced machine of death and destruction in this explosive 2.5D side-scrolling shoot ‘em up, and sequel to the critically acclaimed Valfaris. Strap in and unleash mecha hell!\nTAGS BEFORE: 2D, Action, Dark, Difficult, Fantasy, Gore, Gothic, Horror, Metroidvania, Old School, Pixel Graphics, Platformer, Retro, Sci-fi, Shoot 'Em Up, Side Scroller, Violent\nTAGS AFTER: 2.5D, Action, Aliens, Arcade, Bullet Hell, Colorful, Dark, Difficult, Gore, Gothic, On-Rails Shooter, Retro, Sci-fi, Sequel, Shoot 'Em Up, Shooter, Side Scroller, Violent\n\n 11 2489670 岛叶游戏 Internet Generation → The witch and the bottle of concepts\nBEFORE: A story about a dying chat server. Moderators are abusing their power, the owner has neglected the server, and the lives of members are miserable. And you, as a newcomer, will pick up your keyboard, and overthrow the owner and moderators' tyranny!\nAFTER: A fantasy pixel-style Metroidvania roguelike game. The amnesiac witch picks up her reliable broom and embarks on a journey with Elf companions to defeat the Villain and regain their homeland, unveil the conspiracy. Through the magic mirror, break the 4th wall and find the truth!\nTAGS BEFORE: 2D, 2D Fighter, 2D Platformer, Action, Action-Adventure, Adventure, Anime, Beat 'em up, Colorful, Comedy, Cute, Dialogue Heavy, Fantasy, Female Protagonist, Funny, Hack and Slash, Pixel Graphics, Spectacle fighter, Story Rich, Top-Down\nTAGS AFTER: 2D, Action Roguelike, Adventure, Dialogue Heavy, Lore-Rich, Medieval, Multiple Endings, Mystery, Pixel Graphics, RPG, Roguelike, Story Rich, Surreal\n\n 12 2078450 Caged Element GRIP: Combat Racing → Warhammer 40,000: Speed Freeks\nBEFORE: Harness gravity-defying speed in classic combat racing taken to the EXTREME. Master wall-riding, ceiling-surfing tracks, and unleash devastating weapons across 29 blistering circuits. Race solo, split-screen, or online with up to 10 players. GRIP delivers white-knuckle racing like nothing else. \nAFTER: High-orktane combat racing in the Warhammer 40,000 universe! Drive buggies, tanks and helicopters, mastering advanced movement and explosive weaponry online or against bots. Unlock gear, customize your rides and use the NEW Creation Workshop to create, share, and play custom maps from the community!\nTAGS BEFORE: 4 Player Local, Action, Arcade, Automobile Sim, Combat, Combat Racing, Driving, Fast-Paced, Futuristic, Local Multiplayer, Multiplayer, Racing, Sci-fi, Shooter, Split Screen, VR, Vehicular Combat\nTAGS AFTER: 3D, Arcade, Character Customization, Class-Based, Combat, Combat Racing, Dark Fantasy, Destruction, Driving, Futuristic, Multiplayer, Online Co-Op, PvP, Racing, Sci-fi, Shooter, Team-Based, Third Person, Third-Person Shooter, Vehicular Combat\n\n 13 2719130 YEARS 川建国同志想要连任/Comrade Trump's Re-election → 东方裁判梦~Rookie Attorney Legendry\nBEFORE: This game is for entertainment only. It does not represent any group or individual, nor does it represent or agree with any political views.\r\nA game simulating the re-election of Comrade Trump, let's write a hymn of loyalty.\nAFTER: \"Touhou Saibanmu ~Rookie Attorney Legendry\" is a puzzle-solving detective game based on the Touhou series. It mainly tells the story of Marisa Kirisame who, by chance, becomes a lawyer. The story is divided into three chapters, each with multiple endings.\nTAGS BEFORE: 2D, Cartoony, Casual, Choices Matter, Colorful, Cute, Funny, Memes, Multiple Endings, Political Sim, RPG, Sexual Content, Strategy, Strategy RPG, Visual Novel\nTAGS AFTER: 2D, 2D Platformer, Anime, Colorful, Cute, Detective, Faith, Fantasy, Female Protagonist, Funny, Logic, Puzzle, RPG, Story Rich, Strategy, Visual Novel\n\n 14 1693980 Motive STAR WARS™: Squadrons → Dead Space\nBEFORE: Master the art of starfighter combat in the authentic piloting experience STAR WARS™: Squadrons. Feel the adrenaline of first-person multiplayer space dogfights alongside your squadron, and buckle up in a thrilling STAR WARS™ story.\nAFTER: The sci-fi survival-horror classic returns, completely rebuilt to offer an even more immersive experience — including visual, audio, and gameplay improvements — while staying faithful to the original game’s thrilling vision.\nTAGS BEFORE: Action, Adventure, Co-op, FPS, Female Protagonist, First-Person, Flight, Multiplayer, Online Co-Op, PvP, Sci-fi, Shooter, Simulation, Space, Space Sim, Spaceships, Strategy, VR\nTAGS AFTER: Action, Action-Adventure, Adventure, Aliens, Atmospheric, Combat, Dark, Gore, Horror, Psychological Horror, Remake, Sci-fi, Shooter, Space, Survival, Survival Horror, Third Person, Third-Person Shooter, Violent\n\n 15 2296550 Element Studios Element TD 2 - Tower Defense → Axon TD: Uprising - Tower Defense\nBEFORE: Element TD 2 sees you combine powerful elements to create unique towers. Prove your defenses against challenging foes, and compete or cooperate with others to determine who has mastered the elements.\nAFTER: Axon TD lets you expand the map boundaries themselves. A huge arsenal of towers and a wide variety of enemies clash in a battlefield that grows to your whims. Engage in single, cooperative, and roguelite gameplay.\nTAGS BEFORE: 3D, Building, Character Customization, Co-op, Competitive, Fantasy, Multiplayer, Online Co-Op, PvE, PvP, RTS, Real Time Tactics, Replay Value, Score Attack, Strategy, Tactical, Team-Based, Tower Defense\nTAGS AFTER: 3D, Building, Co-op, Multiplayer, Online Co-Op, PvE, RTS, Real Time Tactics, Robots, Roguelike, Roguelite, Sci-fi, Score Attack, Strategy, Stylized, Tactical, Tower Defense, War\n\n 16 292000 Torn Banner Studios Chivalry 2 → No More Room in Hell 2\nBEFORE: Chivalry 2 is a multiplayer first person slasher inspired by epic medieval movie battles. Players are thrust into the action of every iconic moment of the era - from clashing swords, to storms of flaming arrows, to sprawling castle sieges and more.\nAFTER: No More Room in Hell 2 is an intense 8 player zombie shooter.\r\nOvercome the endless undead. Avoid infection. Survive hell together.\r\n\nTAGS BEFORE: 3D, 3D Fighter, Action, Atmospheric, Combat, Comedy, FPS, First-Person, Gore, Hack and Slash, Historical, Massively Multiplayer, Medieval, Military, Multiplayer, PvP, Swordplay, Third Person, War, Wargame\nTAGS AFTER: Action, Adventure, Atmospheric, Character Customization, Co-op, Dark, FPS, First-Person, Horror, Looter Shooter, Lore-Rich, Multiplayer, Online Co-Op, Outbreak Sim, PvE, Shooter, Survival, Survival Horror, Zombies\n\n 17 1866120 Ludosity Slap City → Heartwood Heroes\nBEFORE: Slap City is a Seriously Fun platform fighter with a perfect mix of characters from the Ludosity universe. Play Locally or Online. Slap it hard in Ranked or fast and loose in Slap Ball game mode. We are the original inventors of Clutch Technology.\nAFTER: A fast-paced board game for up to 4 players featuring humorous quests, quick turn-based combat and multiple-choice events. Explore the board, equip your character, get a job, and fight to become Hero of the Week! Play locally or online with Steam Remote Play.\nTAGS BEFORE: 2.5D, 2D Fighter, 4 Player Local, Action, Co-op, Fighting, Funny, Local Multiplayer, Memes, Multiplayer, Platformer\nTAGS AFTER: 4 Player Local, Board Game, Cartoony, Casual, Character Customization, Choose Your Own Adventure, Co-op, Co-op Campaign, Colorful, Family Friendly, Funny, Local Co-Op, Local Multiplayer, Multiplayer, Online Co-Op, Party, Party Game, PvE, RPG, Tabletop\n\n 18 3251240 White Star Studio Thriving City: Song → Skewer of Heroes: Bubble Survivors\nBEFORE: In 960 AD, the new Song Dynasty needs you! As magistrate, build a prosperous city from scratch. Develop bustling streets filled with businesses and entertainment, while navigating hidden crises of political strife and foreign invasions. How will you lead through these challenges?\nAFTER: Bubble Survivor is a hack-and-slash game that combines Roguelike elements with team-based gameplay. You will summon like-minded heroes to conquer challenges and defeat the malevolent slimes, restoring light to the world!\nTAGS BEFORE: Agriculture, Base Building, Building, Casual, Choices Matter, City Builder, Colorful, Economy, Hand-drawn, Historical, Inventory Management, Management, Mining, Quick-Time Events, Real Time Tactics, Resource Management, Simulation, Strategy, Strategy RPG, Trading\nTAGS AFTER: 2D, Action, Action Roguelike, Adventure, Anime, Bullet Hell, Cartoony, Casual, Colorful, Combat, Deckbuilding, Fantasy, Hack and Slash, Magic, Relaxing, Roguelike, Shooter, Strategy, Survival\n\n 19 1953070 NatsumeAtari Eiyuden Chronicle: Rising → SD Shin Kamen Rider Rumble\nBEFORE: Eiyuden Chronicle: Rising is an action RPG set in the same world as Eiyuden Chronicle: Hundred Heroes. Features town-upgrading mechanics, fast-paced combat. A prequel to the world of the Eiyuden Chronicle: Hundred Heroes.\nAFTER: The world of Shin Kamen Rider can now be experienced as an epic, exhilarating action game combining elements of hack and slash with beat 'em ups.\nTAGS BEFORE: 2.5D, 2D, Action, Action RPG, Action-Adventure, Adventure, Anime, Arcade, Building, Casual, City Builder, Dungeon Crawler, Exploration, JRPG, Pixel Graphics, RPG, Side Scroller, Story Rich\nTAGS AFTER: 1990's, 2D, Action, Action RPG, Action RTS, Action Roguelike, Action-Adventure, Beat 'em up, Casual, Hack and Slash\n\n 20 3547010 Tributary Games King Of The Castle → Occlude\nBEFORE: Take your rightful place on the throne in this multiplayer-only party game for 4+ players. Rule the kingdom with an iron fist or silver tongue as you steer through bloodshed, intrigue & madcap disaster! Can you keep your crown when others plot your downfall? More importantly, can you keep your head?\nAFTER: Occlude is a ritual disguised as Solitaire; a cosmic-horror card game where the rules are obscured and reality is at stake. Play through cryptic challenges to reveal forbidden lore in this singleplayer narrative puzzle game for the terminally curious. Don’t worry; you can always start over.\nTAGS BEFORE: 2D, Casual, Character Customization, Choices Matter, Choose Your Own Adventure, Co-op, Colorful, Diplomacy, Fantasy, Interactive Fiction, Management, Medieval, Minimalist, Multiplayer, Party Game, Political Sim, Story Rich, Strategy, Stylized, Turn-Based Tactics\nTAGS AFTER: Atmospheric, Card Game, Hand-drawn, Horror, Investigation, Minimalist, Puzzle, Solitaire, Story Rich, Strategy, Stylized, Surreal\n\n 21 3494520 People Can Fly OUTRIDERS → Lost Rift\nBEFORE: Outriders’ brutal and bloody combat combines frenetic gunplay, violent powers and deep RPG systems to create a true genre hybrid.\nAFTER: Lost Rift is a first person survival adventure shooter that blends cooperative base building with multiplayer PvE or PvP extraction expeditions. Journey solo or in teams of up to 5 players across a vast archipelago, craft gear, build bases, defeat raiders, and become legendary.\nTAGS BEFORE: Action, Adventure, Character Customization, Co-op, Gore, Loot, Looter Shooter, Multiplayer, Online Co-Op, Open World, Post-apocalyptic, RPG, Sci-fi, Sexual Content, Shooter, Story Rich, Third Person, Third-Person Shooter, Violent\nTAGS AFTER: Action, Adventure, Base Building, Character Customization, Combat, Crafting, Exploration, Extraction Shooter, FPS, First-Person, Massively Multiplayer, Multiplayer, Nature, Online Co-Op, Open World Survival Craft, PvE, PvP, Shooter, Survival\n\n 22 2229940 Striking Distance Studios The Callisto Protocol™ → [REDACTED]\nBEFORE: Survive to escape the horrors of Callisto and uncover the dark secrets of Jupiter’s dead moon.\nAFTER: Get ready to massacre monsters, race your rivals, and combat corpses of your fellow guards as you race to escape Black Iron Prison in the critically acclaimed roguelike, [REDACTED].\nTAGS BEFORE: 3D, Action, Action-Adventure, Adventure, Cinematic, Dark, Futuristic, Horror, Linear, PvE, Realistic, Robots, Sci-fi, Shooter, Space, Story Rich, Survival Horror, Third Person, Third-Person Shooter\nTAGS AFTER: 3D, Action, Action Roguelike, Action-Adventure, Comedy, Comic Book, Dark Comedy, Dark Humor, Difficult, Dungeon Crawler, Hack and Slash, Isometric, Roguelike, Roguelite, Sci-fi, Shooter, Stylized, Survival, Top-Down Shooter\n\n 23 1769980 Failbetter Games Sunless Skies: Sovereign Edition → Mask of the Rose\nBEFORE: SAIL THE STARS. BETRAY YOUR QUEEN. MURDER A SUN. \r\n\r\nSunless Skies is a Gothic Horror roleplay game with a focus on exploration and exquisite storytelling.\nAFTER: A marvellous romance with a hint of murder, where every playthrough is different! Seek love, friendship, or more... intimate encounters. Help a murdered man find justice. And watch out for the bats.\nTAGS BEFORE: 2D, Adventure, Alternate History, Atmospheric, Choices Matter, Dark, Exploration, Lovecraftian, RPG, Roguelike, Sequel, Space, Steampunk, Story Rich, Survival, Top-Down, Trains\nTAGS AFTER: 2D, Alternate History, Atmospheric, Choices Matter, Dark Fantasy, Dating Sim, Detective, Dialogue Heavy, Hand-drawn, Investigation, LGBTQ+, Lore-Rich, Lovecraftian, Multiple Endings, Otome, RPG, Romance, Story Rich, Visual Novel\n"}

**Checking release-date metadata**

**Checking release-date metadata**

**Inspecting original-date fields**

Script completed
Wall time 1.3 seconds
Output:

{}
{"chunk_id":"b81897","wall_time_seconds":0.179024937,"exit_code":0,"original_token_count":24,"output":"judgment\nmajor_activity_change    16\nrelated_variation         6\nunclear                   2\n"}
{"chunk_id":"68b1b3","wall_time_seconds":0.808936763,"exit_code":0,"original_token_count":1134,"output":"                developer                         previous_name                                 name              judgment  prev_reviews  reviews  positive_pct\n                doinksoft                      Devolver Bootleg                            Gunbrella               unclear        1031.0      922            85\n            moi rai games                     Monster Sanctuary                         Aethermancer     related_variation       10398.0     2561            91\n        AMPLITUDE Studios ENDLESS Dungeon™ - Definitive Edition                    ENDLESS Legend™ 2 major_activity_change        2950.0     2385            82\n           Auroch Digital    Brewmaster: Beer Brewing Simulator            Warhammer 40,000: Boltgun major_activity_change        1045.0    15641            90\n          iFAction Studio                    River Town Factory                           Big Winner major_activity_change        1273.0      549            82\n               Sonic Team            SONIC X SHADOW GENERATIONS            Sonic Racing: CrossWorlds major_activity_change       12117.0    11397            93\n           System Erasure                            ZeroRanger                        Void Stranger major_activity_change        2030.0     3280            91\n        RocketPunch Games                        HARDCORE MECHA                     Anomaly Collapse major_activity_change        2329.0      567            85\n             Wanba Studio                        Wanba Warriors                   Bing in Wonderland major_activity_change        2337.0     2019            97\n          SNK CORPORATION            METAL SLUG ATTACK RELOADED              THE KING OF FIGHTERS XV major_activity_change        1630.0     8352            83\n             Steel Mantis                              Valfaris              Valfaris: Mecha Therion major_activity_change        1771.0      300            94\n                     岛叶游戏                   Internet Generation The witch and the bottle of concepts     related_variation        1379.0       57            57\n            Caged Element                   GRIP: Combat Racing       Warhammer 40,000: Speed Freeks     related_variation        3665.0      491            85\n                    YEARS 川建国同志想要连任/Comrade Trump's Re-election       东方裁判梦~Rookie Attorney Legendry               unclear        1174.0       96            63\n                   Motive                 STAR WARS™: Squadrons                           Dead Space major_activity_change       25071.0    62597            90\n          Element Studios          Element TD 2 - Tower Defense    Axon TD: Uprising - Tower Defense     related_variation        2847.0      705            85\n      Torn Banner Studios                            Chivalry 2               No More Room in Hell 2 major_activity_change       46462.0    16065            62\n                 Ludosity                             Slap City                     Heartwood Heroes major_activity_change        2409.0       32            96\n        White Star Studio                   Thriving City: Song   Skewer of Heroes: Bubble Survivors major_activity_change        3052.0       18            83\n             NatsumeAtari             Eiyuden Chronicle: Rising           SD Shin Kamen Rider Rumble     related_variation        2069.0       53            88\n          Tributary Games                    King Of The Castle                              Occlude major_activity_change        1401.0      341            87\n           People Can Fly                             OUTRIDERS                            Lost Rift     related_variation       46137.0     1458            67\nStriking Distance Studios                The Callisto Protocol™                           [REDACTED] major_activity_change       35095.0      328            81\n         Failbetter Games      Sunless Skies: Sovereign Edition                     Mask of the Rose major_activity_change        2656.0      260            76\n           restriction  older1000    n  ge100  ge1000  median_reviews\nat_least_third_release      False 2100    189      36             7.0\nat_least_third_release       True  103     52      22           107.0\nthirdplus_smallcatalog      False 1843    182      35             7.0\nthirdplus_smallcatalog       True   84     44      18           110.5\n  thirdplus_price5to20      False  943    128      20            10.0\n  thirdplus_price5to20       True   54     26       6            95.5\n"}

**Verifying release-date precedence**

**Verifying release-date precedence**

**Testing credit-link sensitivity**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"38053f","wall_time_seconds":1.001014173,"session_id":46155,"original_token_count":0,"output":""}

**Checking date-field controls**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"bc69b0","wall_time_seconds":1.0009576,"session_id":90704,"original_token_count":0,"output":""}
{"chunk_id":"65cbcb","wall_time_seconds":0.000002002,"exit_code":0,"original_token_count":168,"output":"Passed 161 checks\nALL CREDITS\n prior_band    n  ge100  ge1000  median_reviews\n     10-99 2743    476     105            18.0\n   100-999 1364    732     164           116.0\n 1000-9999  559    454     223           695.0\n    10000+  183    163     133          3782.0\n       <10 3009    121      20             2.0\nDATE HISTORY\n    period prior_band    n  ge100  ge1000  median_reviews\n2023_2025      10-99 1765    264      59            16.0\n2023_2025    100-999  775    423      82           120.0\n2023_2025  1000-9999  287    244     119           730.0\n2023_2025     10000+   78     71      56          3993.5\n2023_2025        <10 2024     60      10             2.0\n"}

[
  {
    "type": "text",
    "text": "Script completed\nWall time 0.5 seconds\nOutput:\n"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"c66c3c\",\"wall_time_seconds\":0.000001276,\"exit_code\":0,\"original_token_count\":7,\"output\":\"Saved two PNG/SVG figures.\\n\"}"
  },
  {
    "type": "image",
    "note": "[image: shown below]"
  },
  {
    "type": "image",
    "note": "[image: shown below]"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"1f7b54\",\"wall_time_seconds\":0.289032509,\"exit_code\":0,\"original_token_count\":182,\"output\":\"all_eligible 7427 {'peak_half': 82.9, 'peak80': 34.4, 'other100': 25.2, 'other1000': 7.6, 'first_is_peak': 54.6, 'first_is_cohort_peak': 39.9}\\npeak1000 1248 {'peak_half': 86.1, 'peak80': 45.3, 'other100': 85.3, 'other1000': 45.5, 'first_is_peak': 45.9, 'first_is_cohort_peak': 32.9}\\npeak10000 307 {'peak_half': 88.6, 'peak80': 49.5, 'other100': 93.2, 'other1000': 78.2, 'first_is_peak': 43.0, 'first_is_cohort_peak': 30.0}\\nstrong combined 583 belowprevious 485 ratio median 0.24339757316202712 beats agepercentile 186\\nrecorded history a327ex [{'name': 'SNKRX', 'prev_name': 'BYTEPATH', 'gap_days': 1178.3586458333334, 'reviews': 4197, 'prev_reviews': 306.0, 'similarity': 0.058823529411764705, 'similarity_band': 'distant'}]\\n\"}"
  }
]

**Checking genre similarity evidence**

**Checking genre similarity evidence**

**Validating report charts**

**Checking genre similarity evidence**

**Validating report charts**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"f85d31","wall_time_seconds":1.001005002,"session_id":13948,"original_token_count":0,"output":""}

**Awaiting current test results**

**Awaiting current test results**

**Checking cohort statistics**

Script completed
Wall time 1.0 seconds
Output:

{"chunk_id":"45980c","wall_time_seconds":0.793235381,"exit_code":0,"original_token_count":651,"output":"Terry Cavanagh [{'name': 'VVVVVV', 'reviews': 6648, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'Terry Cavanagh'}, {'name': 'Super Hexagon', 'reviews': 18357, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'Terry Cavanagh'}, {'name': 'Dicey Dungeons', 'reviews': 10418, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'Terry Cavanagh'}, {'name': \"Terry's Other Games\", 'reviews': 61, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'Terry Cavanagh'}]\nSuspicious Developments [{'name': 'Gunpoint', 'reviews': 10274, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'Suspicious Developments'}, {'name': 'Floating Point', 'reviews': 3854, 'is_free': 1, 'in_main_study': False, 'all_developer_credits': 'Suspicious Developments'}, {'name': 'Morphblade', 'reviews': 264, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'Suspicious Developments'}, {'name': 'Heat Signature', 'reviews': 6851, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'Suspicious Developments'}, {'name': 'Tactical Breach Wizards', 'reviews': 11457, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'Suspicious Developments'}]\nincrepare games [{'name': 'English Country Tune', 'reviews': 185, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'increpare games'}, {'name': \"Stephen's Sausage Roll\", 'reviews': 1583, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'increpare games'}, {'name': 'Hypnocult', 'reviews': 42, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'increpare games'}, {'name': 'Mask Quest', 'reviews': 43, 'is_free': 0, 'in_main_study': False, 'all_developer_credits': 'increpare games | undef'}, {'name': 'Oeuf', 'reviews': 89, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'increpare games'}]\nMossmouth [{'name': 'Spelunky', 'reviews': 14661, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'Mossmouth'}, {'name': 'Spelunky 2', 'reviews': 19691, 'is_free': 0, 'in_main_study': False, 'all_developer_credits': 'Mossmouth | BlitWorks'}, {'name': 'UFO 50', 'reviews': 7105, 'is_free': 0, 'in_main_study': True, 'all_developer_credits': 'Mossmouth'}]\nnear baseline flags [{'name': 'Re:Night', 'first_date': Timestamp('2026-08-05 03:00:08+0000', tz='UTC'), 'original_release_date': nan, 'original_steam_release_date': nan, 'is_early_access': 0}, {'name': 'THE KING OF FIGHTERS XV', 'first_date': Timestamp('2024-12-12 08:00:00+0000', tz='UTC'), 'original_release_date': nan, 'original_steam_release_date': nan, 'is_early_access': 0}]\n"}
{"chunk_id":"cab489","wall_time_seconds":0.000001627,"exit_code":0,"original_token_count":168,"output":"Passed 162 checks\nALL CREDITS\n prior_band    n  ge100  ge1000  median_reviews\n     10-99 2743    476     105            18.0\n   100-999 1364    732     164           116.0\n 1000-9999  559    454     223           695.0\n    10000+  183    163     133          3782.0\n       <10 3009    121      20             2.0\nDATE HISTORY\n    period prior_band    n  ge100  ge1000  median_reviews\n2023_2025      10-99 1765    264      59            16.0\n2023_2025    100-999  775    423      82           120.0\n2023_2025  1000-9999  287    244     119           730.0\n2023_2025     10000+   78     71      56          3993.5\n2023_2025        <10 2024     60      10             2.0\n"}

**Comparing Steam release dates**

**Comparing Steam release dates**

Script completed
Wall time 1.5 seconds
Output:

THE KING OF FIGHTERS XV is now available! 39 star-studded characters are ready to fight!｜PRESS RELEASE｜SNK Corporation (https://www.snk-corp.co.jp/us/press/2022/021701/)
citeturn0search0 [wordlim: 200] Published: 4.6 years ago; Crawled: today; SNK CORPORATION (Corporate HQ: Suita-city, Osaka, Japan, Company President & CEO：Kenji Matsubara) is proud to announce our latest fighting game, THE KING OF FIGHTERS XV, is now on sale today, Feb 17^{th}, for PlayStation®5, PlayStation®4, Xbox Series X|S, Windows 10, Steam, and the Epic Games Store. ... https://www.snk-corp.co.jp/us/games/kof-xv/ ... Release Date  | February 17^{th}, 2022
--------------------------------------------------------------------------------
THE KING OF FIGHTERS SERIES SITE | SNK (https://www.snk-corp.co.jp/us/games/kof-portal/series/xv/)
citeturn0search1 [wordlim: 200] Crawled: today; THE KING OF FIGHTERS XV ... A total of 39 players, including popular characters from the past, long-awaited revival characters, and characters who play key roles in the story, have joined the game. ... Steam Released February 17th, 2022 (Digital only)
--------------------------------------------------------------------------------
SNK’s newest fighting game
THE KING OF FIGHTERS XV 
launches Feb 17th, 2022. Pre-orders start today!｜PRESS RELEASE｜SNK Corporation (https://www.snk-corp.co.jp/us/press/2021/082601/)
citeturn0search2 [wordlim: 200] Published: 5.0 years ago; Crawled: 2 days ago; SNK CORPORATION (Corporate HQ: Suita-city, Osaka, Japan, Chief Executive Officer: Kenji Matsubara) is proud to announce THE KING OF FIGHTERS (KOF XV) will be launching for PlayStation®5, PlayStation®4, Xbox Series X|S, Windows 10, Steam, and the Epic Games Store on February 17^{th}, 2022. ... https://www.snk-corp.co.jp/us/games/kof-xv/ ... Release Date  | February 17^{th}, 2022 (tentative)
--------------------------------------------------------------------------------
プレスリリース | 株式会社SNK (https://www.snk-corp.co.jp/press/2022/021703/)
citeturn0search3 [wordlim: 200] Published: 4.6 years ago; Crawled: today; 株式会社ＳＮＫ (本社：大阪府吹田市、代表取締役社長 CEO：松原 健二)は、新作対戦格闘ゲーム『THE KING OF FIGHTERS XV（KOF XV）』を、PlayStation®5／PlayStation®4／Xbox Series X|S／Windows 10／Steam／Epic Gamesストアで本日（2022年2月17日）に発売開始したことをお知らせいたします。 ... https://www.snk-corp.co.jp/official/kof-xv/
--------------------------------------------------------------------------------
THE KING OF FIGHTERS SERIES SITE | SNK (https://www.snk-corp.co.jp/official/kof-portal/series/xv/)
citeturn0search4 [wordlim: 200] Published: 4.6 years ago; Crawled: today; THE KING OF FIGHTERS XV ... Steam 2022年2月17日発売
--------------------------------------------------------------------------------
THE KING OF FIGHTERS XV | SNK (https://www.snk-corp.co.jp/official/kof-xv/index.html)
citeturn0search5 [wordlim: 200] Published: 4.6 years ago; Crawled: last month; Image: THE KING OF FIGHTERS XV ... Xbox Series X|S / Windows 10 / Steam / Epic Gamesストア(デジタル版)❙ 発売日 : 2022年2月17日 木曜日 ... +Fighter Pass+DLC Character "マチュア"+"バイス"+ DLCコスチューム "餓狼MotWテリー"+DLCコスチューム "CLASSIC LEONA"
--------------------------------------------------------------------------------
对战格斗游戏『THE KING OF FIGHTERS XV』今日发售！ 总计39名可玩角色的梦之对决！｜新闻发布｜株式会社SNK (https://www.snk-corp.co.jp/cn/press/2022/021701/)
citeturn0search6 [wordlim: 200] Published: 4.6 years ago; Crawled: 6 days ago; 株式会社SNK (本社：大阪府吹田市、代表取缔役社长 CEO：松原 健二)宣布，新作对战格斗游戏『THE KING OF FIGHTERS XV（拳皇15）』今日（2022年2月17日）在PlayStation®5／PlayStation®4／Xbox Series X|S／Windows 10／Steam／Epic Games商店正式发售。 ... http://www.snk-corp.co.jp/cn/games/kof-xv/
--------------------------------------------------------------------------------
PRODUCTS | THE KING OF FIGHTERS XV (https://snk-corp.co.jp/kr/games/kof-xv/products/)
citeturn0search7 [wordlim: 200] Published: 4.6 years ago; Crawled: 6 days ago; ❙ 타이틀명 : 더 킹 오브 파이터즈 XV （THE KING OF FIGHTERS XV） ... / Windows 10 / Steam / Epic Games Store (디지털 버전)❙ 발매일 : 2022년 2월 17일 (목) 예정 ... ■게임 본편 + Team Pass 1 + Team Pass 2 + Fighter Pass + DLC Character "MATURE"+"VICE"+ DLC Costume "GAROU: MotW TERRY"+ DLC Costume "CLASSIC LEONA"
--------------------------------------------------------------------------------
新作対戦格闘ゲーム『THE KING OF FIGHTERS XV』2022年2月17日（木）発売。本日より予約開始！ | 株式会社SNKのプレスリリース (https://prtimes.jp/main/html/rd/p/000000523.000028651.html)
citeturn0search8 [wordlim: 200] Published: 5.0 years ago; Crawled: last week; 株式会社ＳＮＫ (本社：大阪府吹田市、CEO：松原 健二)は、新作対戦格闘ゲーム『THE KING OF FIGHTERS XV（KOF XV）』を、PlayStation®5／PlayStation®4／Xbox Series X|S／Windows 10／Steam／Epic Gamesストアで2022年2月17日（木）に発売することをお知らせいたします。 ... https://www.snk-corp.co.jp/official/kof-xv/
... [43 more lines]

**Applying verified date override**

**Applying verified date override**

**Adding source override helper**

**Applying verified date override**

**Adding source override helper**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"87b283","wall_time_seconds":1.001327045,"session_id":64521,"original_token_count":0,"output":""}

**Updating README scope**

Script completed
Wall time 0.7 seconds
Output:

{"chunk_id":"b37324","wall_time_seconds":0.672093298,"exit_code":0,"original_token_count":1806,"output":"ACCOUNTING {\n  \"source_games\": 184664,\n  \"valid_released\": 128878,\n  \"paid_nonexplicit\": 102394,\n  \"paid_multicredit_or_missing\": 6515,\n  \"paid_ambiguous_name\": 2814,\n  \"paid_promotional_excluded\": 52,\n  \"included\": 93013,\n  \"developers\": 59259,\n  \"ambiguous_normalized_names\": 519\n}\nSHAPES\n        group size    n  median_peak_share  peak_half  peak80  other100  other1000  median_second  first_is_peak  first_is_cohort_peak  median_games\nall_eligible 2-10 7427              0.689       6156    2552      1870        568           16.0           4057                  2960           2.0\nall_eligible  2-5 6775              0.714       5937    2525      1557        452           14.0           3855                  2849           2.0\nall_eligible 6-10  652              0.411        219      27       313        116           87.0            202                   111           7.0\n    peak1000 2-10 1248              0.769       1074     565      1065        568          847.0            573                   410           3.0\n    peak1000  2-5 1076              0.803        978     546       896        452          763.0            527                   388           3.0\n    peak1000 6-10  172              0.534         96      19       169        116         1629.5             46                    22           7.0\n   peak10000 2-10  307              0.796        272     152       286        240         4595.0            132                    92           3.0\n   peak10000  2-5  266              0.831        247     148       245        201         3855.0            122                    91           3.0\n   peak10000 6-10   41              0.539         25       4        41         39        12117.0             10                     1           7.0\nTRANSITIONS\n    period            axis     value    n  developers  under10  ge100  ge1000  liked100  median_reviews  rate100  rate1000  liked100_rate  median_cohort_pct  median_gap  median_price  expected100  ratio100  expected_liked100_dateprice  liked_ratio_dateprice  full_cell_support  median_prior_reviews\n2023_2025      prior_band     10-99 2345        2345      792    389      87       268            17.0    0.166     0.037          0.114              0.648     507.869          6.99      526.018     0.740                      361.224                  0.742              0.963                  26.0\n2023_2025      prior_band   100-999 1125        1125       97    609     136       426           116.0    0.541     0.121          0.379              0.856     751.000          9.99      352.768     1.726                      236.535                  1.801              0.934                 251.0\n2023_2025      prior_band 1000-9999  458         458       16    378     187       264           742.0    0.825     0.408          0.576              0.944    1129.030         14.99      180.002     2.100                      117.221                  2.252              0.860                2500.0\n2023_2025      prior_band    10000+  125         125        4    113      89        67          3148.0    0.904     0.712          0.536              0.981    1441.500         24.99       55.822     2.024                       34.077                  1.966              0.752               21983.0\n2023_2025      prior_band       <10 2668        2668     2069     99      15        55             2.0    0.037     0.006          0.021              0.318     384.343          4.99      473.389     0.209                      330.942                  0.166              0.975                   2.0\n2023_2025        gap_band      1-2y 1532        1532      676    369     108       242            14.0    0.241     0.070          0.158              0.608     518.832          6.99      347.607     1.062                      237.254                  1.020              0.751                  16.0\n2023_2025        gap_band  181-365d 1263        1263      622    203      51       142            10.0    0.161     0.040          0.112              0.557     265.791          4.99      211.921     0.958                      167.251                  0.849              0.826                  12.0\n2023_2025        gap_band      2-4y 1497        1497      554    436     144       296            20.0    0.291     0.096          0.198              0.664    1027.075          8.99      432.895     1.007                      273.603                  1.082              0.658                  26.0\n2023_2025        gap_band   30-180d 1278        1278      767    130      37        92             5.0    0.102     0.029          0.072              0.452      97.846          4.99      162.892     0.798                      156.170                  0.589              0.841                   8.0\n2023_2025        gap_band     4-10y 1151        1151      359    450     174       308            37.0    0.391     0.151          0.268              0.750    1995.960         11.99      432.685     1.040                      245.722                  1.253              0.560                  57.0\n2023_2025 similarity_band     close  617         617      249    157      52       110            17.0    0.254     0.084          0.178              0.651     360.191          7.99      173.963     0.902                      104.595                  1.052              0.663                  30.0\n2023_2025 similarity_band   distant 4657        4657     2190   1062     345       705            12.0    0.228     0.074          0.151              0.579     573.083          6.99     1007.635     1.054                      730.108                  0.966              0.749                  14.0\n2023_2025 similarity_band    middle 1447        1447      539    369     117       265            17.0    0.255     0.081          0.183              0.651     496.959          6.99      406.402     0.908                      245.297                  1.080              0.693                  30.0\nREFERENCE\n      period               group     n  developers  under10  ge100  ge1000  liked100  median_reviews  rate100  rate1000  liked100_rate  median_cohort_pct  median_gap  median_price\n  2019_2022 first_observed_paid 16057       16057     8286   2748     856      1726             9.0    0.171     0.053          0.107              0.431         NaN          4.99\n  2019_2022 returning_developer  4532        4532     1663   1265     488       824            20.0    0.279     0.108          0.182              0.601     491.500          6.99\n  2023_2025 first_observed_paid 22388       22388    13247   2842     722      1860             5.0    0.127     0.032          0.083              0.452         NaN          4.99\n  2023_2025 returning_developer  6721        6721     2978   1588     514      1080            13.0    0.236     0.076          0.161              0.604     531.955          6.99\n2026_JanAug first_observed_paid  9044        9044     6246    765     157       510             3.0    0.085     0.017          0.056              0.470         NaN          4.99\n2026_JanAug returning_developer  3060        3060     1605    538     140       354             8.0    0.176     0.046          0.116              0.624     392.195          6.99\n"}

Script completed
Wall time 1.2 seconds
Output:

{"chunk_id":"51694f","wall_time_seconds":1.00177318,"session_id":74597,"original_token_count":0,"output":""}
{}

Script completed
Wall time 1.2 seconds
Output:

{"chunk_id":"3eb687","wall_time_seconds":0.000001511,"exit_code":0,"original_token_count":1973,"output":"Warning: truncated output (original token count: 1973)\nTotal output lines: 55\n\nROBUSTNESS\n    period                restriction prior_band   n  ge100  ge1000  liked100  under10  median_reviews  below_predecessor  median_raw_ratio  beats_predecessor_cohort\n2023_2025                        all  1000-9999 458    378     187       264       16           742.0                374             0.257                       154\n2023_2025                        all     10000+ 125    113      89        67        4          3148.0                112             0.152                        31\n2023_2025               catalog2to10  1000-9999 438    363     176       259       15           717.5                360             0.252                       147\n2023_2025               catalog2to10     10000+ 116    104      82        64        4          3200.5                103             0.171                        30\n2023_2025 first2010plus_catalog2to10  1000-9999 431    356     171       254       15           701.0                356             0.247                       140\n2023_2025 first2010plus_catalog2to10     10000+ 111     99      78        64        4          3253.0                 99             0.171                        28\n2023_2025             no_recorded_ea  1000-9999 290    247     121       180        8           735.5                242             0.286                        97\n2023_2025             no_recorded_ea     10000+  78     71      56        44        3          3993.5                 68             0.197                        21\n2023_2025        no_shared_franchise  1000-9999 278    223     110       160       13           681.5                221             0.243                       102\n2023_2025        no_shared_franchise     10000+  73     65      46        37        3          1906.0                 66             0.090                        14\n2023_2025             unshared_pages  1000-9999 364    303     146       226       12           742.0                298             0.246                       124\n2023_202…973 tokens truncated…  3           754.0                 74             0.179                        25\n2023_2025 no_shared_franchise        gap_band  30-180d  16     14       8         9        0           882.0                 14             0.534                         3\n2023_2025 no_shared_franchise        gap_band    4-10y 155    131      69        91        8           843.0                124             0.211                        61\n2023_2025 no_shared_franchise similarity_band    close  26     24      19        15        1          3764.5                 19             0.359                        11\n2023_2025 no_shared_franchise similarity_band  distant 237    187      98       128       12           704.0                196             0.167                        75\n2023_2025 no_shared_franchise similarity_band   middle  88     77      39        54        3           884.0                 72             0.245                        30\nSIM20\n    period   predecessor   value    n  ge100  ge1000  liked100  under10  median_reviews  below_predecessor  median_raw_ratio  beats_predecessor_cohort\n2023_2025           all   close  790    232      59       158      284            24.0                532             0.558                       307\n2023_2025           all distant 4165    907     318       603     2021            10.0               2106             0.833                      2261\n2023_2025           all  middle 1766    449     137       319      673            18.0               1060             0.617                       816\n2023_2025 prior1000plus   close   97     90      50        58        1          1141.0                 86             0.262                        25\n2023_2025 prior1000plus distant  289    226     129       155       16           741.0                234             0.185                        95\n2023_2025 prior1000plus  middle  197    175      97       118        3           959.0                166             0.306                        65\n"}
{"chunk_id":"dcb4cd","wall_time_seconds":1.000970116,"session_id":13911,"original_token_count":0,"output":""}

**Updating the network specification**

Script completed
Wall time 1.7 seconds
Output:

{"chunk_id":"f5d668","wall_time_seconds":1.000857991,"session_id":85181,"original_token_count":0,"output":""}
{"chunk_id":"1b7d8d","wall_time_seconds":0.355645036,"exit_code":0,"original_token_count":156,"output":"   value   n  ge100  ge1000  liked100  median_reviews\n    1-2y 111     88      55        60           959.0\n181-365d  44     39      23        28          1021.5\n    2-4y 173    148      84        99           914.0\n 30-180d  24     17      10        10           574.0\n   4-10y 231    199     104       134           854.0\nstrong 583 below 486 median ratio 0.2417083854818523 percentile higher 185\npair_still_valid  judgment             \nFalse             major_activity_change     1\nTrue              major_activity_change    15\n                  related_variation         6\n                  unclear                   2\n"}

**Refining claim qualifiers**

Script completed
Wall time 0.0 seconds
Output:

{}
{"chunk_id":"b7ccb9","wall_time_seconds":0.000001598,"exit_code":0,"original_token_count":168,"output":"Passed 163 checks\nALL CREDITS\n prior_band    n  ge100  ge1000  median_reviews\n     10-99 2744    477     106            18.0\n   100-999 1364    732     164           116.0\n 1000-9999  558    453     222           695.0\n    10000+  183    163     133          3782.0\n       <10 3009    121      20             2.0\nDATE HISTORY\n    period prior_band    n  ge100  ge1000  median_reviews\n2023_2025      10-99 1765    264      59            16.0\n2023_2025    100-999  775    423      82           120.0\n2023_2025  1000-9999  287    244     119           730.0\n2023_2025     10000+   78     71      56          3993.5\n2023_2025        <10 2024     60      10             2.0\n"}

**Auditing final reported numbers**

[
  {
    "type": "text",
    "text": "Script completed\nWall time 0.0 seconds\nOutput:\n"
  },
  {
    "type": "text",
    "text": "{}"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"3738df\",\"wall_time_seconds\":0.000001636,\"exit_code\":0,\"original_token_count\":7,\"output\":\"Saved two PNG/SVG figures.\\n\"}"
  },
  {
    "type": "image",
    "note": "[image: shown below]"
  }
]

**Preparing comparison tables**

**Defining review sample**

**Scoping tool feasibility**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"63a431","wall_time_seconds":0.000002666,"exit_code":0,"original_token_count":839,"output":"\"\"\"One bounded diagnostic sample, not a review-history or whole-market crawl.\"\"\"\nfrom pathlib import Path\nimport json,sys,csv\nsys.path.insert(0,str(Path(__file__).resolve().parent.parent))\nfrom collect import BASE,request,utc,dump\nR=Path(__file__).resolve().parent\nTARGETS={3767850:'SpiritVale',2001760:'Beast of Reincarnation',2244210:'Echoes of Aincrad',4423580:'Shadow Dungeon',4364910:'Ruins of Dawn'}\npolicy_path=BASE/'network_policy.json';old=json.loads(policy_path.read_text())\npolicy=dict(old);policy['network_paused']=False;policy['reason']='Bounded build-expression research: 5 fixed games x 2 review sentiments x at most 30 English reviews, no media.'\ndump(policy_path,policy)\nrows=[];requests_used=0\ntry:\n for appid,name in TARGETS.items():\n  for sentiment in ['positive','negative']:\n   path=R/'review_samples'/f'{appid}_{sentiment}.json'\n   if path.exists():data=json.loads(path.read_text())\n   else:\n    params={'json':1,'filter':'recent','language':'english','purchase_type':'steam','review_type':sentiment,'num_per_page':30,'filter_offtopic_activity':1}\n    r=request(f'https://store.steampowered.com/appreviews/{appid}',params,interval=2)\n    d=r.json();assert d.get('success')==1\n    data={'fetched_at':utc(),'url':r.url,'selection':params,'query_summary':d.get('query_summary'),'reviews':[]}\n    for v in d.get('reviews',[]):\n     # No profile names, avatars or account IDs are needed for this diagnostic.\n     data['reviews'].append({k:v.get(k) for k in ['recommendationid','review','voted_up','timestamp_created','timestamp_updated','steam_purchase','received_for_free','written_during_early_access']}|{'playtime_at_review_minutes':v.get('author',{}).get('playtime_at_review')})\n    assert len(data['reviews'])<=30\n    dump(path,data);requests_used+=1\n   for v in data['reviews']:\n    if any(w in (v['review'] or '').lower() for w in ['howtomarketagame','marketing blog','steam market research']):continue\n    rows.append({'appid':appid,'game':name,'sample_sentiment':sentiment,**v})\nfinally:dump(policy_path,old)\nassert len(rows)<=300 and requests_used<=10\nwith (R/'review_sample.csv').open('w',newline='') as f:\n w=csv.DictWriter(f,fieldnames=list(rows[0]));w.writeheader();w.writerows(rows)\ndump(R/'review_sample_manifest.json',{'at':utc(),'requests':requests_used,'reviews':len(rows),'targets':TARGETS,'sample':'Latest available English Steam-purchase reviews, up to 30 of each sentiment per game. Deliberately balanced by sentiment; not a prevalence estimate or random population sample.'})\nprint('Sample complete',len(rows),'reviews;',requests_used,'requests. Bulk network collection disabled again.',flush=True)\nfrom pathlib import Path\nimport json\nimport pandas as pd\nR=Path(__file__).resolve().parent\ndef load_games():\n    g=pd.read_parquet(R.parent/'analysis-2026-09-05/games.parquet')\n    for appid,record in json.loads((R/'date_overrides.json').read_text()).items():\n        date=pd.Timestamp(record['date'],tz='UTC');mask=g.appid.eq(int(appid))\n        g.loc[mask,'first_date']=date\n        g.loc[mask,'year']=date.year;g.loc[mask,'month']=date.month\n        g.loc[mask,'quarter']=f'{date.year}-Q{date.quarter}'\n    return g\n{\n  \"network_paused\": true,\n  \"metadata_min_interval_seconds\": 2.0,\n  \"stop_on_http_429\": true,\n  \"reason\": \"Metadata snapshot finished. Network collection closed; no media downloader exists.\"\n}"}
[{"name":"list_mcp_resource_templates","description":"Lists resource templates provided by MCP servers. Parameterized resource templates allow s"},{"name":"list_mcp_resources","description":"Lists resources provided by MCP servers. Resources allow servers to share data that provid"},{"name":"mcp__codex_app__open_in_codex","description":"Tools provided by the Codex app.\n\nShow a workspace file, browser tab, terminal, or review "},{"name":"mcp__codex_app__wait_threads","description":"Tools provided by the Codex app.\n\nWait for the first of up to eight Codex threads to compl"},{"name":"mcp__codex_apps__codex_document_control_get_document_tool_schemas","description":"Use Codex Document Control to find connected document sessions, inspect the tools supporte"},{"name":"mcp__codex_apps__hotline_get_local_hotline","description":"Look up local suicide or self harm helpline information for the user based on country infe"},{"name":"mcp__codex_apps__plugin_management_get_app_permissions","description":"Manage plugins, settings, permissions, and connections. Prefer available built-in tools or"},{"name":"mcp__codex_apps__plugin_management_get_plugin_dependencies","description":"Manage plugins, settings, permissions, and connections. Prefer available built-in tools or"},{"name":"mcp__codex_apps__plugin_management_uninstall_app","description":"Manage plugins, settings, permissions, and connections. Prefer available built-in tools or"},{"name":"mcp__codex_apps__plugin_management_update_app_permissions","description":"Manage plugins, settings, permissions, and connections. Prefer available built-in tools or"},{"name":"mcp__codex_apps__sites_list_sites","description":"Use Sites to build, save, deploy, and inspect websites such as landing pages, portfolios, "},{"name":"request_plugin_install","description":"# Suggest a recommended plugin installation\n\nUse this tool only when all of the following "},{"name":"web__run","description":"Tools in the web namespace.\n\nTool for accessing the internet.\n\n\n---\n\n## Examples of differ"}]

**Review official Steam API docs**

Script completed
Wall time 1.7 seconds
Output:

{"chunk_id":"640257","wall_time_seconds":0.000003062,"exit_code":0,"original_token_count":1860,"output":"#!/usr/bin/env python3\n\"\"\"Resumable Steam-only catalog snapshot. Raw responses are the source of truth.\"\"\"\nimport argparse\nimport concurrent.futures as cf\nimport datetime as dt\nimport gzip\nimport hashlib\nimport html\nimport json\nimport math\nimport os\nfrom pathlib import Path\nimport re\nimport sqlite3\nimport threading\nimport time\nfrom urllib.parse import urlparse\nimport requests\n\nBASE = Path(__file__).resolve().parent\nROOT = BASE / '2026-09-05'\nQUERY = 'https://api.steampowered.com/IStoreQueryService/Query/v1/'\nCDN = 'https://shared.fastly.steamstatic.com/store_item_assets/'\nPRIMARY_COUNTRIES = ['US', 'BR', 'JP', 'CN', 'DE']\nEXTRA_COUNTRIES = ['GB', 'FR', 'CA', 'AU', 'KR', 'TW', 'HK', 'RU', 'UA',\n                   'TR', 'IN', 'SG', 'MX', 'AR', 'TH', 'ID', 'VN', 'ZA']\nCOUNTRIES = PRIMARY_COUNTRIES + EXTRA_COUNTRIES\nDATA_REQUEST = {k: True for k in (\n    'include_assets', 'include_release', 'include_platforms',\n    'include_all_purchase_options', 'include_screenshots', 'include_trailers',\n    'include_ratings', 'include_reviews', 'include_basic_info',\n    'include_supported_languages', 'include_full_description',\n    'include_assets_without_overrides', 'include_links')}\nDATA_REQUEST.update(include_tag_count=100, apply_user_filters=False)\nlocal = threading.local()\nrate_lock = threading.Lock()\nnext_request = 0.0\nadaptive_interval = 0.0\n\ndef utc():\n    return dt.datetime.now(dt.timezone.utc).isoformat()\n\ndef dump(path, obj):\n    path.parent.mkdir(parents=True, exist_ok=True)\n    tmp = path.with_suffix(path.suffix + f'.{os.getpid()}.{threading.get_ident()}.tmp')\n    tmp.write_text(json.dumps(obj, indent=2, ensure_ascii=False))\n    tmp.replace(path)\n\ndef request(url, params, interval=0.4):\n    global next_request, adaptive_interval\n    policy_path = BASE / 'network_policy.json'\n    policy = json.loads(policy_path.read_text()) if policy_path.exists() else {}\n    if policy.get('network_paused'):\n        raise RuntimeError('Network collection is paused by network_policy.json')\n    interval = max(interval, policy.get('metadata_min_interval_seconds', 0))\n    if not hasattr(local, 'session'):\n        local.session = requests.Session()\n        local.session.headers['User-Agent'] = 'SteamCatalogResearch/0.1 (public catalog snapshot)'\n    for attempt in range(7):\n        with rate_lock:\n            delay = max(0, next_request - time.monotonic())\n            next_request = max(time.monotonic(), next_request) + max(interval, adaptive_interval)\n        if delay:\n            time.sleep(delay)\n        try:\n            response = local.session.get(url, params=params, timeout=(15, 100))\n            if response.status_code in (401, 403):\n                raise PermissionError(f'{response.status_code} at {url}; not retried')\n            if response.status_code == 429 or response.status_code >= 500:\n                if response.status_code == 429 and policy.get('stop_on_http_429'):\n                    raise PermissionError(f'HTTP 429 at {url}; endpoint stopped by collection policy')\n                if response.status_code == 429:\n                    adaptive_interval = max(adaptive_interval, 1.8)\n                delay = max(float(response.headers.get('Retry-After', 0) or 0), min(120, 5 * 2 ** attempt))\n                with rate_lock:\n                    next_request = max(next_request, time.monotonic() + delay)\n                print(json.dumps({'event': 'backoff', 'status': response.status_code, 'seconds': delay}), flush=True)\n                time.sleep(delay)\n                continue\n            response.raise_for_status()\n            return response\n        except (requests.Timeout, requests.ConnectionError, requests.exceptions.ChunkedEncodingError):\n            if attempt == 6:\n                raise\n            time.sleep(min(60, 2 ** attempt))\n    raise RuntimeError(f'Retries exhausted: {url}')\n\ndef connection():\n    db = sqlite3.connect(ROOT / 'catalog.sqlite', timeout=60)\n    db.execute('PRAGMA journal_mode=WAL')\n    db.execute('PRAGMA synchronous=NORMAL')\n    db.executescript('''\n      CREATE TABLE IF NOT EXISTS pages (\n        phase TEXT, country TEXT, offset INTEGER, raw_path TEXT, fetched_at TEXT,\n        total INTEGER, returned INTEGER, unique_ids INTEGER,\n        PRIMARY KEY(phase,country,offset));\n      CREATE TABLE IF NOT EXISTS discovery (\n        appid INTEGER, country TEXT, source TEXT, PRIMARY KEY(appid,country,source));\n      CREATE TABLE IF NOT EXISTS games (\n        appid INTEGER PRIMARY KEY, name TEXT, type INTEGER, visible INTEGER,\n        is_free INTEGER, is_coming_soon INTEGER, is_early_access INTEGER,\n        steam_release_date INTEGER, original_release_date INTEGER,\n        original_steam_release_date INTEGER, release_from_early_access_date INTEGER,\n        reviews_filtered_count INTEGER, reviews_filtered_positive_percent INTEGER,\n        reviews_unfiltered_count INTEGER, reviews_english_count INTEGER,\n        short_description TEXT, source_country TEXT, fetched_at TEXT, data_json TEXT);\n      CREATE TABLE IF NOT EXISTS regional (\n        appid INTEGER, country TEXT, success INTEGER, visible INTEGER,\n        best_purchase_option_json TEXT, purchase_options_json TEXT,\n        reviews_json TEXT, fetched_at TEXT, PRIMARY KEY(appid,country));\n      CREATE TABLE IF NOT EXISTS app_aliases (\n        requested_appid INTEGER, canonical_appid INTEGER, country TEXT,\n        success INTEGER, visible INTEGER, fetched_at TEXT,\n        PRIMARY KEY(requested_appid,country));\n      CREATE TABLE IF NOT EXISTS game_tags (\n        appid INTEGER, tagid INTEGER, weight INTEGER, rank INTEGER, PRIMARY KEY(appid,tagid));\n      CREATE TABLE IF NOT EXISTS creators (\n        appid INTEGER, role TEXT, ordinal INTEGER, name TEXT, clan_id INTEGER,\n        PRIMARY KEY(appid,role,ordinal));\n      CREATE TABLE IF NOT EXISTS languages (\n        appid INTEGER, language_id INTEGER, additional_language_id INTEGER,\n        supported INTEGER, full_audio INTEGER, subtitles INTEGER,\n        PRIMARY KEY(appid,language_id,additional_language_id));\n      CREATE TABLE IF NOT EXISTS game_categories (\n        appid INTEGER, category_group TEXT, categoryid INTEGER,\n        PRIMARY KEY(appid,category_group,categoryid));\n      CREATE TABLE IF NOT EXISTS image_sources (\n        image_id TEXT PRIMARY KEY, url TEXT UNIQUE);\n      CREATE TABLE IF NOT EXISTS game_images (\n        appid INTEGER, image_id TEXT, kind TEXT, ordinal INTEGER,\ntotal 72\ndrwxr-xr-x 1 adn adn   496 Sep  6 09:00 .\ndrwxr-xr-x 1 adn adn   198 Sep  5 16:01 ..\ndrwxr-xr-x 1 adn adn    76 Sep  5 16:28 .venv\ndrwxr-xr-x 1 adn adn  1214 Sep  5 18:15 2026-09-05\n-rw-r--r-- 1 adn adn  8105 Sep  5 18:06 README.md\ndrwxr-xr-x 1 adn adn   252 Sep  5 17:48 __pycache__\ndrwxr-xr-x 1 adn adn  4476 Sep  5 19:03 analysis-2026-09-05\ndrwxr-xr-x 1 adn adn  5718 Sep  5 20:44 build-expression-2026-09-05\n-rw-r--r-- 1 adn adn 16575 Sep  5 17:44 collect.py\ndrwxr-xr-x 1 adn adn  1578 Sep  6 01:25 continuation-2026-09-06\n-rw-r--r-- 1 adn adn  2626 Sep  5 17:51 export.py\n-rw-r--r-- 1 adn adn  3312 Sep  5 17:48 finish_metadata.py\n-rw-r--r-- 1 adn adn  4318 Sep  5 17:58 inventory.py\n-rw-r--r-- 1 adn adn   194 Sep  5 20:26 network_policy.json\ndrwxr-xr-x 1 adn adn  1708 Sep  6 09:25 portfolios-2026-09-06\n-rw-r--r-- 1 adn adn  6391 Sep  5 17:46 reconcile.py\n-rw-r--r-- 1 adn adn  3561 Sep  5 17:32 repair_aliases.py\n-rw-r--r-- 1 adn adn  3536 Sep  5 16:16 search_check.py\n-rw-r--r-- 1 adn adn  2623 Sep  5 17:34 validate.py\n"}
User Reviews - Get List (Steamworks Documentation) (https://partner.steamgames.com/doc/store/getreviews?language=english)
citeturn1search0 [wordlim: 200] Crawled: 2 weeks ago; filter  | string  | ✔  | recent – sorted by creation time ... all – (default) sorted by helpfulness, with sliding windows based on day_range parameter, will always find results to return. ... language  | string  | ✔  | see https://partner.steamgames.com/doc/store/localization/languages (and use the API language code list) or pass “all” for all reviews ...     * `timestamp_created ` - date the review was created (unix timestamp) ...     * `steam_purchase ` - true if the user purchased the game on Steam

User Reviews - Get List

Steamworks Documentation > Store Presence > User Reviews - Get List

If you wish to get a dump of reviews on an application in Steam, you can use the following method with the parameters below.

## Parameters:

GET store.steampowered.com/appreviews/<appid>?json=1

Name  | Type  | Required  | Description
--- | --- | --- | ---
filter  | string  | ✔  | recent – sorted by creation time
updated – sorted by last updated time
all – (default) sorted by helpfulness, with sliding windows based on day_range parameter, will always find results to return.
If paging through the reviews with `cursor` then choose either the recent option or the updated option to eventually receive an empty response list.
language  | string  | ✔  | see https://partner.steamgames.com/doc/store/localization/languages (and use the API language code list) or pass “all” for all reviews
day_range  | string  | ✔  | range from now to n days ago to look for helpful reviews.       * `playtime_forever ` - lifetime playtime tracked in this app
      * `playtime_last_two_weeks ` - playtime tracked in the past two weeks for this app
      * `playtime_at_review` - playtime when the review was written
      * `deck_playtime_at_review` - playtime on Steam Deck when the review was written
      * `last_played ` - time for when the user last played
... [101 more lines]

**Implementing scoped histogram collection**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"937f71","wall_time_seconds":1.000478982,"session_id":81531,"original_token_count":472,"output":"byte_hist bytes 7794 keys ['success', 'results', 'count_all_reviews', 'expand_graph']\n{\"success\": 1, \"results\": {\"start_date\": 1519430400, \"end_date\": 1785628800, \"weeks\": [], \"rollups\": [{\"date\": 1517443200, \"recommendations_up\": 61, \"recommendations_down\": 0}, {\"date\": 1519862400, \"recommendations_up\": 17, \"recommendations_down\": 0}, {\"date\": 1522540800, \"recommendations_up\": 9, \"recommendations_down\": 2}, {\"date\": 1525132800, \"recommendations_up\": 1, \"recommendations_down\": 0}, {\"date\": 1527811200, \"recommendations_up\": 15, \"recommendations_down\": 0}, {\"date\": 1530403200, \"recommendations_up\": 9, \"recommendations_down\": 0}, {\"date\": 1533081600, \"recommendations_up\": 4, \"recommendations_down\": 0}, {\"date\": 1535760000, \"recommendations_up\": 1, \"recommendations_down\": 0}, {\"date\": 1538352000, \"recommendations_up\": 3, \"recommendations_down\": 0}, {\"date\": 1541030400, \"recommendations_up\": 6, \"recommendations_down\": 0}, {\"date\": 1543622400, \"recommendations_up\": 5, \"recommendations_down\": 0}, {\"date\": 1546300800, \"recommendations_up\": 7, \"recommendations_down\": 0}, {\"date\": 1548979200, \"recommendations_up\": 4, \"recommendations_down\": 0}, {\"date\": 1551398400, \"recommendations_up\": 2, \"recommendations_down\": 0}, {\"date\": 1554076800, \"recommendations_up\": 1, \"recommendations_down\": 1}, {\"date\": 1556668800, \"recommendations_up\": 2, \"recommendations_down\": 0}, {\"date\": 1559347200, \"recommendations_up\": 9, \"recommendations_down\": 0}, {\"date\": 1561939200, \"recommendations_up\": 14, \"recommendations_down\": 1}, {\"date\": 1564617600, \"recommendations_up\": 2, \"recommendations_down\": 0}, {\"date\": 1567296000, \"recommendations_up\": 1, \"recommendations_down\": 0}, {\"date\": 1569888000, \"recommendations_up\": 2, \"recommendations_down\": 0}, {\"date\": 1572566400, \"recommendations_up\": 8, \"recommendations_down\": 1}, {\"date\": 1575158400, \"recommendations_up\": 9, \"recommendations_dow\n"}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"020804","wall_time_seconds":0.000001643,"exit_code":0,"original_token_count":696,"output":"snkrx_hist bytes 6690 keys ['success', 'results', 'count_all_reviews', 'expand_graph']\n{\"success\": 1, \"results\": {\"start_date\": 1621209600, \"end_date\": 1788307200, \"weeks\": [], \"rollups\": [{\"date\": 1619827200, \"recommendations_up\": 115, \"recommendations_down\": 2}, {\"date\": 1622505600, \"recommendations_up\": 1343, \"recommendations_down\": 34}, {\"date\": 1625097600, \"recommendations_up\": 814, \"recommendations_down\": 16}, {\"date\": 1627776000, \"recommendations_up\": 222, \"recommendations_down\": 9}, {\"date\": 1630454400, \"recommendations_up\": 144, \"recommendations_down\": 5}, {\"date\": 1633046400, \"recommendations_up\": 108, \"recommendations_down\": 2}, {\"date\": 1635724800, \"recommendations_up\": 432, \"recommendations_down\": 3}, {\"date\": 1638316800, \"recommendations_up\": 70, \"recommendations_down\": 7}, {\"date\": 1640995200, \"recommendations_up\": 70, \"recommendations_down\": 14}, {\"date\": 1643673600, \"recommendations_up\": 35, \"recommendations_down\": 6}, {\"date\": 1646092800, \"recommendations_up\": 31, \"recommendations_down\": 5}, {\"date\": 1648771200, \"recommendations_up\": 29, \"recommendations_down\": 5}, {\"date\": 1651363200, \"recommendations_up\": 36, \"recommendations_down\": 5}, {\"date\": 1654041600, \"recommendations_up\": 44, \"recommendations_down\": 8}, {\"date\": 1656633600, \"recommendations_up\": 43, \"recommendations_down\": 3}, {\"date\": 1659312000, \"recommendations_up\": 33, \"recommendations_down\": 4}, {\"date\": 1661990400, \"recommendations_up\": 24, \"recommendations_down\": 9}, {\"date\": 1664582400, \"recommendations_up\": 28, \"recommendations_down\": 4}, {\"date\": 1667260800, \"recommendations_up\": 31, \"recommendations_down\": 10}, {\"date\": 1669852800, \"recommendations_up\": 36, \"recommendations_down\": 7}, {\"date\": 1672531200, \"recommendations_up\": 38, \"recommendations_down\": 2}, {\"date\": 1675209600, \"recommendations_up\": 12, \"recommendations_down\": 2}, {\"date\": 1677628800, \"recommendation\nbyte_summary bytes 1449 keys ['success', 'query_summary', 'reviews', 'cursor']\n{\"success\": 1, \"query_summary\": {\"num_reviews\": 1, \"review_score\": 8, \"review_score_desc\": \"Very Positive\", \"total_positive\": 282, \"total_negative\": 24, \"total_reviews\": 306}, \"reviews\": [{\"recommendationid\": \"231925114\", \"timestamp_created\": 1785704529, \"timestamp_updated\": 1785704529, \"voted_up\": true, \"steam_purchase\": true}], \"cursor\": \"AoJ4xpeS4p8Detfc9AY=\"}\nsnkrx_summary bytes 1193 keys ['success', 'query_summary', 'reviews', 'cursor']\n{\"success\": 1, \"query_summary\": {\"num_reviews\": 1, \"review_score\": 8, \"review_score_desc\": \"Very Positive\", \"total_positive\": 3969, \"total_negative\": 228, \"total_reviews\": 4197}, \"reviews\": [{\"recommendationid\": \"234308559\", \"timestamp_created\": 1788360722, \"timestamp_updated\": 1788360722, \"voted_up\": true, \"steam_purchase\": true}], \"cursor\": \"AoJw5eCmsaADf7zo/QY=\"}\n"}
{"chunk_id":"288914","wall_time_seconds":0.000003396,"exit_code":0,"original_token_count":139,"output":"byte_hist.json total 328 count_all_reviews False last [{'date': 1775001600, 'recommendations_up': 2, 'recommendations_down': 0}, {'date': 1780272000, 'recommendations_up': 2, 'recommendations_down': 0}, {'date': 1785542400, 'recommendations_up': 1, 'recommendations_down': 0}]\nsnkrx_hist.json total 4469 count_all_reviews False last [{'date': 1782864000, 'recommendations_up': 6, 'recommendations_down': 1}, {'date': 1785542400, 'recommendations_up': 4, 'recommendations_down': 0}, {'date': 1788220800, 'recommendations_up': 1, 'recommendations_down': 0}]\n"}

**Probing review histogram variants**

**Probing review histogram variants**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"045024","wall_time_seconds":1.000692625,"session_id":97331,"original_token_count":0,"output":""}

**Checking historical review dates**

**Checking historical review dates**

**Bounding historical review counts**

**Checking historical review dates**

**Bounding historical review counts**

**Checking historical review dates**

**Bounding historical review counts**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"18bafe","wall_time_seconds":0.000001155,"exit_code":0,"original_token_count":1173,"output":"byte_all_summary bytes 1449 keys ['success', 'query_summary', 'reviews', 'cursor']\n{\"success\": 1, \"query_summary\": {\"num_reviews\": 1, \"review_score\": 8, \"review_score_desc\": \"Very Positive\", \"total_positive\": 292, \"total_negative\": 25, \"total_reviews\": 317}, \"reviews\": [{\"recommendationid\": \"231925114\", \"timestamp_created\": 1785704529, \"timestamp_updated\": 1785704529, \"voted_up\": true, \"steam_purchase\": true}], \"cursor\": \"AoJ4xpeS4p8Detfc9AY=\"}\nsnkrx_all_summary bytes 1193 keys ['success', 'query_summary', 'reviews', 'cursor']\n{\"success\": 1, \"query_summary\": {\"num_reviews\": 1, \"review_score\": 8, \"review_score_desc\": \"Very Positive\", \"total_positive\": 4099, \"total_negative\": 238, \"total_reviews\": 4337}, \"reviews\": [{\"recommendationid\": \"234308559\", \"timestamp_created\": 1788360722, \"timestamp_updated\": 1788360722, \"voted_up\": true, \"steam_purchase\": true}], \"cursor\": \"AoJw5eCmsaADf7zo/QY=\"}\nbyte_hist_steam bytes 7794 keys ['success', 'results', 'count_all_reviews', 'expand_graph']\n{\"success\": 1, \"results\": {\"start_date\": 1519430400, \"end_date\": 1785628800, \"weeks\": [], \"rollups\": [{\"date\": 1517443200, \"recommendations_up\": 61, \"recommendations_down\": 0}, {\"date\": 1519862400, \"recommendations_up\": 17, \"recommendations_down\": 0}, {\"date\": 1522540800, \"recommendations_up\": 9, \"recommendations_down\": 2}, {\"date\": 1525132800, \"recommendations_up\": 1, \"recommendations_down\": 0}, {\"date\": 1527811200, \"recommendations_up\": 15, \"recommendations_down\": 0}, {\"date\": 1530403200, \"recommendations_up\": 9, \"recommendations_down\": 0}, {\"date\": 1533081600, \"recommendations_up\": 4, \"recommendations_down\": 0}, {\"date\": 1535760000, \"recommendations_up\": 1, \"recommendations_down\": 0}, {\"date\": 1538352000, \"recommendations_up\": 3, \"recommendations_down\": 0}, {\"date\": 1541030400, \"recommendations_up\": 6, \"recommendations_down\": 0}, {\"date\": 1543622400, \"recommendations_up\": 5, \"recommendations_down\": 0}, {\"date\": 1546300800, \"recommendations_up\": 7, \"recommendations_down\": 0}, {\"date\": 1548979200, \"recommendations_up\": 4, \"recommendations_down\": 0}, {\"date\": 1551398400, \"recommendations_up\": 2, \"recommendations_down\": 0}, {\"date\": 1554076800, \"recommendations_up\": 1, \"recommendations_down\": 1}, {\"date\": 1556668800, \"recommendations_up\": 2, \"recommendations_down\": 0}, {\"date\": 1559347200, \"recommendations_up\": 9, \"recommendations_down\": 0}, {\"date\": 1561939200, \"recommendations_up\": 14, \"recommendations_down\": 1}, {\"date\": 1564617600, \"recommendations_up\": 2, \"recommendations_down\": 0}, {\"date\": 1567296000, \"recommendations_up\": 1, \"recommendations_down\": 0}, {\"date\": 1569888000, \"recommendations_up\": 2, \"recommendations_down\": 0}, {\"date\": 1572566400, \"recommendations_up\": 8, \"recommendations_down\": 1}, {\"date\": 1575158400, \"recommendations_up\": 9, \"recommendations_dow\nsnkrx_hist_steam bytes 6690 keys ['success', 'results', 'count_all_reviews', 'expand_graph']\n{\"success\": 1, \"results\": {\"start_date\": 1621209600, \"end_date\": 1788307200, \"weeks\": [], \"rollups\": [{\"date\": 1619827200, \"recommendations_up\": 115, \"recommendations_down\": 2}, {\"date\": 1622505600, \"recommendations_up\": 1343, \"recommendations_down\": 34}, {\"date\": 1625097600, \"recommendations_up\": 814, \"recommendations_down\": 16}, {\"date\": 1627776000, \"recommendations_up\": 222, \"recommendations_down\": 9}, {\"date\": 1630454400, \"recommendations_up\": 144, \"recommendations_down\": 5}, {\"date\": 1633046400, \"recommendations_up\": 108, \"recommendations_down\": 2}, {\"date\": 1635724800, \"recommendations_up\": 432, \"recommendations_down\": 3}, {\"date\": 1638316800, \"recommendations_up\": 70, \"recommendations_down\": 7}, {\"date\": 1640995200, \"recommendations_up\": 70, \"recommendations_down\": 14}, {\"date\": 1643673600, \"recommendations_up\": 35, \"recommendations_down\": 6}, {\"date\": 1646092800, \"recommendations_up\": 31, \"recommendations_down\": 5}, {\"date\": 1648771200, \"recommendations_up\": 29, \"recommendations_down\": 5}, {\"date\": 1651363200, \"recommendations_up\": 36, \"recommendations_down\": 5}, {\"date\": 1654041600, \"recommendations_up\": 44, \"recommendations_down\": 8}, {\"date\": 1656633600, \"recommendations_up\": 43, \"recommendations_down\": 3}, {\"date\": 1659312000, \"recommendations_up\": 33, \"recommendations_down\": 4}, {\"date\": 1661990400, \"recommendations_up\": 24, \"recommendations_down\": 9}, {\"date\": 1664582400, \"recommendations_up\": 28, \"recommendations_down\": 4}, {\"date\": 1667260800, \"recommendations_up\": 31, \"recommendations_down\": 10}, {\"date\": 1669852800, \"recommendations_up\": 36, \"recommendations_down\": 7}, {\"date\": 1672531200, \"recommendations_up\": 38, \"recommendations_down\": 2}, {\"date\": 1675209600, \"recommendations_up\": 12, \"recommendations_down\": 2}, {\"date\": 1677628800, \"recommendation\n"}
{"chunk_id":"75c760","wall_time_seconds":0.000002818,"exit_code":0,"original_token_count":23,"output":"byte_hist.json 328\nbyte_hist_steam.json 328\nsnkrx_hist.json 4469\nsnkrx_hist_steam.json 4469\n"}

Script completed
Wall time 1.2 seconds
Output:

{"chunk_id":"0a04b7","wall_time_seconds":0.000002919,"exit_code":0,"original_token_count":7,"output":"byte 303 25\nsnkrx 4235 234\n"}
{}
{"chunk_id":"2e7cb4","wall_time_seconds":1.001439631,"session_id":20254,"original_token_count":0,"output":""}

**Scoping review census**

**Scoping review census**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"ca2f98","wall_time_seconds":0.000015905,"exit_code":0,"original_token_count":229,"output":"byte_unfiltered_summary bytes 1449 keys ['success', 'query_summary', 'reviews', 'cursor']\n{\"success\": 1, \"query_summary\": {\"num_reviews\": 1, \"review_score\": 8, \"review_score_desc\": \"Very Positive\", \"total_positive\": 292, \"total_negative\": 25, \"total_reviews\": 317}, \"reviews\": [{\"recommendationid\": \"231925114\", \"timestamp_created\": 1785704529, \"timestamp_updated\": 1785704529, \"voted_up\": true, \"steam_purchase\": true}], \"cursor\": \"AoJ4xpeS4p8Detfc9AY=\"}\nsnkrx_unfiltered_summary bytes 1193 keys ['success', 'query_summary', 'reviews', 'cursor']\n{\"success\": 1, \"query_summary\": {\"num_reviews\": 1, \"review_score\": 8, \"review_score_desc\": \"Very Positive\", \"total_positive\": 4099, \"total_negative\": 238, \"total_reviews\": 4337}, \"reviews\": [{\"recommendationid\": \"234308559\", \"timestamp_created\": 1788360722, \"timestamp_updated\": 1788360722, \"voted_up\": true, \"steam_purchase\": true}], \"cursor\": \"AoJw5eCmsaADf7zo/QY=\"}\n"}
{"chunk_id":"207124","wall_time_seconds":0.000003138,"exit_code":0,"original_token_count":1226,"output":"byte ['start_date', 'end_date', 'weeks', 'rollups', 'rollup_type', 'recent'] {'start_date': 1519430400, 'end_date': 1785628800, 'rollup_type': 'month', 'recent': [{'date': 1783123200, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1783209600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1783296000, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1783382400, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1783468800, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1783555200, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1783641600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1783728000, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1783814400, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1783900800, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1783987200, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1784073600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1784160000, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1784246400, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1784332800, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1784419200, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1784505600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1784592000, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1784678400, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1784764800, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1784851200, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1784937600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1785024000, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1785110400, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1785196800, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1785283200, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1785369600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1785456000, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1785542400, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1785628800, 'recommendations_up': 1, 'recommendations_down': 0}]} {'success': 1, 'count_all_reviews': False, 'expand_graph': False}\nsnkrx ['start_date', 'end_date', 'weeks', 'rollups', 'rollup_type', 'recent'] {'start_date': 1621209600, 'end_date': 1788307200, 'rollup_type': 'month', 'recent': [{'date': 1785801600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1785888000, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1785974400, 'recommendations_up': 1, 'recommendations_down': 0}, {'date': 1786060800, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1786147200, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1786233600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1786320000, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1786406400, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1786492800, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1786579200, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1786665600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1786752000, 'recommendations_up': 1, 'recommendations_down': 0}, {'date': 1786838400, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1786924800, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787011200, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787097600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787184000, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787270400, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787356800, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787443200, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787529600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787616000, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787702400, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787788800, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787875200, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1787961600, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1788048000, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1788134400, 'recommendations_up': 1, 'recommendations_down': 0}, {'date': 1788220800, 'recommendations_up': 0, 'recommendations_down': 0}, {'date': 1788307200, 'recommendations_up': 1, 'recommendations_down': 0}]} {'success': 1, 'count_all_reviews': False, 'expand_graph': False}\n"}

**Testing census requests**

**Testing census requests**

**Testing review census pagination**

**Testing census requests**

**Testing review census pagination**

**Testing census requests**

**Testing review census pagination**

**Computing candidate counts**

Script completed
Wall time 1.4 seconds
Output:

{"chunk_id":"99003a","wall_time_seconds":1.196546317,"exit_code":0,"original_token_count":3289,"output":"                             dev                        name                                       anchor  anchor_id  anchor_reviews                                                       quiet  quiet_id  quiet_reviews                                     last  last_id  last_reviews  total\n                      name:rejet                       Rejet                                  Ken ga Kimi    1162650            1059                                                        與魔共舞 2206980.0           39.0                                與魔共舞 我的頌歌  2593680             4 1102.0\n            name:sugarcane games             Sugarcane Games                               Myths of Rules    2539350            1009                                                 Snake Crush 2574030.0           95.0                             Picnic Peril  2773600            21 1125.0\n                       name:冷笑黑妖                        冷笑黑妖           寇莎梅特：困世迷情 Consummate:Missing World     630060            1000                                          Where Pandas 熊猫在哪里 2685220.0           87.0                      Where Ferrets 雪貂在哪里  2685260            43 1130.0\n        name:satur entertainment         Satur Entertainment               Microcosmum: survival of cells     386260            1118                                          Tale of Dark Lands 3475860.0           35.0                 Astro Flame: Starfighter  3990840            17 1170.0\n                  name:mad mimic                   Mad Mimic                                    Dandy Ace    1037130            1052                                            No Heroes Here 2 2424480.0           13.0                         Mark of the Deep  1933370           229 1294.0\n              name:tuatara games               Tuatara Games                                Let Them Come     505630            1082                                            Bare Butt Boxing 1186660.0           13.0                 Let Them Come: Onslaught  2050800           216 1311.0\n               name:sinned games                Sinned Games                    Mineirinho Director's Cut    1201630            1366                                   Hitobashira - Human Suika 2904810.0            7.0    RAGMONTON ENDLESS for save indie game  2740030             1 1374.0\n       name:crypton future media        Crypton Future Media                   Hatsune Miku Logic Paint S    2089350            1076 Hatsune Miku - The Planet Of Wonder And Fragments Of Wishes 2477920.0           90.0              Hatsune Miku Logic Paint S+  3446190           277 1443.0\n        name:shinachiku-castella         Shinachiku-castella                              Scars of Summer    1405470            1296                                     Kotoko's a Little Weird 2657710.0           68.0                   Scars of Summer: After  4041630            95 1459.0\n            name:zoo corporation             Zoo Corporation    Delicious! Pretty Girls Mahjong Solitaire     540610            1426             Smith & Dungeon: Crafting the Legendary Weapons 3402300.0           39.0                        Merge Match March  3764410             0 1465.0\n                name:taito corp.                 TAITO CORP.                               Groove Coaster     744060            1437                                    NINJA HAYATE HD-Remaster 3052550.0           14.0     Bubble Bobble Sugar Dungeons Boosted  2937770            15 1466.0\n                     name:gustav                      Gustav                               Cat in the Box    1257410            1426                                         Death of Marionette 2618380.0           52.0                             Stray Savior  3282040            56 1534.0\n                       name:橙光游戏                        橙光游戏 蓝宝石般的被害妄想少女/Damsel with persecutory delusion     970510            1084                                                        巅峰之路 3375950.0            0.0                                    TOUQI  3686930           461 1545.0\n                  name:rad codex                   Rad Codex                               Horizon's Gate    1224290            1212                                                      Azalea 1366100.0           40.0                                Kingsvein  2276830           303 1555.0\n           name:istom games kft.            isTom Games Kft.                      Nyan Cat: Lost In Space     415420            1637                                        Junkyard Rush Racing 3363510.0           20.0                     To The Top, Mammoth!  3446750            12 1669.0\n                 name:vorh games                  Vorh Games                                   Screen Cat    3402530            1692                                                 Nasubi Room 3806370.0           12.0                              Dirty Piggy  3913550            12 1716.0\n            name:parallel studio             Parallel Studio                              Under The Waves    1975440            1655                                                   Dark Days 2856570.0           56.0                                     EQQO  2940170            11 1722.0\n           name:terminal reality            Terminal Reality                     BloodRayne: Terminal Cut    1373510            1583                         Terminal Velocity™: Boosted Edition 1956430.0           52.0                             Terminal One  3197060            96 1731.0\n                    name:arrowiz                     Arrowiz        寄居隅怪奇事件簿 Hermitage Strange Case Files     836640            1841                               Hermitage: Strange Case Files 1225820.0           33.0                           Mato Anomalies  2024390           131 2005.0\n                 name:desert fox                  Desert Fox                              Bad Dream: Coma     538070            1966                                    Faded Stories: Full Moon 2540460.0           41.0                     Bad Dream: Afterlife  3202410            32 2039.0\n                  name:paul bird                   Paul Bird                           Wild Animal Racing     389510            2039                                   The Haunted House of Doom  781920.0           17.0                    Puzzling Wild IQ Test  2078070            54 2110.0\n                    name:joy way                     Joy Way                                       STRIDE    1292040            1975                                                   Dead Hook 2342360.0           22.0                            STRIDE: Fates  2597880           150 2147.0\n              name:studio fizbin               Studio Fizbin                                 Say No! More    1191900            1758                                                 Lost At Sea 1278750.0           43.0                             Reignbreaker  2129810           440 2241.0\n           name:games for humans            Games For Humans                                   WAVESHAPER     562260            2231                                         Stronghold Conquest 2464220.0           10.0                                Game Room  3460520             2 2243.0\n          name:hello there games           Hello There Games    Kung Fury: Street Rage - Ultimate Edition     373180            2251                                     Invector: Rhythm Galaxy 2009450.0           87.0               The Ultimate BezzerQuizzer  4010090             0 2338.0\n             name:projectorgames              ProjectorGames                       FortressCraft Evolved!     254200            2652                                                 Astro Heist 2510540.0            0.0  Forklift Certified : Stacked and Loaded  3628890             3 2655.0\n                   name:andy law                    Andy Law                    Crazy Flasher Series 2021    1540150            2600                                                 A ARMY BASE 2673730.0           51.0                                  X Run 3  2996320            19 2670.0\n               name:cradle games                Cradle Games                                    Hellpoint     628670            2581   Teenage Mutant Ninja Turtles Arcade: Wrath of the Mutants 2527580.0           67.0          Snoopy & The Great Mystery Club  3328180            60 2708.0\n           name:resolution games            Resolution Games                                        Demeo    1484280            1702                                                 Racket Club 2407910.0           58.0 Demeo x Dungeons & Dragons: Battlemarked  3124340          1223 2983.0\n             name:crater studios              Crater Studios                               Dwarven Realms    2015240            2840                  Truck Town: Kids and Toddlers Driving Game 3584120.0            8.0                                Doomspire  3621050           140 2988.0\nname:tasharen entertainment inc. Tasharen Entertainment Inc.                                     Windward     326410            2667                                                    Celestis 3215480.0           14.0                         Windward Horizon  2665460           399 3080.0\n            name:lan - games ltd             LAN - GAMES LTD  Rento Fortune: Online Dice Board Game (大富翁)     663390            3062                                                Farm Builder 1275060.0           12.0  Carnival Heroes - Super party minigames  3142290            39 3113.0\n                   name:seedwall                    SeedWall                            Hentai Vs Furries    1208260            3055                                             Idol VS Furries 2579020.0           65.0                      Trump is in trouble  3295480             6 3126.0\n              name:acquire corp.               ACQUIRE Corp.             AKIBA'S TRIP: Undead ＆ Undressed     333980            3072                                        Ancient Weapon Holly 2403260.0           42.0    C.A.R.D.S. RPG: The Misty Battlefield  2539880            76 3190.0\n      name:whiskeybarrel studios       Whiskeybarrel Studios        Swords and Sandals Classic Collection    1055430            2887                                               Gorgon Shield 2446920.0           83.0               Swords and Sandals 3 Redux  2116620           231 3201.0\n     name:megapixel studio s. a.      MegaPixel Studio S. A.                THE HOUSE OF THE DEAD: Remake    1694600            2697                                                    Hollow 2 1055240.0           24.0          THE HOUSE OF THE DEAD 2: Remake  3376690           603 3324.0\n            name:digital eclipse             Digital Eclipse   Street Fighter 30th Anniversary Collection     586200            2300                                  Golden Tee Arcade Classics 3698680.0           45.0         Mortal Kombat: Legacy Kollection  3454980          1204 3549.0\n              name:sobaka studio               Sobaka Studio                   Redeemer: Enhanced Edition     447290            2633                                                    REMEDIUM 1659090.0           68.0                                   KIBORG  2405060          1220 3921.0\n                 name:walnut llc                  Walnut LLC                           CASE: Animatronics     489360            3801                                                Whisper Trip 1551920.0           90.0          CASE RECORDS: Fear of Abduction  2826400            86 3977.0\n        name:talesshop co., ltd.         TALESSHOP Co., Ltd.                  Some Some Convenience Store    1217390            2892                               The Lament of a Lifeless Star 2749650.0           51.0                   Sani Yang's Laboratory  3464370          1050 3993.0\n            name:take it studio!             Take IT Studio!                        Electrician Simulator    1080020            4042                                    Electrician Simulator VR 3256020.0           35.0                 Before Exit: Gas Station  3268520            32 4109.0\n            name:mooneye studios             Mooneye Studios                                   LOST EMBER     563840            4195                                     LOST EMBER - VR Edition 1880890.0           80.0            LOST EMBER: Rekindled Edition  3086580            99 4374.0\n                    name:triband                     Triband                               WHAT THE GOLF?     785790            4319                                               WHAT THE BAT? 2012100.0           84.0                            WHAT THE CAR?  2727650           805 5208.0\n                    name:varagtp                     VaragtP                                     Plantera     421040            4512                                              Boom Box Blue!  705240.0           24.0                 Plantera 2: Golden Acorn  1091920          1185 5721.0\n"}

**Shortlisting four studios**

**Shortlisting four studios**

**Collecting earlier portfolios**

Script completed
Wall time 0.9 seconds
Output:

{"chunk_id":"acac17","wall_time_seconds":0.748737611,"exit_code":0,"original_token_count":497,"output":"a327ex 2 4503\n appid     name                first_date  reviews\n760330 BYTEPATH 2018-02-24 00:51:44+00:00      306\n915310    SNKRX 2021-05-17 09:28:11+00:00     4197\nRad Codex 5 1961\n  appid              name                first_date  reviews\n 415920 Voidspire Tactics 2015-11-02 18:23:16+00:00      271\n 643900    Alvora Tactics 2017-06-01 13:46:12+00:00      135\n1224290    Horizon's Gate 2020-03-09 16:59:01+00:00     1212\n1366100            Azalea 2020-11-09 13:51:56+00:00       40\n2276830         Kingsvein 2024-01-16 11:33:03+00:00      303\nTuatara Games 3 1311\n  appid                     name                first_date  reviews\n 505630            Let Them Come 2017-10-03 16:50:01+00:00     1082\n1186660         Bare Butt Boxing 2023-05-04 12:43:26+00:00       13\n2050800 Let Them Come: Onslaught 2025-11-11 15:22:29+00:00      216\nStudio Fizbin 6 4329\n  appid                                 name                first_date  reviews\n 251430                      The Inner World 2013-09-27 17:43:00+00:00      920\n 613470 The Inner World - The Last Wind Monk 2017-10-20 16:58:39+00:00      455\n1191900                         Say No! More 2021-04-09 13:00:55+00:00     1758\n1049710                    Minute of Islands 2021-06-13 21:01:17+00:00      713\n1278750                          Lost At Sea 2021-07-15 16:00:55+00:00       43\n2129810                         Reignbreaker 2025-03-18 16:58:36+00:00      440\nDesert Fox 7 2336\n  appid                     name                first_date  reviews\n 538070          Bad Dream: Coma 2017-03-09 20:02:29+00:00     1966\n1527880    Darkness Under My Bed 2021-02-19 17:35:31+00:00       85\n1692060       Bad Dream: Stories 2021-09-30 20:57:54+00:00      117\n1837820 Faded Stories: Greenberg 2022-06-18 14:32:45+00:00       54\n2313780     Bad Dream: Purgatory 2023-06-23 17:28:04+00:00       41\n2540460 Faded Stories: Full Moon 2023-11-14 17:43:04+00:00       41\n3202410     Bad Dream: Afterlife 2025-10-10 17:01:27+00:00       32\n"}

**Finalizing catalog scope**

**Finalizing catalog scope**

Script completed
Wall time 2.1 seconds
Output:

{"chunk_id":"864c3d","wall_time_seconds":1.9167167410000001,"exit_code":0,"original_token_count":2471,"output":"\nAPP 1366100 Azalea \n A short, cute puzzle action-adventure through the tense and dangerous Forbidden Forest. \n Azalea is a short puzzle adventure with a cute yet spooky atmosphere. Famine has taken the land, forcing you to enter the Forbidden Forest to save your family. Explore snowy fields and dark caverns. Find and use magic flowers to solve puzzles and defeat bosses. Uncover the true nature of the forest.\n\nAPP 1186660 Bare Butt Boxing \n Fight in completely unregulated alien boxing bouts and punch out players online in BARE BUTT BOXING, a chaotic multiplayer brawler with\r\nphysics-based characters. \n Take control of mischievous aliens mesmerized by the sport of boxing as they battle it out all over our beloved planet Earth, unaware of the laws and traditions of human society. Enjoy the chaos and whimsy of simulated characters and environments in this action arena brawler that will keep you hooked with surprising combat twists and gameplay variety. Board the spaceship, we're off to the next scrimmage! MAIN FEATURES SIMULATED CHAOS - Take control of funny, physics based alien fighters and use crazy power ups to turn the tide of battle. ARENA COMBAT - Fight to dominate in different levels, each with their own quirks and hazards. ONLINE AND LOCAL MULTIPLAYER - Battle it out against players around the globe or invite your friends to local multiplayer matches. CUSTOMISATION - Unlock new player skins, gloves and name tags to customize your characters and stunt on your opponents.\n\nAPP 1278750 Lost At Sea \n What are the most important moments in your life? Lost At Sea is a game about life and death, about coming to terms with your past and about family. Set on a beautiful island, this game will force you to face your fears and take stock of your life through imaginative puzzles and a relatable story. \n “Life is what happens to you while you’re busy making other plans” - John Lennon Lost At Sea tells the story of Anna, who, in the later stages of her life, has come to be alone. To live her new future, she will first have to take stock of her past. Construct her memories by finding objects and solving puzzles on a strange island, and face the question we all face eventually: “Did I do it right?” In Lost At Sea players experience poignant moments in Anna’s life, moments that we can all relate to, moments that make us all wonder how life can suddenly take a hold of us, instead of the other way around. Set on a mysterious island, where every biome represents a phase of life, you will explore your surroundings and solve riddles to reclaim memories. All the while your fears are chasing you, trying to trap you in your misery and prevent you from moving on. In the end, this is a game about the things you want to remember, the things you wish for and about life as a whole. A personal story: Experience moments from someone’s life first-hand and help the main character relive and re-assess them to move on Island of the mind: Explore the different phases of life, each of them with their own beautiful biome on the island Beautiful memories: Gorgeous illustrations capture the essence of the most important moments in Anna’s life, while elaborate sound design brings her deepest fears to life Face your fears: Did I make the right choices? Did I fail my loved ones? As Anna, face all fears head-on, or walk away from them until you’re ready Game to remember: Unlock memories with specifically designed minigames and puzzles that capture these memories in unique gameplay\n\nAPP 2540460 Faded Stories: Full Moon \n Unburied and forgotten undead gather in the city center, while the citizens of Greenberg go about their day-to-day lives. \n Once approximately every 100 years, the unburied and forgotten undead gather in the city of Greenberg to undertake their final march. The missing, accident victims, and the murdered—everyone who hasn't been properly laid to rest participates in this event. The ultimate destination is the city catacombs, whose gates are opened by the city authorities during an event called the Full Moon. Faded Stories: Full Moon is an adventure game set in the surreal and unpredictable city of Greenberg. The entire series consists of short stories whose main aim is to showcase the culture, customs, and laws that govern this extraordinary place. Take on the role of an ordinary citizen and experience unusual events that could only happen in a city like this. \\ Unique and creative puzzles Varied art and gameplay styles for different chapters Constantly fresh and surprising gameplay mechanics Secrets Surreal and atmospheric world\n\nAPP 716500 Eliza \n Eliza is a visual novel about an AI counseling program, the people who develop it, and the people who use it. Follow Evelyn Ishino-Aubrey as she reconnects with people from her past, gets to know the people of Seattle who use Eliza for counseling, and decides the course of her future. \n Eliza is a visual novel about an AI counseling program, the people who develop it, and the people who use it. Follow Evelyn Ishino-Aubrey as she reconnects with people from her past, gets to know the people of Seattle who use Eliza for counseling, and decides the course of her future. After abandoning her high-powered tech career and a mysterious three-year absence, Evelyn Ishino-Aubrey resurfaces working as a proxy for a virtual counseling app called Eliza. Her job consists solely of reading a script provided to her in real-time by an AI, leaving her no autonomy over what she says. Is Eliza a technological marvel that brings therapy to those who otherwise couldn’t access it? Or is it an ineffective replacement and another vector for technology companies to gain control over our lives and humanity? As she reconnects with people from her past and gets to know the ordinary people of Seattle who use Eliza for counseling, Evelyn finds herself caught in the middle between differing sides of the argument— and as she begins to question everything about her past life and future purpose in the world, it becomes clear she must confront the turmoil and darkness inside. A thought-provoking story told in the form of a visual novel from an award-winning indie team. Follow Evelyn’s journey through a present-day Seattle brought to life with beautiful hand-painted art. Get to know Eliza’s clients and the employees of Skandha, the technology giant behind it, with over a dozen fully-voiced characters and powerful performances from a highly regarded cast. An evocative, ambient original score by the composer for Infinifactory, SHENZHEN I/O, Opus Magnum, and EXAPUNKS. Is AI-driven counseling better than nothing? Is it worse than nothing? What responsibilities come with technical skills? Consider your own answers to these questions and more. Counseling sessions getting a little too real? Relax with Kabufuda Solitaire, a new take on solitaire using Japanese kabufuda cards.\n\nAPP 760330 BYTEPATH \n BYTEPATH is a replayable arcade shooter with a focus on build theorycrafting. Use a massive skill tree, many classes and ships to create your own builds and defeat an ever increasing amount of enemies. \n BYTEPATH is a replayable arcade shooter with a focus on varied play styles with RPG elements. The game has a huge passive skill tree, classes and different ships through which you'll be able to try out lots of different builds to achieve ever increasing high scores and eventually beat the game. Expect BYTEPATH to be a mix of Bit Blaster XL and Path of Exile , created with the intention of expanding Bit Blaster XL's relaxing and addictive gameplay with Path of Exile's build depth, build diversity and RPG elements. Gameplay Your ship can't stop moving, so you must turn it left and right or use your boost and brakes to prevent it from crashing on enemies. Your ship can't stop shooting, so you must continually kill enemies and collect ammo so you can keep using your powerful attacks. Items, resources and new attacks will be spawned randomly. Those will provide you with great boosts that may help you achieve higher scores in the current run. Features 900+ nodes passive skill tree, allowing for a huge amount of varied play styles and builds 40+ classes, each giving stat boosts and modifiers which further enhance your build 10+ ships, each providing unique stats and modifiers (both positive and negative) that dramatically change the way you play 15+ enemy types which are spawned with increasing frequency as the run gets harder 40+ Steam achievements Entrancing soundtrack by AIRGLOW\n\nAPP 915310 SNKRX \n SNKRX is an arcade shooter roguelite where you control a snake of heroes that automatically attack nearby enemies. Combine different heroes to unlock class bonuses and create unique builds, and steer your unstoppable party as they ravage through endless waves of enemies. \n SNKRX is an arcade shooter roguelite where you control a snake made of multiple heroes, with each one having their own attacks, passives and classes. Inspired by the auto-battler genre, each hero in the snake has a set of classes, and combining enough heroes of the same class together grants additional class bonuses. Heroes can also be upgraded to be more powerful when copies are bought from the shop. Gameplay Your snake can't stop moving, turn it left or right to steer Your heroes attack automatically when close to enemies Combine heroes of the same classes together to unlock unique class bonuses Clearing arenas grants gold, which can be used to hire heroes at the shop Buying enough copies of the same heroes levels them up, making them more powerful Features 40+ heroes, each with unique attacks and passives 12+ classes, each giving stat boosts and modifiers to your snake 40+ passive items, each giving strong global effects to your snake 25+ levels of increasing difficulty as the run progresses 15+ Steam achievements Soundtrack by Kubbi\n"}

**Verifying 24 review dates**

**Verifying 24 review dates**

**Filtering Steam review records**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"267d09","wall_time_seconds":1.001313247,"session_id":29984,"original_token_count":0,"output":""}

**Checking pagination completeness**

**Checking pagination completeness**

**Reviewing store descriptions**

**Checking pagination completeness**

**Reviewing store descriptions**

Script completed
Wall time 5.5 seconds
Output:

{"chunk_id":"32c643","wall_time_seconds":5.004869078,"session_id":29984,"original_token_count":127,"output":"Bad Dream: Coma complete 1966 summary 1967 requests 21\nDarkness Under My Bed complete 85 summary 85 requests 23\nBad Dream: Stories complete 116 summary 116 requests 26\nFaded Stories: Greenberg complete 54 summary 54 requests 28\nBad Dream: Purgatory complete 41 summary 41 requests 30\nFaded Stories: Full Moon complete 41 summary 41 requests 32\nBad Dream: Afterlife complete 31 summary 31 requests 34\nVoidspire Tactics complete 271 summary 271 requests 38\nAlvora Tactics complete 135 summary 135 requests 41\n"}
{"chunk_id":"e22e96","wall_time_seconds":0.358268612,"exit_code":0,"original_token_count":1216,"output":"\n 415920 Voidspire Tactics \n An indie turn-based Tactics RPG with a heavy emphasis on exploration, character progression, and highly tactical combat. Create your own deeply customizable party of Warriors, Sorcerers, and Shiftcloaks to take on the Spire! \n Voidspire Tactics is an indie turn-based Tactics RPG with a heavy emphasis on exploration, character progression, and highly tactical combat. Create your own deeply customizable party of Warriors, Sorcerers, and Shiftcloaks to take on the Spire! Open-ended exploration – traverse a nonlinear world full of secrets Refined tactical combat fit for beginners and veterans alike – easy to learn, hard to master Choose from 4 playable races, unlock up to 18 classes, learn & upgrade 100+ abilities, and equip your choice of 50+ passive abilities Destructible terrain and elemental interactions – flood the battlefield, then shock everyone with lightning! Abilities are good for more than just combat – freeze a river to cross it, or grow seeds into vines to scale mountains! Atypical fantasy setting – not one elf, orc or goblin in sight\n\n 643900 Alvora Tactics \n Alvora Tactics is an indie Tactics RPG that combines strategic combat with immersive exploration. Explore and conquer Great Serpent Alvora with your own customizable band of warriors! \n Alvora Tactics is an indie Tactics RPG that combines strategic combat with immersive exploration. Explore and conquer Great Serpent Alvora with your own customizable band of warriors! Deep, yet straightforward & easy to understand tactical combat Explore a mix of procedural and handcrafted areas full of varied & challenging encounters Destructible terrain and elemental interactions – set a jungle ablaze with fire magic, or summon water then electrify it! Combine 10 races, 23 classes, 150+ upgrade-able abilities, and 50+ passives to create your ideal party Unique fantasy setting; no elves, orcs, or dwarves Built using the same engine, mechanics, and setting as the highly-rated Voidspire Tactics\n\n 1224290 Horizon's Gate \n Voyage into the wide world of Eral as an explorer, trader, or privateer. Defeat your enemies on land or at sea in deep but straightforward tactical combat. Unravel the secrets behind Dominio's betrayal – or sail off in search of your own adventure. \n Voyage into the wide world of Eral as an explorer, trader, or privateer. Unravel the secrets behind Dominio's betrayal - or sail off in search of your own adventure. Engage in deep but straightforward tactical combat on land or at sea Explore the world in your ship, then disembark to explore dungeons with your crew Create your crew from 8 playable races, unlock 36 classes, learn & upgrade 150+ abilities, and equip your choice of 100+ passive abilities Build a devastating armada of Ice Reavers and Sea Serpents – or sail past trouble with a small but swift fleet Serve under one of three factions, amass wealth as a trader, or become a reviled pirate Abilities are good for more than just combat – freeze a river to cross it, or grow seeds into vines to scale mountains! Unique fantasy setting – not one elf, orc or goblin in sight Built using the same engine, mechanics, and setting as the highly-rated Voidspire Tactics and Alvora Tactics (however, knowledge of those games is NOT required for Horizon's Gate!)\n\n 1366100 Azalea \n A short, cute puzzle action-adventure through the tense and dangerous Forbidden Forest. \n Azalea is a short puzzle adventure with a cute yet spooky atmosphere. Famine has taken the land, forcing you to enter the Forbidden Forest to save your family. Explore snowy fields and dark caverns. Find and use magic flowers to solve puzzles and defeat bosses. Uncover the true nature of the forest.\n\n 2276830 Kingsvein \n Explore and reclaim the infested castle-city of Kingsvein! Ride into its depths on your trusty wyvern, and lead your customizable party of Knights and Mystics in deep tactical combat. \n All contact with the city of Kingsvein has been lost. As Wisp Slayer, your role is to investigate and exterminate whatever evil lurks there. Adventure through the castle and its outer reaches as you see fit - secret treasures and skulking beasts hide around every corner! Dynamic, deep, and challenging tactical combat - toss your enemies into spikes then electrify their blood, or cover the battlefield in flames Customize your team by mixing and matching 16 different classes, each with upgradeable abilities and equippable passives Use your abilities to explore - freeze rivers with your magic, blast through cave walls, and throw chains to topple statues Ride your loyal steed in and out of combat, and train it to be as fierce as any Knight Set in the realm of Graven, a weird subterranean world of living soil and half-stone Diecast From the maker of the highly-rated Horizon's Gate, Voidspire Tactics and Alvora Tactics\n"}

**Comparing review counts**

**Comparing review counts**

**Preparing official source checks**

Script completed
Wall time 6.9 seconds
Output:

Zachtronics | Eliza (https://www.zachtronics.com/eliza/)
citeturn2search0 [wordlim: 200] Crawled: last month; Eliza is a visual novel about an AI counseling program, the people who develop it, and the people who use it.

# About Eliza

Eliza is a visual novel about an AI counseling program, the people who develop it, and the people who use it.

PURCHASE ON STEAM

PURCHASE ON GOG

PURCHASE ON ITCH.IO

PURCHASE ON SWITCH

DISCUSS ON REDDIT

After abandoning her high-powered tech career and a mysterious three-year absence, Evelyn Ishino-Aubrey resurfaces working as a proxy for a virtual counseling app called Eliza. Her job consists solely of reading a script provided to her in real-time by an AI, leaving her no autonomy over what she says. Is Eliza a technological marvel that brings therapy to those who otherwise couldn’t access it? Or is it an ineffective replacement and another vector for technology companies to gain control over our lives and humanity?
--------------------------------------------------------------------------------
Zachtronics (https://www.zachtronics.com/)
citeturn2search1 [wordlim: 200] Crawled: 5 months ago; Möbius Front '83 Defend America from an alternate-universe America using state-of-the-art 1980s military hardware in this intricate and tactical turn-based strategy game. ... Eliza Eliza is a visual novel about an AI counseling program, the people who develop it, and the people who use it.

MOLEK-SYNTEZ Use the MOLEK-SYNTEZ to create small molecules with various pharmacological effects from the comfort of your small Romanian apartment.

Eliza Eliza is a visual novel about an AI counseling program, the people who develop it, and the people who use it.

... [367 more lines]

**Verifying creator evidence**

**Verifying creator evidence**

**Sampling reviews manually**

**Verifying creator evidence**

**Sampling reviews manually**

Script completed
Wall time 5.4 seconds
Output:

{"chunk_id":"079130","wall_time_seconds":0.269627794,"exit_code":0,"original_token_count":4675,"output":"\nAPP 1366100 ENGLISH REVIEWS 30\n\nID 200077729 positive True \n Short but well done little indie game that provides a bite sized experience for a reasonable price\n\nID 190266378 positive True \n This game is excellent, it's like what I wanted TUNIC to be.\n\nI really hope they make a bigger version in the future with a Metroidbrainia / Zelda style open-world and some more items / abilities.\n\nWorks perfectly on the Steam Deck.\n\nThe only minor issue is the near pixel-perfect alignment you need to hit things sometimes, it could do with being more lenient.\n\nID 188872054 positive True \n Very short, practically casual. This monster vs humans story, souls of previous wanderers and a bunch of other stuff hit with Undertale vibes a little bit. Overall - nice\n\nID 175168472 positive False \n I don't know how I accidentally picked this up.  It's very short and feels like a demo.  As far as those go, it's fine.  Linear, top down sokoban adventure that has very tiny areas.  Didn't run into any problems, it was competently made.  \n\nNot sure why \"games\" like this exist on Steam.  Buying and playing it is merely an act of charity to the dev, It's worthless to the consumer.  Belongs on itch io.\n\nID 174607220 positive True \n Short and sweet\n\nID 156105938 positive True \n \"Azalea\" is a charming and short pixelated top-down action-adventure about a kid searching for healing in a forbidden glade. \n\nThe style and atmosphere is great and whimsical and while it is quite short (around 45 min) and quite on the easy side, it is a well rounded experience and plays like an interactive fairy-tale.\n\nID 154835592 positive True \n The platform creatures are cute.\n\nID 154397045 positive True \n Soundtrack is greaet, gameplay is nice\n\nID 146305962 positive True \n Not a whole lot here, but it's a pleasant way to spend an hour. The puzzles are simple, but the atmosphere is nice. I enjoyed the music and the ending had a nice poetic quality to it. It's worth the buck fifty and hour of your life if you're looking for something chill.\n\nID 140030930 positive True \n Neat little game. It's got some nice art and music. I wish it did a bit more with it's mechanics (the second item gets like, 3 puzzles) but for it's length maybe that's for the best. It's pretty cheap so I think for the price I got what I wanted out of it. Main thing to complain about is wanting more and that's not a bad thing.\n\nID 134700551 positive True \n It's short. It's cute. The story is nice. The bosses are decently clever, although you've probably seen the formula before.\n\nSeem to be some issues getting the settings to save. Unclear what the point of the difficulty settings is if there are no associated achievements (not much of an achievement hunter nor interested in higher difficulties, but figured I'd point it out). The puzzles aren't particularly interesting for what I've seen of other Rad Codex games.\n\nThat's about it. It's less than 2 dollars. I really like Rad Codex in general.\n\n(Says I have ~5.7 hours in the game, but it actually took 45 minutes to complete. I just forget to close things sometimes.)\n\nID 127088489 positive True \n What a nice  little journey. Not a hard puzzle game by any means, but some of the fire seed puzzles were new to me. If you're looking for a small adventure game inspired by zelda, undertale, and studio ghibli, then this is worth your spare change.\n\nID 122461008 positive True \n Short and sweet adventure game with a pleasant aesthetic and fun, simple puzzles. Took me about 1:10 to beat it while poking around for secrets (didn't find any). Another 15 minutes to run halfway through again to go down a side path that got locked behind me the first time. I had fun with it and definitely think it's worth the $1.50.\n\nThe developer's RPGs Voidspire Tactics and Alvora Tactics are also quite good, and I plan to play the newest one as well.\n\nID 120530071 positive True \n Very short. Very sweet.\n\nID 108072716 positive True \n Good. Music especially.\n\nID 101254608 positive True \n This little game was pure bliss. Thank you for making it!! \n\nBuy this at its full price. It is worth it and then some!\n\nID 96149153 positive True \n excellent adventure for an hour! thanks devs!\nотличная игруля на часок перед сном!!\n\nID 84677717 positive True \n Really enjoyed this.  Nice retro graphics.  Lovely music.  Nice story.\n\nWell worth it.\n\nID 82532083 positive True \n [url=http://store.steampowered.com/curator/32482033/]Crim's Humble Opinion Curator[/url]\n\n[h1]Overview[/h1]\nAzalea is a really neat little indie game. You take the role of a small child who needs to venture into a dangerous forest for medicine for their dying parents. There are some puzzles you will have to overcome Sokoban style, but they aren't anything too complicated. There are several enemies on your journey including bosses. For smaller enemies you don't really have any defenses aside from running or leading them to a light. You get two abilities to get you through puzzles and there are some chase scenes. The checkpoints are very forgiving and there are five different difficulties you can set the game to. The colors are vibrant and dark appropriately where they need to be. There are several different biomes to experience in your journey and the music is fitting. The sound design never gets annoying either. You can play with a controller or mouse and keyboard. The game is a short and simple experience. \n\n[h1]The Minor Gripes[/h1]\nThis game is a pretty well-rounded project. I don't have too many gripes about it. The movement can look a bit awkward at times, but I think that's just the engine. If you aren't into chase scenes you may not like having to solve puzzles during a chase scene. With the forgiving checkpoints, this isn't too bad though. There are some areas that look like you can explore them but you can't. The only other snag I ran into is some of the limited explanation of the powers. An indicator above where you use it the first time you use it could be beneficial.\n\n[code]\n[h1](っ◔◡◔)っ ♥ pros ♥[/h1]\n+Short & Simple\n+Enjoyable Music Design\n+Variable difficulties\n\n[h1](っ>▬<)っ x cons x[/h1]\n*Nothing major to note\n[/code]\n\nFor the very low price, this is an excellent little game. I would recommend it as a short little one session game. It took me around 40 minutes. If you like adventure games with simple puzzles then check it out.\n\nhttps://www.youtube.com/watch?v=bcY6dCvXcTw\n\n𝕀𝕗 𝕐𝕠𝕦 𝔼𝕟𝕛𝕠𝕪 𝕋𝕙𝕚𝕤 𝕋𝕣𝕪\nhttps://store.steampowered.com/app/1371940/Milos_Quest/\n\nThis review has been shortened for the curator reader base. If you are the developer or you enjoy diving into the details check out the long-form review [url=https://crimson-albedo.wixsite.com/crimshumbleopinion/post/azalea-detailed-opinions]here[/url].\n\nID 81764221 positive True \n I'd recommend this game to people with an hour of free time and can enjoy a good soundtrack\n\nID 80570529 positive True \n While incredibly short, this game is entirely beautiful. A linear game, but the way it is arranged makes that an asset rather than a detriment to the game. Minimal gameplay and puzzle solving, but the story and world set up more than made up for that and I enjoyed my quiet journey through the world.\n\nThe conclusion was very satisfying, though I cannot help but wish that this was a much longer game with a much bigger story, that is how much I enjoyed it.\n\nMore than worth its low asking price of $1.49.\n\nID 79806254 positive True \n £1 for an hours gameplay what can you say, a short quick game, well worth the value, the control of wasd can be a tad awkward when using skills but changing to arrow key movement sorts that, the graphics are great and music 2, the puzzles are simple, only real issue you may have is the chases, where you have to get the sequence right\n\nID 79627600 positive True \n It's a fun little thing.\n\nID 79584501 positive True \n This is a great little puzzle and action game with [b]delightful artwork.[/b]  The challenge was just right for a quick romp, and the story elements played nicely toward a satisfying ending. Well done! to the programmer and artist!\n\nID 79583730 positive True \n Pet dog.\n\nDied.\n\n10/10\n\nID 79439960 positive True \n this game is adorable. the puzzles are good, the story is interesting, the art and music are really good. its a very short game, but more than worth the dollar price tag.\n\nID 79331499 positive True \n I have made a YT series based around games for £1 and this was my first pick! What a beautiful game!\nCheck out my review here - https://youtu.be/xpFlSZAh5QY\n\nWell done guys, i loved it!\n\nID 79087787 positive True \n Short and straightforward, nice atmosphere, the story is simple but kinda neat.\n\n\n\nID 79086603 positive True \n It's a nice simple short game. It's got its charm and it does its job.\n\nID 79074448 positive True \n Cute little game.  Very short, very easy.  Maybe a little too easy.\n\nAPP 2540460 ENGLISH REVIEWS 25\n\nID 231995200 positive True \n Again, pretty much just more Bad Dream. This one's a bit creepier than the previous one. But, I really like how it continued the theme of undeads from the previous game, which extended into Bad Dream: Afterlife, which came out two years later.\r\n\r\nBut the one thing I praise most about this game, is that in chapter 3 and chapter 4, we go into the eyes of different characters, and get to see our two protagonists through their eyes. This is huge to me, because this is the first time in any Bad Dream or Faded Stories game that we actually get a full-on look at the character we are controlling. It makes me wonder if our protagonist in the Bad Dream series is actually either of these characters. I personally like to think that we are Sarah in the Bad Dream series. (Except in Bad Dream: Memories, since we see a boy in the drawings. Though maybe Sarah was just boyish when she was younger and is also lesbian or bisexual? I dunno.)\r\n\r\nEither way, I also noticed in chapter 4 where the game switched to a 3D freeroam style, which is what the original release of Bad Dream: Purgatory was. Again, I prefer the 2D or 2.5D point-and-click style, but I like how it switched to that just from wearing glasses.\r\n\r\nEither way, I hope to see more Bad Dream or Faded Stories games in the future. Desert Fox, please make more. 💛🥹\n\nID 224814209 positive True \n The game is fantastic! Desert Fox really outdid themselves.\n\nID 221865165 positive False \n I love this developper game, really (even though it doesn't show on my profile cuz I didn't play most of them on steam)  but this one was a deception. The point and click chapters were great, no notes, I thoroughly enjoyed them, but the others? Heh. \"move slowly left and right\" and worse \"move slowly forward while nothing happens\" were really not my jam. I ended up putting a book on my joystick to have my character move forward on it's own\n\nID 221078136 positive False \n I already played Bad Dream: Fever and really liked it, but this Game wasn't for me.\nThe artstile is again really cool. I liked the Chapters Monday, Wednesday and Friday, even tho there were a little too easy. But the Chapter between them are pretty boring and I were just waiting for them to end. The Chapter between Monday and Wednesday is just walking for a pretty long time. Sometimes you need to change your direktion when you can't get further. Fun! The Chapter between Wednesday and Friday is in 3D, but the controls are super wierd.\nI don't know. It took me two hours to finish the game, and I interrupted it twice during that time because I was simply bored.\n\nID 208901106 positive True \n Unlike any other game I have played.\n\nID 207590006 positive True \n The game is fine, I much preferred Bad Dream Coma & Fever. I just didn't really enjoy the gameplay for \"Undead Girl\" and \"Starring Contest\" (mainly because it was confusing navigating through it). The story telling and the setting are amazing as always though! I haven't played all games by Desert Fox, so it might be gimmicks taken from previous games that simply do not fit my taste. I recommend the game, but not for the full price.\n\nID 201831786 positive False \n Bad and slow and terrible\n\nID 199760626 positive True \n Really enjoyed this game. Love how you can use past games story elements to find Easter eggs in the other games. Love the repeatability of levels and being able to experience a completely different set of events from the first time you play. Love the different levels too. Short, but all levels are pretty very fun to play. Again I didn't enjoy the 2.5D movement level and having to walk from place to place took instead of clicking to travel, didn't feel like it fit in with the genre and style of the game, but it was still possible to complete. Maybe I will one day grow to like it, but I have been so accustomed to the 2D style that its difficult to transition.\n\nID 198253125 positive False \n i really enjoy the game in the beginning, but later imo it is boring, bc its so short and i don't really like this, also the game is not worth the price, for something like that i think that the price must be lower. if you like spending a lot of money on short games then here you go.\n\nID 192207128 positive True \n At this point you probably know if you like the ganes from Desert Fox or not. There are some new things this time around that change things up, that's always welcome, but overall it's more of the same, and IMO that's what I expect. Alt endings where you need to use information from the first run, point and click gameplay with a fair bit of weird dream like logic, a weird and depressing mood and very vague storytelling.\n\nIf you like the previous games, you will like this, if you don't then this will probably not convince you otherwise. I like them, so as long as they keep making them, I will keep buying them.\n\nID 191729171 positive True \n i like it a lot\nthe story is intresting\nthe puzzels are logic and cool\n:333333\n\nID 186581562 positive True \n I enjoyed it quite a bit more than I thought I would, it surprised me. \r\n\r\nIt pretends it doesn't save but really it does.  ;)\n\nID 170501235 positive True \n Love these games from Desert Fox.  A real puzzler to follow the different paths to various achievements.\n\nID 167446520 positive True \n 👍\n\nID 167439691 positive True \n Honestly, seeing how these series expands with lore and from simple point and click adventures to full 3D exploration with same formula is a great play. This one goes with a collection of chapters all reminiscent to previous days as a follow up. Let it be eerie or let it be confusing but the story is engaging and fun to explore alongside with the achievements to collect meaning it also has secrets to unfold with fun twists and ends.\r\nLovely game! Can't wait to see more of these.\n\nID 159339018 positive True \n Amazing game. \nI really love this game and how it expands on the lore of the other games. I also love how many new character interactions you have. \nThe puzzles in this game are very entertaining I love point & click games from this dev all of the other games are amazing too.  I liked how different each \"level\" was but that some were still connected.  Seeing scenes from previous games awakened my nostalgia from playing them. It's a very fun experience.\n\nID 159220058 positive False \n I have enjoyed most of this developers point and click titles, so I do not think I am being too harsh here: \n\nThe total content is around 2-3 hrs.. absolute tops, probably far less if you aren't high/wasted. There are some interesting point and click segments that involve strange interactions with the environment, but there is maybe a grand total of one-two hours of these segments and at least what feels like 30+ minutes of holding down a key and waiting for something to happen. \n\nI honestly feel like not much effort was put into this game, since compared to previous titles, or even other point and click titles the interactions seem sparse and there isn't an iota of the detail that you will find in the previous gems from this developer. Not to mention that there are no real puzzles and the game is much more of a walking simulator than a point and click. \n\nAll in a all a pretty big disappointment.\n\n@Desert Fox: I just give honest opinions. You have to understand that I just bought a game, played it and gave it a review. I did not do so with malicious intent. If I had found the game good and worth the price I would have given a great review. As the person buying it is only my obligation to pay the money and be completely honest if I write a review. Take my review with a grain of salt given my expectations and preferences and use it as information to improve future developments if you care about players with similar preferences.\n\nID 159164388 positive False \n Nope. Not today.\n\nID 158285991 positive True \n Excellent!\n\nID 157587906 positive True \n My favorite type of point & click. Tricky enough that I have to use my brain but not so difficult that I pull all the hair out of my head in frustration. Deceptively simple art that is fun to look at but also just the right amount of creepy. A cool story line and world. Minimal dialogue (except the creepy guy who wants a date - hahahaha!). Nothing on here is overkill, and it's as fun to play as the other Desert Fox titles I've played so far (I'm burning through all of them). I clocked a little over 3 hours on this, but I took my time looking around and had a couple times I had to walk away from the computer for \"responsibilities.\" There would still be plenty for me to go back and do now that I've gone through all the chapters. Worth the full price. I'll keep buying these as long as he keeps creating them.\n\nID 156307692 positive True \n you know the deal a classic point n click adventure\n\nID 156004690 positive True \n It's such an amazing game. I love all of the puzzles and mysteries implemented\n\nID 150230198 positive True \n I recommend ALL of the Bad Dream/Faded stories games to any puzzle lover. I love them all, and I love the replayability of them to find different solutions. I'm having a time with this one, but i'm getting there! The Dev of these games is also so sweet and tries to fix any problems you might run into. All of these games are 10/10 for me and are my comfort games when my anxiety is high and I just need to relax and jump into a different world. Thank you Dev for all your hard work and dedication to bring us more games! <3\n\nID 150186618 positive True \n Another amazing puzzle game from Desert Fox. Keeping the history of Point and Click alive with a spooky twist. Got to love it.\n\nID 150175291 positive True \n Very fun point and click puzzle story. I've always liked this developer's games.\n"}
{"chunk_id":"8f43a1","wall_time_seconds":5.000156742,"session_id":29984,"original_token_count":95,"output":"Say No! More complete 1757 summary 1757 requests 98\nMinute of Islands complete 714 summary 714 requests 107\nLost At Sea complete 43 summary 43 requests 109\nReignbreaker complete 440 summary 440 requests 115\nLet Them Come complete 1082 summary 1082 requests 127\nBare Butt Boxing complete 13 summary 13 requests 129\nLet Them Come: Onslaught complete 216 summary 216 requests 133\n"}

**Labeling 25 reviews manually**

**Labeling 25 reviews manually**

**Paraphrasing review evidence**

**Labeling 25 reviews manually**

**Paraphrasing review evidence**

**Labeling 25 reviews manually**

**Paraphrasing review evidence**

**Validating release-timing metrics**

**Labeling 25 reviews manually**

**Paraphrasing review evidence**

**Validating release-timing metrics**

Script completed
Wall time 0.0 seconds
Output:

{}

Script completed
Wall time 0.0 seconds
Output:

{"chunk_id":"d3c732","wall_time_seconds":0.000001841,"exit_code":0,"original_token_count":47,"output":"Eliza complete 1007 summary 1007 requests 145\nBYTEPATH complete 306 summary 306 requests 150\nSNKRX complete 4197 summary 4197 requests 193\nCollection closed: 193 requests, 15445 reviews.\n"}
{"chunk_id":"ee9c93","wall_time_seconds":0.00000325,"exit_code":0,"original_token_count":77,"output":"-rw-r--r-- 1 adn adn 4.2M Sep  6 10:48 steam-research/career-history-2026-09-06/reviews.parquet\n  },\n  {\n    \"appid\": 915310,\n    \"game\": \"SNKRX\",\n    \"complete\": true,\n    \"pages\": 43,\n    \"collected\": 4197,\n    \"summary_total\": 4197,\n    \"old_snapshot_total\": 4197,\n    \"earliest_created\": 1621252070\n  }\n]"}

Script completed
Wall time 1.4 seconds
Output:

Warning: truncated output (original token count: 10108)
Total output lines: 2

{"chunk_id":"cee10e","wall_time_seconds":0.84549269,"exit_code":0,"original_token_count":1946,"output":"HISTORY\n                                 name    n  english  summary_total  earliest_lag_days  clock_usable  count_by_day30  count_by_day90  count_by_day365\n                     Bad Dream: Coma 1966      565           1967               0.08          True              74              91            182.0\n               Darkness Under My Bed   85       46             85               0.16          True              10              13             40.0\n                  Bad Dream: Stories  116       60            116               0.29          True              12              22             36.0\n            Faded Stories: Greenberg   54       31             54               0.11          True              12              17             29.0\n                Bad Dream: Purgatory   41       24             41               0.13          True              11              12             20.0\n            Faded Stories: Full Moon   41       25             41               0.93          True               6              10             19.0\n                Bad Dream: Afterlife   31       18             31               0.83          True              12              17              NaN\n                   Voidspire Tactics  271      244            271               0.10          True              13              21             36.0\n                      Alvora Tactics  135      123            135               3.19          True               8              13             24.0\n                      Horizon's Gate 1212     1048           1212               0.22          True              88             166            296.0\n                              Azalea   40       30             40               0.12          True              16              19             22.0\n                           Kingsvein  303      279            304               0.13          True             141             165            225.0\n                     The Inner World  920      468            920               0.14          True              10              26             77.0\nThe Inner World - The Last Wind Monk  455      186            455               1.67          True              12              34             57.0\n                        Say No! More 1757      995           1757               0.05          True             174             235            379.0\n                   Minute of Islands  714      368            714               0.11          True              96             158            243.0\n                         Lost At Sea   43       24             43               0.42          True               4               4              6.0\n                        Reignbreaker  440      251            440               0.09          True             201             244            417.0\n                       Let Them Come 1082      617           1082               0.07          True              58              88            246.0\n                    Bare Butt Boxing   13        9             13               0.03          True              11              12             12.0\n            Let Them Come: Onslaught  216      147            216               0.02          True             152             178              NaN\n                               Eliza 1007      771           1007               0.03          True             103             142            405.0\n                            BYTEPATH  306      259            306               0.04          True              71              85            134.0\n                               SNKRX 4197     3620           4197               0.10          True             590            2298           3322.0\nPAIRS\n     developer                                 name                        previous_name  previous_reviews_at_launch  previous_reviews_now  prior_catalog_reviews_at_launch  prior_catalog_reviews_now  count_by_day90\n   Desert Fox                Darkness Under My Bed                      Bad Dream: Coma                        1116                  1966                             1116                       1966              13\n   Desert Fox                   Bad Dream: Stories                Darkness Under My Bed                          26                    85                             1298                       2051              22\n   Desert Fox             Faded Stories: Greenberg                   Bad Dream: Stories                          29                   116                             1520                       2167              17\n   Desert Fox                 Bad Dream: Purgatory             Faded Stories: Greenberg                          29                    54                             1736                       2221              12\n   Desert Fox             Faded Stories: Full Moon                 Bad Dream: Purgatory                          14                    41                             1817                       2262              10\n   Desert Fox                 Bad Dream: Afterlife             Faded Stories: Full Moon                          30                    41                             2157                       2303              17\n    Rad Codex                       Alvora Tactics                    Voidspire Tactics                          58                   271                               58                        271              13\n    Rad Codex                       Horizon's Gate                       Alvora Tactics                          53                   135                              175                        406             166\n    Rad Codex                               Azalea                       Horizon's Gate                         231                  1212                              448                       1618              19\n    Rad Codex                            Kingsvein                               Azalea                          33                    40                             1179                       1658             165\nStudio Fizbin The Inner World - The Last Wind Monk                      The Inner World                         199                   920                              199                        920              34\nStudio Fizbin                         Say No! More The Inner World - The Last Wind Monk                         174                   455                              585                       1375             235\nStudio Fizbin                    Minute of Islands                         Say No! More                         217                  1757                              820                       3132             158\nStudio Fizbin                          Lost At Sea                    Minute of Islands                          98                   714                              952                       3846               4\nStudio Fizbin                         Reignbreaker                          Lost At Sea                          24                    43                             3014                       3889             244\nTuatara Games                     Bare Butt Boxing                        Let Them Come                         866                  1082                              866                       1082              12\nTuatara Games             Let Them Come: Onslaught                     Bare Butt Boxing                          13                    13                             1062                       1095             178\n       a327ex                                SNKRX                             BYTEPATH                         219                   306                              219                        306            2298\n"}
{"chunk_id":"b61338","wall_time_seconds":0.25163405,"exit_code":0,"original_token_count":8956,"output":"Warning: truncated output (original token count: 8956)\nTotal output lines: 303\n\n\nAPP 915310 EXPLICIT REFERENCE CANDIDATES 32\n\nID 221781570 positive True \n SNKRX, created by solo developer a327ex, is a deceptively small game built around a very sharp idea: take the movement logic of classic Snake, fuse it with roguelite progression and auto-battler synergies, and strip everything down to its essentials. The result is a game that looks almost primitive at first glance but quickly reveals itself to be tightly designed, highly replayable, and surprisingly strategic. It doesn’t rely on flashy presentation or narrative hooks—instead, it succeeds through clarity of concept and the elegance of its systems.\n\nAt its core, the game asks you to control a constantly moving chain of heroes arranged like a snake, weaving through compact arenas while your units attack automatically. This creates an immediate tension between movement and survival. You’re not aiming or attacking directly; instead, your focus is on positioning, spacing, and reading enemy patterns while your party does the fighting for you. That constraint becomes the foundation for everything else, turning what could have been a passive system into something that feels active and skill-driven. A slight miscalculation in movement can cause your formation to collapse, while precise steering can turn even a fragile build into something surprisingly effective.\n\nThe real depth comes from how you assemble your team. Each hero belongs to one or more classes, and stacking similar classes unlocks powerful bonuses that shape your overall strategy. You might lean into ranged units that fill the screen with projectiles, or build around area damage that punishes enemies for clustering, or even create setups that rely on summoned entities or status effects. The shop system, which appears between waves, constantly feeds you new options while forcing you to make trade-offs—do you spend your limited currency upgrading a key unit, or take a risk on a new addition that could complete a synergy? These decisions are frequent, meaningful, and often irreversible, which gives each run a strong sense of identity.\n\nWhat makes this system especially effective is how approachable it feels. You don’t need to understand every interaction to start having fun, and the game rarely overwhelms you with information. Instead, it gradually reveals its possibilities through play. Early runs might feel chaotic or even lucky, but over time, patterns emerge and strategies become clearer. There’s a satisfying progression from simply surviving to deliberately crafting builds that work exactly as intended. Even when things fall apart, it often feels like a consequence of your choices rather than something arbitrary.\n\nThe pacing plays a big role in maintaining that engagement. Runs are broken into short waves that keep the action moving, while the pauses between them give just enough breathing room to make decisions without pressure. This rhythm makes the game easy to pick up for a quick session, but it also supports longer stretches of play when you get pulled into chasing a particularly strong build. As difficulty increases, the game introduces tougher enemy compositions and scaling challenges, pushing you to refine your approach without drastically changing the core mechanics.\n\nVisually, SNKRX commits fully to minimalism. Units are represented by simple shapes and colors, environments are clean and uncluttered, and effects are functional rather than decorative. This might seem understated, but it serves a clear purpose: readability. In a game where positioning and awareness are everything, the lack of visual noise becomes an advantage. The audio design complements this approach, with a punchy electronic soundtrack that reinforces the game’s tempo without overpowering it. It’s a presentation style that understands exactly what it needs to do and avoids anything unnecessary.\n\nThat said, the game’s simplicity does come with some limitations. Over time, the core loop remains largely unchanged, and while different builds offer variety, they don’t fundamentally alter how the game is played. Some combinations can feel significantly stronger than others, which can reduce experimentation once you discover what works best. The randomness of the shop can also be frustrating, occasionally preventing you from completing a build you’ve been aiming for. There’s also a sense that the system, while clever, is operating near its limits—there’s only so much variation that can be extracted from its central mechanic.\n\nEven with those constraints, SNKRX remains a remarkably focused and effective experience. It’s a game that understands its scope and executes its idea with precision, offering just enough complexity to stay engaging without losing its clarity. The combination of movement-based control and automated combat creates a unique gameplay rhythm that stands apart from more traditional roguelites, and its emphasis on synergy-building gives each run a sense of purpose.\n\nIn the end, SNKRX is a strong example of how far a well-defined concept can go when it’s refined and delivered without excess. It may not be expansive, but it is cohesive, satisfying, and consistently fun to return to. For players who appreciate tight design and strategic experimentation in a compact package, it’s an easy game to recommend.\n\nRating: 9/10\n\nID 216242458 positive True \n Simple, fun, hard, and somewhat addicting. However, the lack of meta progression makes the whole play-die-repeat loop feel a bit pointless. Synergies are not as evidently helpful as in other games, and it is very challenging to create a build that works for all enemies—but that’s part of what makes it interesting. There are no meta unlocks, which means doing better is a matter of understanding the game mechanics, some skill, and tons of chance.\n\nID 215440154 positive False \n Will randomly drop below 1fps during runs, and by extension cause my entire PC to slowdown as I desperately try to minimize the game window and close the application, and even after the window finally closes, my PC is so slow that I have to restart the entire thing. I think that something is wrong with the current build of this game. None of my other games do or have done anything like this.\n\nID 185006542 positive True \n Always come back to it, great game to have a few rounds on when youre burnt out on other games.\n\nID 158054318 positive True \n It's a fun little game by a good developer. It's like snake with tower defense. Quite fun!\nThere's not really any meta-progression, which is fine, not every game needs to have that. And the balance is definitely interesting, in that there's a lot of relics that can change things a LOT, but some base units just suck period (Host I'm looking at you).\nPerformance can be bad when you have too many swarmers (they span little critters that attack your enemies), but you need a somewhat specific build to really grind the game to a halt. It seems the pathfinding/AI is not built to scale well with more and more units.\nA SNKRX 2 with meta progression, better performance for pathfinding, and a \"survival\" mode where instead of stages you gain xp in an open environment, etc, etc. That's just a cheap way to add more content tbh.\nIf the dev is reading this, I like Bytepath too, but it was a lot thrown at the player at once. This game eases you in a lot better. I'm excited to see what you make next!\n\nID 138574519 positive False \n Google \"a327ex website NFT\".\nClick on the first link.\n\nYou can choose for yourself if you would like to support this man. But do so in knowing that about him.\n\nID 130406587 positive True \n Holy cow!!! There's so much depth here. Totally awesome little arcade game-- gorgeous design, aesthetically pleasing as hell, and the gameplay is genuinely engaging. I've lost *hours* playing this game. (In a good way!) Not another mindless arcade grind. Don't get me wrong, you can zone out and have a blast piloting your snake around, wreaking havoc on the enemy bleps-- but I've actually started to think pretty hard about how I can optimize my build between runs, and all I have to say is-- wow. Simple concept, beautifully executed! If you pay attention to this game's little complexities, it'll really reward you. ♥♥♥♥ as heck. And the MUSIC???!!! Dang!!! It honestly chills me out. For the past few days, my routine has been-- get home from work, pour myself a nice glass of scotch, sit down with my laptop, and open SNKRX. If I could give you six stars, I would. Listen, dude-- I feel like you probably put a lot of time and effort and heart into writing this game, and I just want to say: hell yeah. Keep at it. If game design's your passion, you're doing the right thing.   This game is my *♥♥♥♥.* I don't care if other games are bigger, or have larger budgets-- I like *this* one, and I've played it every day since I bought it. ^-^\n\nID 128135772 positive True \n Now this was a slapper. I was late for dinner playing this one! The sound track was neat too. Party composition is also pretty cool too. you can also get artifacts or relcs. I I was not aware I had their other game but i guess Bytepath just moved up on my to play list.\n\nID 122591274 positive False \n creator spent a year making no progress on reworking the game while posting about how much he loves nfts before ceasing the promised update altogether, a disappointment and the game feels unfinished at best, don't recommend the game and don't recommend giving the creator another chance on any other games\n\nID 122252593 positive True \n A simple little game with a solid hook. Didn't really get it at first but after a couple of runs of experimenting & learning it is a lot of fun. Well worth the tiny asking price for what you get to play with :)\n\nNo issues at all on steamdeck\n\nGreat job a327ex!\n\nEdit, 24/09/2022: This game really is a lot of fun that you can easily keep coming back to. It has a lot more depth than I initially saw & the gameplay loop is super satisfying. I'm still trying to wrap my head around such a flawless execeution of a \"snake, shooter roguelite\".\n\nCame here after a couple of real good, fun runs wanting to sing high praises. But turns out I'd already left a review after having a blast with the first couple of hours. So an edit will have to suffice.\n\nWill be keeping an eye on this developer for sure. Super fun, unique little game!\n\nID 119802659 positive True \n A single stick (or mouse) autobattler in the vein of vampire survivors that's progression and mechanics are much more rich. It also came out prior to vampire survivor by like a year but the other game clearly aped some of the aspects of this game and more people should play this one. I hope a327ex does 2 things for this game: improve performance (I've lost twice now in three games when the framerate went to zero and the cpu usage spiked to +70%, my PC is high end and this should not happen) and I hope for better controller/steam deck support. Pick this game up, it's worth more than the $3.\n\nID 110339674 positive False \n I am a huge fan of the indie-arcade style genre that has become somewhat a huge success on Steam. This seemed to fill every check box. Indie, simple, arcade, rogue-like, perks, etc. etc. But it just does not hit that same like so many other games that follow this formula do. I continued through trying to die by just clicking any combination of \"heroes\" and yet it still seemed to take forever. Maybe I did not give it enough time....but I mean...its snake (on steroids) what's to get in any amount of time?\n\nID 109050086 positive False \n Game's good, fun, easily replayable, and dirt cheap. Dev's a dirtbag convinced most people against NFTs aren't actually against them for the reasons they say they are, and instead just don't wan…108 tokens truncated… RPG but you can imagine it as being one considering the amount of combinations you can do. I have played this more than any of my other games within the day. I get hyped thinking about it. Amazing it only has one creator.\n\nID 96036344 positive True \n It's a short, fun game that offers a ton of replay value. What I like about the game is that while the choices is random, the amount of choices still makes you feel like you have a decent control over your run. In other games, runs can end simply because of bad luck while in this game skill still plays a big factor to how 'successful' your run is.\n\nIf you have the time, please give this game a try. There are not that many indie developers nowadays much less those who continue to update their games with new content. The game is decently-priced, and it helps support the indie game industry too.\n\n\nID 95986334 positive True \n Amazing game, absolutely worth the price, plenty of fun to play and theory craft. \nI would only have one suggestion: it would be nice if you could choose between the previous game modes at start.\nI'm at ng+4 and it is is already quite difficult, it would be nice to be able to return to the calm levels of previous game modes. Sometimes I want a challenge, sometimes I just want to play and relax.\n\nID 95412717 positive True \n EDIT (19/7/2024): Finally beat the game at ng :D\nIt sounds stupid but this here, this little 3$ game is one of my favorite games. I don't even like roguelikes and am kinda sick of them, but this little gem is special. The soundtrack is phenomenal and sometimes I just keep playing the game because of the music. \nThanks a327ex :)\n\nyou can see that I don't have enough hours to write a full review. But it's enough to recommend this game. It's so cheap, and the irony is that it has so much more replayability than many of the well known AAA games in my opinion. This is the beauty of indie games that are done right :)\nIt is simple, but will be more complex the more you think about it. It's easy to pick up, but it's SO MUCH HARD to even win for the first time, at least for me xd\nDefinitely highly recommended for killing time and even serious gaming sessions, definitely worth every penny, and at the end, definitely DO NOT HIT THOSE F-ING BEANS\n\nID 95224877 positive True \n Another great and addictive game by a developer that obviously knows how to string together several mechanics to make a compelling gaming experience in a small, cheap package;\n\nBytepath - A top-down shooter with several customization options through a skill tree and many ship options;\nSNKRX - A top-down shooter with centipede like controls and auto-chess type units for each segment of the centipede.  There are plenty of builds and ways to play while you mix and match upgradeable units and items.  I easily blew through 4 hours in 1 sitting, well worth the price and then some  This is a developer to keep your eye out for more content!\n\nID 94996875 positive True \n This is a fantastic game. As the creator states, it was only made in three months, yet the game is solid, enjoyable, and endlessly re-playable. While a little rough around the edges, the creator has been consistently updating the game with helpful patches and more quality content. I'm very excited for all the new stuff they have planned! Overall, a round of applause to the creator a327ex for a great game, as well as a quick thanks to Northernlion for introducing it to me via his streams :)\n\nID 94706164 positive True \n Autochess+snake from the guy that made bytepath. If that meant anything to you, get it.\n\nID 94239790 positive True \n Yo real talk, I got a lil crush on whoever made this game.\nIt's the same dev as BYTEPATH, a little arcade-shooter that's filled with love.\nEverything about this game oooooozes charm and…956 tokens truncated….5 hours I have logged. The game is too simple and has almost no (functional) build variety and zero replay value outside of NG+5. I expected a lot more gametime out of a roguelike/lite\n\nI beat the game once with a warrior-focused build, a mage build and a ranger build. There is never any reason to go anything other than ranger/rogue build except getting bored of ranger.\n\nHaving said that, the game has a lot of potential. It needs some sort of permanent unlock system for the dopamine hits and a sense of progression and interesting new modes. The class system and the mixing/matching needs to be fleshed out a lot more than it is currently. It's not particularly satisfying as it is, at all.\n\nI'm not holding my breath for the game to get better as the developer clearly stated at the end of the game that he is now making other games and to be on the lookout for those (as opposed to making this one better?)\n\nID 93712391 positive True \n This game is basically Snake meets some kind of idle tower defense game like BTD. You move your snake around and your units kill enemies in waves. You can upgrade units and switch them out and stuff. Its has similar aspects to other games in the rogue-like genre. Each run is different and you can strategize differently each time. Considering this game is $3 and I would happily spend $10 on it, I think it's definitely worth playing.\n\nAll of that being said, there are some things I would like to see in this game in the future. One of them being support for 21:9 displays in full screen mode. That's just a nitpick though and this game is still amazing even if that is never added.\n\nID 93497676 positive True \n [h1]A New Twist on an Old Snake[/h1]\n\n[b]Full review on 2 Headed Hero:[/b]\nhttps://youtu.be/YOUNcO9CPns\n\nSNKRX is a weird chimera of an auto-battler, a rogue-lite, and the classic arcade game SNAKE. As in [b]Bytepath[/b], a327ex's prior game, SNKRX merges a simple two-button control scheme with a layer of skill-building strategy to make an entirely engaging and addictive coffee break game. \n\nNG+ unlocks another segment for your battlesnake and ups the difficulty, a great way to provide limitless replay value. A full run only takes minutes from start to finish, so it is clear that SNKRX is meant to be enjoyed over and over for short periods of time. Or if you're like me... for hours on end.\n\nUpdate: Many changes and additions have been made since release, reinforcing that this game is SO GOOD.\n\nBest 3 bucks I've spent in a long, long time.\n\nID 93241020 positive True \n This is a very fun an challenging game where you can find yourself sinking hours on end trying to unlock the achievements. I have never played snake but enjoy the leveling and rougue-lite aspects of the game.\n\nYou find yourself putting the game down after a few runs only to come back to it later to pick it back up and try to unlock some more achievements. The game offers the right amount of challenge and with the appropriate RNG rolls of passives and characters you can easily get an OP build to steamroll the rest of the game the challenge is either with the white machine gunners or the orange lobbers that typically pick off your party in a single shot. typical dying points are about round 6, 9, 16ish at least for me you can pretty much faceroll the first 1-2 rounds and afk it which is nice.\n\nThe dev did a great job on putting out an appropriately priced retail 2.99 game. Some may say that this is still too high priced but compared to mobile and other games that are abandonware alphas and betas that the devs dissappear from, this dev is continuing to update the game with additional features classes etc free of charge!\n\nID 92900640 positive True \n A new entry into the rogue-like revolution, SNKRX stands strong on its own among Loop Hero, Luck Be a Landlord, and Spelunky 2, and Returnal. \n\nLike Loop Hero and Luck Be a Landlord, SNKRX brings its own unique take on the \"one more run\" mind control technique and builds on what many games did before. Much more \"rogue-like\" than \"rogue-lite\" each run of SNKRX is it's own; there are no persistent upgrades or unlocks that carry over run to run. In that aspect, it is fairly unique among the more modern rogue-inspired games (notwithstanding games like Pixel Dungeon that are masterpieces unto themselves). \n\nSeemingly drawing inspiration from the auto-battler genre (ex, TFT and Dota's Auto Chess), every level provides you with a form of currency. You can save up currency to get interest on it (1 currency per 5 saved up, up to a maximum of 5) or spend it on \"units\". Units in this game are spots on your snake, in the old Nokia phone game style. You add units onto your snake that, like auto-battlers, have synergies (ex, 3 rogues give you higher crit damage, 3 warriors give your warriors increased defense). In non-new game plus, you can have a maximum of 7 units. There is also an upgrade mechanic, to a maximum level of 3 per unit.\n\nAddictive, tightly designed, incredibly \"one-more-run\" driving, and an absolute banging soundtrack by Kubbi brings a complete package of a game. Highly recommend following a327ex to see what they have up their sleeve next. I can't imagine it won't be long until, if they desire, they'll be working for a larger studio with a bigger budget. Excited to see what comes next.\n\nFantastic game.\n\np.s. only found out about this game by watching Dan Gheesling's twitch channel #596\n\nID 92150794 positive True \n By the dev who did Bytepath. Pretty interesting and unique game.\n\nPros:\n-Good music.\n-Lots of cool abilities and combos\n-Art is simple and easy on eyes.\n-Many enemy types only based on colors.\n-Runs are different every time based on class choices.\n\nCons:\n-Beat half of NG+1 by not even touching keyboard.\n-UI layout is extremely confusing but finally figured it out.\n-Way too easy to get OP.\n-Needs more modes, like a Survival/Endless mode.\n\nAPP 760330 EXPLICIT REFERENCE CANDIDATES 9\n\nID 221865358 positive True \n BYTEPATH, developed and published by a327ex, is a minimalist arcade shooter that builds surprising depth out of extremely simple mechanics. At first glance, it looks almost barebones—a small ship, a scrolling field of enemies, and a constant stream of projectiles—but beneath that stripped-down surface lies a system-driven experience that thrives on experimentation and incremental progression. Rather than overwhelming players with complex controls or flashy systems, it focuses on a tight core loop and expands outward through customization and build diversity.\n\nThe gameplay is intentionally straightforward. Your ship automatically fires, leaving you to concentrate entirely on movement—dodging incoming threats, positioning for survival, and navigating the increasingly chaotic battlefield. This design removes the need for precise aiming and instead emphasizes spatial awareness and timing. As runs progress, enemies become more aggressive, patterns more dense, and the margin for error smaller. The result is a steady escalation of tension that feels both manageable and intense, especially as the screen fills with hazards that demand constant attention.\n\nWhere the game truly distinguishes itself is in its progression system. Rather than relying solely on reflexes, BYTEPATH introduces a wide array of upgrades, ships, and passive abilities that can dramatically alter how each run plays out. Over time, players unlock new options that allow for different approaches—some builds might focus on survivability and regeneration, while others prioritize speed or overwhelming firepower. This layer of customization transforms the experience from a simple arcade shooter into something closer to a roguelite, where the real engagement comes from discovering and refining combinations that suit your playstyle.\n\nEach playthrough becomes an opportunity to experiment. The game encourages players to try different configurations, observe how systems interact, and gradually optimize their approach. This emphasis on build crafting gives the experience a sense of longevity that belies its simple mechanics. Even though the core actions remain the same, the way those actions feel can vary significantly depending on the choices you make. It’s this constant cycle of experimentation and refinement that keeps the game engaging over time, especially for players who enjoy exploring systems and uncovering hidden synergies.\n\nVisually, BYTEPATH embraces a retro-inspired aesthetic that complements its design philosophy. The graphics are minimal and functional, focusing on clarity rather than detail. While this simplicity may not immediately stand out, it ensures that gameplay remains readable even during the most chaotic moments. The interface and overall presentation reinforce the game’s utilitarian approach, stripping away unnecessary elements to keep the focus on mechanics. The soundtrack, while not overly prominent, provides a steady backdrop that supports the game’s rhythm without becoming distracting.\n\nThat said, the game’s simplicity can also be its biggest limitation. Because the core interaction is limited to movement and positioning, players who are not interested in the deeper progression systems may find the experience repetitive after a while. Without an investment in the build mechanics, the gameplay loop can feel somewhat shallow. Additionally, the lack of a narrative or contextual framework means that the experience is entirely mechanical, offering little in the way of story or thematic engagement.\n\nAnother minor drawback is that the visual clarity, while generally effective, can occasionally become strained during more intense sequences. As enemies and projectiles fill the screen, it can be difficult to track everything at once, leading to moments where failure feels unavoidable rather than earned. While this is partly a characteristic of the genre, it can still be frustrating, particularly during longer runs.\n\nDespite these shortcomings, BYTEPATH succeeds by fully committing to its design philosophy. It doesn’t try to be more than it is; instead, it refines its core idea and builds depth through systems rather than surface complexity. The result is a game that is easy to pick up but offers enough variation to keep players engaged over time, especially those who enjoy experimenting with different strategies.\n\nIn the end, BYTEPATH is a compact but cleverly designed experience that demonstrates how much can be achieved with a focused concept and thoughtful execution. It may not appeal to players looking for narrative depth or mechanical variety in moment-to-moment gameplay, but for those drawn to system-driven progression and arcade-style challenges, it offers a rewarding and surprisingly lasting experience.\n\nRating: 8/10\n\nID 213631890 positive False \n This game honestly looks really cool, and the style of it is great. But I am really just frustrated with the how the menu is designed, it is very unique in the sense that it is in the form of a concole command. But by golly is it hard to navigate it. Also, I am not certain yet, but I don't see an option to control the ship with a mouse, I might be blind, since the option was available on another game this developer made, SNKRX. And even whe nI turn distortions, glitching, and sound to 0, the music and static is still there after launching the game. I really want to give this game a fair judgment before I reviewed it, but my initial playthrough was just a pain. The distortions, the glitching, and effects are just too much for me.\n\nID 119044211 positive True \n You cannot have enough of this dev's games, excited to see what they pull after this and snkrx 👀!\n\nID 116097298 positive True \n I finished the game (once, of course) and went to read some reviews, maybe write another myself. Someone pointed out at the developer. I clicked the name and went into the page, and saw they have two games. I already owned the other one, SNKRX which is FANTASTIC even if it was more expensive. And I loved this one, so I think I'm a big fan of this guy now. \n\nThe game is unique, at least I didn't play anything like it. Sure it has similar ideas here and there, but the visuals and feel of the game was unmatched for me. I really liked the style of this game, and despite being short, it's cheap as hell and even if I never come back to it, It's money well spent. Not that I mind short games, in fact I like them more. They respect our time and the value it has. Of course they have to make for it in the quality, and I'm happy to say that this game does!\n\nBuy it, it's short. If you didn't like it, refund it. You'll probably can even if you finished it once (like me).\n\nID 95034038 positive True \n Though less popular and less polished than a327ex's latest game SNKRX (which I enjoy a lot), I found lots of fun with the various build options that BYTEPATH has to offer. From ships which send homing projectiles towards basically every rock, ship, and big floating square it sees, to nuclear cores that create several screen-wide explosions every second, to exploding tornadoes full of self-replicating projectiles which, depending on the powerup weapon, can bring the framerate and even the fabric of the game itself to its knees, there are many ways to play BYTEPATH. These descriptions are not hyperbole, and are actual builds I have played with. It's fun, and only like two dollars.\n\nID 94855577 positive False \n This game is essentially Roguelite Asteroids, with a purposely massive skill tree and overpowered classes. But, you'll probably very quickly end up overpowered if you carefully consider the class choices, and easily reach level 40 to \"complete\" the game. It lacks the difficulty and challenge that good roguelite games have. From there, there isn't much to do.\n\nThe 1 million point achievements are really the only other thing to go for after quickly reaching the ending, and they end up being quite a grind. Getting that many points basically REQUIRES a degen build that's nearly invulnerable and can generally reach level 80 without any user input, and at such high difficulty levels, the FPS starts to drop, to the point that it enforces a time limit of sorts to get 1 million points in before the FPS goes below 1. A game like this would be better if it could handle being truly endless...\n\nAlso, I ran into a bug where your ship gets stuck past the boundaries of the screen, which ruined a few good runs.\n\nYou could probably have a lot of fun with this if you do self-imposed challenge runs, but the main game lacks the kind of \"fight to the top\" energy a good roguelite needs, and the 1 million point achievements are boring to get. But, for the price, it can be worth a try, and I really enjoy this developer's style.\n\nFor a much more fun and balanced roguelike game with many similarities (controlling a constantly moving/firing object bouncing along a rectangular arena getting more and more powerful), try the dev's next game, SNKRX!\n\nID 94239722 positive True \n Yo real talk, I got a lil crush on whoever made this game.\nIt's the same dev as SNKRX, the little arcade-shooter I've been grinding for a short while now.\nEverything about this game oooooozes charm and passion.\nIt's very fun to explore and get over-powered!\n\nOverall score: Beeeeaaaaauuutiful.\n\nID 92476092 positive True \n Another one from a327ex, again worth every penny.\n\nJust a few phrases to describe this game:\n\n- Astroids-like\n- mini-rpg\n- large skill tree\n- polished, robust, bug-free\n- very fairly priced\n- casual\n- visually engaging\n\nID 85528080 positive True \n Learning the mechanics in play and finding their synergies takes some time but eventually you'll digest enough to navigate the game's class system and the giant, intimidating skill tree. Before long you'll be sitting back with your own broken build in a sea of particles and chromatic aberration.\n\nA very cool thing to see is that the developer wrote a tutorial detailing how to program the game for yourself: https://github.com/a327ex/BYTEPATH\n"}

**Checking review independence**

Script completed
Wall time 0.4 seconds
Output:

{"chunk_id":"e75290","wall_time_seconds":0.284102036,"exit_code":0,"original_token_count":11334,"output":"Warning: truncated output (original token count: 11334)\nTotal output lines: 391\n\n\nAPP 915310 CANDIDATES 19\n221781570 True SNKRX, created by solo developer a327ex, is a deceptively small game built around a very sharp idea: take the movement logic of classic Snake, fuse it with roguelite progression and auto-battler synergies, and strip everything down to its essentials. The resu\n158054318 True en environment, etc, etc. That's just a cheap way to add more content tbh.\nIf the dev is reading this, I like Bytepath too, but it was a lot thrown at the player at once. This game eases you in a lot better. I'm excited to see what you make next! | epath too, but it was a lot thrown at the player at once. This game eases you in a lot better. I'm excited to see what you make next!\n138574519 False Google \"a327ex website NFT\".\nClick on the first link.\n\nYou can choose for yourself if you would like to support this man. But do so in knowing that about him.\n128135772 True so pretty cool too. you can also get artifacts or relcs. I I was not aware I had their other game but i guess Bytepath just moved up on my to play list.\n122252593 True . Well worth the tiny asking price for what you get to play with :)\n\nNo issues at all on steamdeck\n\nGreat job a327ex!\n\nEdit, 24/09/2022: This game really is a lot of fun that you can easily keep coming back to. It has a lot more depth than I initially saw & the gameplay loop is super satisfying. I'm still trying to wrap my head around |  left a review after having a blast with the first couple of hours. So an edit will have to suffice.\n\nWill be keeping an eye on this developer for sure. Super fun, unique little game!\n119802659 True but the other game clearly aped some of the aspects of this game and more people should play this one. I hope a327ex does 2 things for this game: improve performance (I've lost twice now in three games when the framerate went to zero and the cpu usage spiked to +70%, my PC is high end and this should not happen) and I hope for better \n109050086 False lly against them for the reasons they say they are, and instead just don't want to take the risk. https://www.a327ex.com/posts/2022/\nIf you don't care you're giving your money to a cryptobro, it's worth the pocket change. If you do, don't touch it.\n95412717 True pecial. The soundtrack is phenomenal and sometimes I just keep playing the game because of the music. \nThanks a327ex :)\n\nyou can see that I don't have enough hours to write a full review. But it's enough to recommend this game. It's so cheap, and the irony is that it has so much more replayability than many of the well known AAA games\n95224877 True s how to string together several mechanics to make a compelling gaming experience in a small, cheap package;\n\nBytepath - A top-down shooter with several customization options through a skill tree and many ship options;\nSNKRX - A top-down shooter with centipede like controls and auto-chess type units for each segment of the centipede.  T | tems.  I easily blew through 4 hours in 1 sitting, well worth the price and then some  This is a developer to keep your eye out for more content!\n94996875 True ontent. I'm very excited for all the new stuff they have planned! Overall, a round of applause to the creator a327ex for a great game, as well as a quick thanks to Northernlion for introducing it to me via his streams :)\n94706164 True Autochess+snake from the guy that made bytepath. If that meant anything to you, get it.\n94247218 True worked. Perhaps making them buff the defense of other enemies instead of throwing them at you? I'm excited to see what's next for this game, but I'm certain that it will also make me upset.\n94239790 True Yo real talk, I got a lil crush on whoever made this game.\nIt's the same dev as BYTEPATH, a little arcade-shooter that's filled with love.\nEverything about this game oooooozes charm and passion.\nIt's very fun to explore and get over-powered! (And getting your ass handed to you.)\n\nDirt-cheap, wit | Yo real talk, I got a lil crush on whoever made this game.\nIt's the same dev as BYTEPATH, a little arcade-shooter that's filled with love.\nEverything about this game oooooozes charm and passion.\nIt's very fun to explore and get over-powered! (And getting your ass handed to you.)\n\nDirt-cheap, with a great so\n93983393 True ing, the game's creator also seems like someone who really loves games. I caught up on the developers blog at a327ex.com and he seems like a really genuine person who's invested in creating cool games. I think this game has a lot of untapped potential that it will reach in time. Not Anthem roadmap ♥♥♥♥♥♥♥♥ type potential, but real tun\n93920127 True utochess, but Arcade Shooter. Fast. Chaotic. Fun. Someone should do this but with Tower Defense. If you do it a327ex, I will definitively buy it.\n\nUpdates are constant but I do have one QoL request: \n\nINFORMATION. I would really like to know how my units are doing and what are they doing. Is one of them useless? Is the whole team usel | ack speed of whatever\", \"if this unit is at the front it gets +10 defense\", stuff like that.\n\nI can't wait to see what comes next. GG a327ex <3 | \"if this unit is at the front it gets +10 defense\", stuff like that.\n\nI can't wait to see what comes next. GG a327ex <3\n93540628 True n a creative idea, endless mode with higher level would be great but otherwise fantastic :)\n\nIf you'd like to see what we're playing next with regular updates, or request to review a game, feel free to follow us at: \nhttps://store.steampowered.com/curator/35688508-Ember-Works-Reviews/\n93497676 True 9CPns\n\nSNKRX is a weird chimera of an auto-battler, a rogue-lite, and the classic arcade game SNAKE. As in [b]Bytepath[/b], a327ex's prior game, SNKRX merges a simple two-button control scheme with a layer of skill-building strategy to make an entirely engaging and addictive coffee break game. \n\nNG+ unlocks another segment for your batt | s a weird chimera of an auto-battler, a rogue-lite, and the classic arcade game SNAKE. As in [b]Bytepath[/b], a327ex's prior game, SNKRX merges a simple two-button control scheme with a layer of skill-building strategy to make an entirely engaging and addictive coffee break game. \n\nNG+ unlocks another segment for your battlesnake and \n92900640 True n\" driving, and an absolute banging soundtrack by Kubbi brings a complete package of a game. Highly recommend following a327ex to see what they have up their sleeve next. I can't imagine it won't be long until, if they desire, they'll be working for a larger studio with a bigger budget. Excited to see what comes next.\n\nFantastic game.\n\np.s. onl | bsolute banging soundtrack by Kubbi brings a complete package of a game. Highly recommend following a327ex to see what they have up their sleeve next. I can't imagine it won't be long until, if they desire, they'll be working for a larger studio with a bigger budget. Excited to see what comes next.\n\nFantastic game.\n\np.s. only found out about this game by watching Dan Gheesling's twitch channel #596\n92150794 True By the dev who did Bytepath. Pretty interesting and unique game.\n\nPros:\n-Good music.\n-Lots of cool abilities and combos\n-Art is simple and easy on eyes.\n-Many enemy types only based on colors.\n-Runs are different every time based on class choices.\n\nAPP 716500 CANDIDATES 106\n225320384 True imely this seems even now. I can see parts of this storyline becoming reality. Maybe it already has. My first Zachtronics game, I will have to check out the others.\n222456195 True  sound design, etc. were all very satisfying to interact with throughout.\n\nGood story and characters too. Big fan of Evelyn.\n\nCertainly worth checking out when it's on a sale. $15 normal price isn't terrible, but it's a little high, in my opinion.\n\n6.7/10 experience.\n221167331 True cholic and weirdly inspiring story about technology, the people who make it, and people it affects.\r\n\r\n10/10, Zachtronics did it again.\n202371170 True In 2019, Zachtronics' developers, famous for their well-liked card and puzzle games, decided to write and draw a visual novel about technology's relationship with us, society, and our relationship with each other. To call 'Eliza' a great vi | In 2019, Zachtronics' developers, famous for their well-liked card and puzzle games, decided to write and draw a visual novel about technology's relationship with us, society, and our relationship with each other. To call 'Eliza' a great visual novel, written with a distinct touch of vulnerability and \n192557559 True It is still so weird that Zachtronics made this. If I played this when it released I probably would've thought it far fetched to see people latch on to an \"AI\" chatbot that just reflects what you tell it back in the form of a question. But playing it today \n191993380 True This was great, specially when they throw a twist in the plot. GG  Zachtronics.\n191755824 True I'm not a fan of visual novel, but this one is different, specially if you are into tech.\n\nThe fact that the choices have no impact until the last chapter could be a negative point, and the ending could be a little more complete (even i\n190599746 False  I'm glad Eliza was made, but it just didn't land for me. I've played and enjoyed (and often failed at) other Zachtronics games, and I knew what I was getting into when I bought this one on sale. It's not a puzzle game by any stretch (solitaire notwithstanding), it's a pure visual novel.\n\nVisual novels are already not my style of game, but | bout the game itself but are really curious to try out the solitaire variant, I recommend instead buying \"The Zachtronics Solitaire Collection\".\n169782009 True preference, though it does come with one other piece of content being the classic kabafuda solitaire for this Zachtronics game. That being said I truly enjoyed the story and exploration of the characters immensely and it definitely left a lasting impression on me. The one drawback is that initially a lot of choices in the story are surface\n165244803 True I enjoyed other games of the same developer, but it is the first VN I played from them. I just hope they do more VNs with AI theme or even better maybe a cyberpunk VN. I am really fan of their games at this point. | played from them. I just hope they do more VNs with AI theme or even better maybe a cyberpunk VN. I am really fan of their games at this point.\n147041693 True  lack of control is used to brilliant effect to say what it wants to say. It's insane to me how versatile the ZachTronics team is at making games in many genres that force you to think critically about the subject matter and come away with your own solution, puzzle or otherwise. Highly recommend.\n145812937 True I started playing this because I love the Zachtronics puzzles games, and found myself unexpectedly invested in their stories. I don't usually play graphic novels, but I enjoyed this one. I liked the writing, and the story of the protagonist struck a chord with me personall |  counseling, it had a remarkably down-to-earth feeling, almost slice-of-life, similar to the stories from the Zachtronics puzzle games.\n\nIt's not a ground-breaking piece of literature, but certainly an interesting experience if you're interested in the story's themes, and I'm glad I played it. |  it had a remarkably down-to-earth feeling, almost slice-of-life, similar to the stories from the Zachtronics puzzle games.\n\nIt's not a ground-breaking piece of literature, but certainly an interesting experience if you're interested in the story's themes, and I'm glad I played it.\n144889034 True [i] Zachtronics published a visual novel. Not something you see every day. [/i]\n \nEliza is a visual novel that covers psychology and artificial intelligence. Not to brag, but I have degree in psychology and am currently working at an A\n141353521 True  the main character and her effect on the world.  Pleasant soundtrack as usual from Matthew S. Burns, and the Zachtronics solitaire game is fun, too.\n137708435 True I didn't assume i'd enjoy a choose-your-own-adventure game, but this felt prescient. Zachtronics games never go on sale, and it's not necessarily worth upwards of $2/hour as a gaming experience, but as a literary work with a variety of perspectives to share, i recommend this game to anyone who might be interested i\n122144064 True Learnt about this game from Jacob Geller. Didn't know Zachtronics made a VN. I hope more people will find it.\n\nAs with pretty much all Zachtronics games, very high levels of polish, and even full voice acting.\n\nIt has some interesting things I haven't seen in other VNs, but it also do | om Jacob Geller. Didn't know Zachtronics made a VN. I hope more people will find it.\n\nAs with pretty much all Zachtronics games, very high levels of polish, and even full voice acting.\n\nIt has some interesting things I haven't seen in other VNs, but it also doesn't feel as interactive, as it seems that whichever ending you get is decided a\n114924745 False  myself to recommend it.\n\nIt has great graphic design and good voice acting, but that's about all. I'm a huge fan of visual novel+…4834 tokens truncated…ce on a game like this. I very much enjoyed my time with the game, and would like to see Zachtronics produce |  see them take a chance on a game like this. I very much enjoyed my time with the game, and would like to see Zachtronics produce more games like this!\n54932254 True man, and finding happiness. Definitely the best game I've played so far this year, with a typically fantastic Zachtronics solitaire game as a bonus! I'm really glad this game got made, and that I got to experience it.\n54865855 True [i]Eliza[/i] came out of the blue for me; I never would've expected Zachtronics to release a visual novel, of all things. But, having played it through and explored every available scene and ending, it makes a lot of sense.\n\nThis story touches on a lot of interesting (and topical) subjects, includi\n54860514 True s of technology merging into the field of mental health while also tackling the main character's issues. As a fan of Zachronic games, I'm amazing how they are able to incorporate their common themes of technology and development and keep that intact into this narrative experience. I highly recommend this visual novel if you are at all\n54846429 True I love Zachtronics puzzle games, and this game engages deeply and sincerely with many of the themes and ideas that have been lightly floating on the surface of those games. Like those puzzle games, this is a world of the mundane made fant | I love Zachtronics puzzle games, and this game engages deeply and sincerely with many of the themes and ideas that have been lightly floating on the surface of those games. Like those puzzle games, this is a world of the mundane made fantastic, but un | y with many of the themes and ideas that have been lightly floating on the surface of those games. Like those puzzle games, this is a world of the mundane made fantastic, but unlike them it is a journey of reflection rather than problem-solving. Outside of some tough choices (and the hardest Zachtronics version of solitaire yet), it is a ga\n54845213 True ings I've heard in my day to day life that it pulled me into the game's storyline even more. \n\nAlso, wasn't a fan of Solitaire before. But after solving it at the Expert difficulty within the game, I genuinely had a smile on my face. \n\n11/10 would recommend.\n54818973 True It's nice to see Zachtronics move from niche puzzles to addressing something we are all facing - the ethics of Artificial Idiots and the fact that people are more and more disconnected from each other as the \"progress\" treads further.\nA game with s\n54807259 True Wow, a game from Zachtronics that isn't heavy on technical rules and mechanics, just a story to click through. The story is sci-fi, but incredibly close to the world we live in.\n\nIt a great reflection on the state of the world: the modern epidemic \n54770083 True l Health, Ethics and Technology in the Modern World[/h1]\n\nThis was a completely unexpected piece of work from Zachtronics given their past games, but it is surprisingly well done and it couldn't have been more timely.\n\nIf you are a tech worker in your late twenties / early thirties like so many of us these days, this game will speak to you\n54729718 True Zachtronics has traditionally created engineering puzzle games -- games that often come frighteningly close to what I do for a living, but still manage to entertain.  Eliza is a departure.  It explores more the consequences of engi | Zachtronics has traditionally created engineering puzzle games -- games that often come frighteningly close to what I do for a living, but still manage to entertain.  Eliza is a departure.  It explores more the consequences of engineering and technology rather than simulating the m\n54725239 True to avoid providing these additional portraits to support the emotive and well performed dialogue, \n\nI applaud Zachtronics for venturing into unfamiliar territory, but I hope that if the developer works on another title of this ilk that they pay more attention to creating a true branching story, one in which the decisions of the \"player\" ch\n54724543 False I'm a huge fan of choice games. I thought this would involve more meaningful choices, but it didn't. \n\nThe voice acting was great, albeit a bit slow going. \n\nLike someone else said, you have to do a lot of clicking that doesn't really ge\n54724094 True Great g̶a̶m̶e̶ visual novel very much unlike any other Zachtronics genre so far but equally great in polish and love. Of I'm biased since I love the puzzle games but Eliza is more than my other most favorite game series (Life is Strange). So I'm all for games heading into that directio | nlike any other Zachtronics genre so far but equally great in polish and love. Of I'm biased since I love the puzzle games but Eliza is more than my other most favorite game series (Life is Strange). So I'm all for games heading into that direction but also being different.\n\nThanks so much for supporting SteamOS/Linux again! <3\n54692960 True en't experienced a lot of visual novels like this, but it was a rewarding experience.  Very polished like all Zachtronics stuff - fully voice acted, great soundtrack, immersive experience.  I hope that it pays off taking a risk going in a new direction.\n\nThe themes of Eliza fit very well as visual novel, and I really found myself intensely\n54688217 False A visual novel from zachtronics.. I actually quite like the story elements of their games, especially in Opus Magnum. But Eliza is not good.\n\nIt starts off well, with a little sinister twist in the first Eliza session. You think it's going somewhere i\n54667760 True lly trained musician who's been diving into EDM... I've purchased, played, and loved the previous releases by Zachtronics, but this one is very different from past offerings. Some may comment about the length, but I've paid more to see a movie. Eliza is well written, engaging, and it definitely resonated with me personally. Another fantast | liza is well written, engaging, and it definitely resonated with me personally. Another fantastic offering by Zachtronics at a reasonable price. At this point I'll buy anything they put out, sight unseen.\n\n-LeRoy3rd\n54662765 True The first Zachtronics game I actually completed\n54658843 True 9/10 , I'm fans of Zachtronics\n\nAdditional Details\n\n[table]\n    [tr]\n        [td]OS[/td]\n        [td]Ubuntu 18.04 bionic x86_64 Linux 4.15.0-55-generic[/td]\n    [/tr]\n    [tr]\n        [td]GPU[/td]\n        [td]GeForce GTX 1050,nvidia-390.116[/td]\n    \n54654463 True PLEASE do yourself a favor and play this game. Zachtronics have deviated from their great exploration of niche problem-solvers to create a masterpiece. Experiencing Eliza will take you an afternoon and you'll leave as a smarter and wiser human being. One of the best arguments t\n54646438 True I don't often do reviews, but I enjoy all the games Zachtronics have put out.\nI thought this was a wonderful, thought-provoking story. This and Red Strings Club are two games that I finished playing and said \"That was a wonderful story.\"\n54617758 True As a big fan of visual novels and a casual follower of Zachtronics, I was really surprised by how well everything in this game came together to reinforce its central theme of personal alienation. The writing is great-- the person-to-pe | As a big fan of visual novels and a casual follower of Zachtronics, I was really surprised by how well everything in this game came together to reinforce its central theme of personal alienation. The writing is great-- the person-to-person interactions and dialogue alone put it in the \n54607295 True f extra storytelling needed to turn it from a decent, unique read into something that's an absolute must as a fan of the genre, or a fan of modern-problem storytelling in general.\n\nI'd say, give it a read & come to your own conclusions - the game is definitely deserving of that much. | eded to turn it from a decent, unique read into something that's an absolute must as a fan of the genre, or a fan of modern-problem storytelling in general.\n\nI'd say, give it a read & come to your own conclusions - the game is definitely deserving of that much.\n54604246 True out any reviews being posted.  I would never do that for most developers (and especially for a VN), but given Zachtronics' record and how I hold their games in such high esteem I didn't deliberate.\n\nEliza has mostly been enjoyable (I've played it through once which took six hours).  High notes that I would draw attention to is the excellen | 've gone into way more detail on the negative than positive here which might be unfair.  It's a bold move for Zachtronics and does show how far they've come that they can \"pull off\" a VN, but judged as a game or as entertainment on it's own merit I don't find it as satisfying as their other stuff.\n\n6/10\n54600519 True I've had my expectations high for this, since I love both Zachtronics puzzles and visual novels. The writing is clearly influenced by the recent data breaches at a certain tech giants, so the ethics of collecting huge amounts of personal data and critique of \"move fast and break things\" m\n54599215 True This game, unlike any other Zachtronics game, isn't based on a logic-driven puzzle mechanism, but in the form of a Visual Novel. The voice-acting in this game is one of the best that I've came across recently, and the art design is also unique and beautiful.\n\n54595011 True  the next five years.  The game has a satisfying end, yet I would still love to see more from this world that Zachtronics has built. \n\nOn a side note; I feel like this game stole from my life. [spoiler] Born n raised in the PNW, I got a Computer Science degree and landed a lucrative consulting gig.  After a few years it left me incredibly  | ler] Even though I don't make as much anymore, my life is so much more fulfilling and satisfying. \n\nGreat job Zachtronics. Can't wait to see what you come up with next.\n\nAlso please release the soundtrack, it's fantastic.\n54592522 True tl;dr: Zachtronics are the exact opposite of EA. Everything they touch becomes great.\n\nLet's get this out of the way first: with the exception of [spoiler]a new kind of solitaire (as is now a tradition)[/spoiler], this game is a visual no | hat's pretty good for a VN, though of course it's extremely short compared to what one comes to expect from a Zachtronics game.\n\nA common problem with VNs is that the added interactivity, images and sounds of the medium tend to serve to hide mediocre writing. Boy, was that not the case here at all. This story could very well be written on  |  helps placing you inside the story being told.\n\nAll in all, if you were drawn by the story elements in other Zachtronics games such as Spacechem, Infinifactory, Schenzen I/O and especially EXAPUNKS, then this VN is for you. If you like VNs in general and anime style is not a requirement for you, you'll love it. If you absolutely hate VNs \n54592158 True As a long time Zachtronics fan, Eliza to me just cements what I already felt about the team.... It's not just their puzzle creation ability and superb presentation that makes their games so engaging. It's their choice of underlining subject matte | hat, and the difference in the end result as the end user really can be felt.\n\nI don't think there has been a Zachtronics game yet that felt like the game itself wasn't a purely natural extension of the subject matter (Infinifactory came kind of close, but lets forget about that one for now hehe), and Eliza still doesn't disappoint.\n54590269 True Something completely different from Zachtronic. Really enjoyed it. Probably some of the best voice acting I've heard in a game in a long time. Adorable and really likable characters. Interesting story. I just feel it's a bit too short though to really do the themes \n54587854 True Amazing!\nLike any Zachtronics game this is just perfect. And this one especially! Zach has always been great about hitting that mood or feel. This is even more true and more central to this game than any of his other games.\nThis is such a great game | n great about hitting that mood or feel. This is even more true and more central to this game than any of his other games.\nThis is such a great game to really let yourself fall into and just enjoy. It explores so many topics in all the right ways. Giving you the right nudge to think about a lot of stuff without oversharing. Approaching eve\n54585391 True This game proves that the \"Zach-like\" part of Zachtronics' games are not \"just\" the puzzles and quickly ramping complexity / difficulty.\n\nI feel like this game is the result of refining every part of the \"meta\" game found within recent Zachtronics releases. \n\nTake away the puz | ifficulty.\n\nI feel like this game is the result of refining every part of the \"meta\" game found within recent Zachtronics releases. \n\nTake away the puzzles and games like Opus Magnum, Shenzhen IO, and EXAPUNKS still have a compelling story, sharp visuals, and an air of intrigue.\n\nFrom the first half hour of the game, my impression is alrea\n"}

**Checking review evidence**

**Checking review evidence**

**Checking discovery-count history**

**Checking review evidence**

**Checking discovery-count history**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"c74228","wall_time_seconds":0.000002025,"exit_code":0,"original_token_count":2832,"output":"\nCASE 0 BYTEPATH 40343694 \n fun asteroids good music, \nI like the console gimmick.\nNode picking is a bit annoying.\n\nCASE 1 BYTEPATH 40349141 \n Really cool little game that encourages experimentation.\n\nCons:\n-Can get repetitive\n-Limited variety in environment (as in, there's only one)\n-Screen can get messy\nPros:\n-Screen can get messy >:]\n-Decent replayability\n-Great music\n-MANY upgrade paths\n-Upgrades feel like Borderlands meets Path of Exile\n-You can turn down screen shake\n\nI would absolutely recommend this game, particularly for the price point. An extended version of the game with more environments, zone progression and more (maybe randomized?) class upgrade options and ships would be excellent to see.\n\nCASE 2 BYTEPATH 40364921 \n Game is incredible for 2 bucks, defenitely worth it. It is pretty sad you have to work around the crashes though.\n\nCASE 3 BYTEPATH 40368481 \n PoE plus retro arcade shooter. For two bucks, it's a must-buy.\n\nCASE 4 BYTEPATH 40370057 \n Great. I love it.\n\nI think i almost fried my PC though. After getting upgrades and all that, Projectiles flying all over the place, explosions everywhere, zero FPS basically.\n\nGreat fun though, was laughing the whole time.\n\n\n\nCASE 5 BYTEPATH 40371253 \n If you want to feel overpowerd in a pixel world this is the game for you. Can't really complain about anything but minor bugs & screenshake on explosion.\nThe game itself is fine for the price of 1,99€\n\nCASE 6 BYTEPATH 40378529 \n Interesting Asteroids style game with a ton of mechanics going on. Definitely worth it at $2 if it looks like your sort of thing.\n\nI highly recommend toning down some of the visual effects.\n\nCASE 7 BYTEPATH 40402572 \n Asteroids meets Path of Exile. \n\nThis game is good for 15 minutes or 15 hours. There is a depth of possibilities when choosing your skills, and your waking thoughts will be plagued by new ideas of how to build a character. Replayability is excellent.\n\nCan be buggy at times. The game may crash if your build spawns obscene amounts of exploding projectiles.\n\nGo sell your crappy TF2 hats to get the $2 to buy this game!\n\nCASE 8 BYTEPATH 40406931 \n Friend suggested the game, said it was only 2 dollars, so what the heck.\n\nDamn, this is actually a really good game.\n\nI love the console theme and the visuals. The soundtrack fits the game and is good in it's own right. The customization options are insane, you really can do a different crazy ship every time, tons of ways to change up the gameplay.\n\nAs for the negitives, one game enviorment and limited enemy variety. There is also a lot of visual clutter (but I think that's part of the charm.) It may also be a bit too easy? I was able to escape on my first build.\n\nYou can get a lot of mileage with two dollars if you buy this game. Even with my two hours so far, I feel it was worth the purchase.\n\nCASE 9 BYTEPATH 40440688 \n A fun game with a solid pitch and decent execution. An absolute steal at the price. I do think it's missing an essential feature though: A search button on the passive tree. Having to comb through the entire tree to find the one minor node that gives you the effect you need to complete your build is infuriating.\n\nCASE 10 BYTEPATH 40597779 \n This was more fun than i anticipated.\nyou have several different \" device\" ships, a gigantic skill tree thats the honey-comb looking thing you see in the screenshots, it's HUGE.\nyou also have \" classes \" each tier has several to choose from, all have positive and negatives - huge customization, each level brings more classes and the cost goes up/\n\nAnd all of these things, can be re-rolled, and done again.\n\nIt's almost endlessly replayable in this way.\n\nThe main game is a simple Asteroids - type game. You rotate the ship, it propels at it's own velocity (depending on what device type it is, and all the passives you have applied, plue class benefits ) and it auto fires as well under the same rules.\n\nThen, there's barrages ( a huge rapid fire burst of bullets from you ) a shield ring ( more rings = more shields ) \nand a bunch of different weapon types you pickup as powerups in the game, and also these nodes that give you benefits as well !\nThere is a lot going on. It's a slow boil, but when you start boiling.. it's very hard to put down. \n\nit has a good \" just one more round \" gameplay loop going for it.\n\nabsolutley recommend.\n\nCASE 11 BYTEPATH 40599500 \n Worth substantially more than the $2 asking price. Replayable arcadey shoot-em-up with a Path of Exile style passive tree plus ship and multiclass selections make a very simple gameplay loop feel like it has a LOT of depth.\n\nCASE 12 BYTEPATH 40865529 \n Very short and takes a bit of getting used to, but definitely gets to the point real fast.\n\nPlay this if the latest Path of Exile league is a dud.\n\nCASE 13 BYTEPATH 41120286 \n Came for the tech tree, stayed for the terminal.\n\nCASE 14 BYTEPATH 41403451 \n Wow. I heard about this game on the Path of Exile subreddit because it is so heavily inspired by it. I never would have thought that what is essentially an Asteroids clone combined with PoE's customization could possibly work, but it absolutely does.\n\nFor a mere 2 bucks you can't go wrong with this purchase. For reference, it took me 11 hours to 100% the achievements, but I definitely dragged my feet a bit because I was enjoying certain builds so much. The possibilities and sense of discovery when you find awesome new combos is exactly what I was looking for in a new game, but nothing could ever compare to PoE in that sense. This game does.\n\nI challenge anyone who ends up buying this to create the ultimate build: something that scales infinitely and can't die, but also won't crash the damn game. That's the one fatal flaw of Bytepath. You can make an invulnerable OP build that will eventually die to a game crash because the infinitely scaling enemies and millions of projectiles just fry your PC no matter how much of a beast it may be. As much of a flaw as that is, I also look at it as an added challenge. How can you not only survive and kill enemies, but how can you do it without spamming so many projectiles or leaving too many enemies on screen that would make the game crash? If anyone figures it out please let me know!\n\nMy favorite playstyle was stacking boost and cycle speed, and a few crucial keystone passives I'll let you discover on your own! Buy the game if you're a PoE theorycrafting nutcase like me. If you're not into theorycrafting your own builds in ARPGs then the game probably isn't for you, because gameplay-wise it isn't that complex. The magic is in the menus!\n\nCASE 15 BYTEPATH 41513713 \n Incredibly fun, satisfying game. Worth every penny.\n\nCASE 16 BYTEPATH 41605279 \n The gaming equivalent of an epyleptic on LSD shoving a pair of lit firecrackers down his trousers. Hours of mindless fun.\nHighly recommended.\n\nCASE 17 BYTEPATH 42939775 \n This game is great. Check the other reviews for details as to why. I just want there to be more positive reviews on this baby so it becomes more popular.\n\nCASE 18 BYTEPATH 42968840 \n A great asteroids clone that can scratch the itch of creating a build in Path of Exile.\n\nCASE 19 BYTEPATH 43041360 \n This is a great little game. The gameplay is really fun and the overarching powerup system is fun to play around with. The vibe is also great and makes it easy for you to lose yourself in the gameplay. Play it!\n\nCASE 20 BYTEPATH 43594238 \n Addictive, well-rounded extension to the 'asteroids' formula that sees you upgrading your space ship until you survive long enough to reset the game. Lots of depth, replayability (very distinct kinds of builds and paths through the tech-tree and class system).\n\nMy only constructive criticism would be that not much is explained at all in the UI - having beaten the game three times, I still don't know what some of the bars represent, or vague/typo'd terminology is \"area\", \"energy shield\" etc. Didn't take away from my fun, did enjoy inferring some stuff, and ultimately for the price you can't go wrong.\n\nCASE 21 BYTEPATH 43596793 \n 9/10: BYTEPATH takes the blasting fun of Asteroids, styles up the vector graphics with some glitchtastic visual options, and extends the gameplay with an expansive amount of upgradable options. Can you escape the terminal?\n\nBytepath places the player in the role of an AI in a simulation that feels an awful lot like Asteroids. Your only means of escape is the keys, but surviving long enough means building up your device with the right classes and passives to suit your play style. Do you prefer surrounding yourself with shield projectiles? Using boost to get through? Focusing only on lightning and nothing else? All of these options are feasible through the passive tree. While it isn't easy to figure out what every node in the passive tree means at a glance, the letter and color convey a pretty good gist: red H is usually something about HP, orange A has some specific attack focus, yellow S relates to SP points, and so on. While you can deselect modes, be careful to not commit prematurely: you aren't refunded an SP if you deselect one after you commit to a build.\n\nClasses provide a similar sort of customization to passives, though are generally more extreme and unrefundable once selected. Things like Tank increase health and size but slow you down, Raider gives you more resource drops, Absorber prevents damage at the cost of boost, with players being able to buy only a few classes from a set before needing to choose from completely new ones. \n\nThe devices are another way to switch up your play style, though it's free to switch between them once unlocked. Thankfully, you keep all the same classes and passive tree no matter what you switch to, though that also means you only really have one build saved at a time. Still, the ability to respec the passive tree and the boost you get to your base points every time you complete a cycle (finding all the keys and escaping), so there are plenty of ways to add variety to each replay. \n\nEven with the cyberspace context, the game definitely plays like an updated Asteroids. Zipping around, destroying floating debris and hostile ships can feel very zen when your build is working smoothly, but can turn hectic as the challenge level grows and the screen fills up. As a result, BYTEPATH feels like a juiced up version of a retro arcade classic, and a pretty damn good one at that.\n\nCASE 22 BYTEPATH 45625501 \n Great game, good soundtrack, a giant skill tree, lots of classes, and fun gameplay, and an interesting main menu design and art style. Its deffinitely worth the price, I'd write a longer review but this is all I can really say, Its simply that.\n\nCASE 23 BYTEPATH 47592343 \n When Bullet-hell meet Path of Exile, this is the product. At first it may not too appeal, but multiple ship selection plus the Class selection, and mega big skill tree, All makes the game full with replay value. Thinking the combo from ship, classes, skilles in every gameplay just for fun.\n\nPro: Mega Skill tree, Unlimited Replay value.\n\nCon: CRT like screen (is this a con? R U kidding?)\n\nCASE 24 BYTEPATH 47735547 \n As someone that doesn't really like indie games or Geometry Wars I couldn't believe how much Iv'e enjoyed this game. This game is seriously kickass, IMPORTANT: you can turn off the glitch and shake effects in options!!\n"}

**Refining topic labels**

**Refining topic labels**

**Refining coding topics**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"865300","wall_time_seconds":0.000002304,"exit_code":0,"original_token_count":2044,"output":"\nCASE 25 BYTEPATH 47923174 \n It's like Path of Exile,\nbut with more build variety!\n\nCASE 26 BYTEPATH 49347140 \n A game I can play with my left hand only, while I smoke.\n\nCASE 27 BYTEPATH 49756268 \n Hidden gem alert! Don't sleep on this one. It looks like a regular arcade shooter, but everything about this game escalates quickly and in the most beautiful way. The best two coins you can insert, free play forever.\n\nCASE 28 BYTEPATH 52003956 \n Whooooo boy the aesthetic of this game.\n\nIf you like Asteroids/Geometry Wars, Synthwave, and progression systems this is for you.\n\nwhile the gameplay is simple it is rather well done (perfect for pickup-n-play), but there is a suprising amount of depth with the game. \nfor the price you cant beat it. \n\nP.S.(Sentinal with the lighting weapon will take you far.)\n\nP.P.S.(While the sound design is great I recommend playing some RetroSynthwave (something like Kevinsky))  [;-)]\n\nCASE 29 BYTEPATH 52307741 \n A very confusing but fun arcade game.\n\nThe Good:\n- Awesome visuals.\n- Fun gameplay.\n- Tons of upgrade paths, not only do you have plenty of classes and ships, or \"devices\" to pick, you get a massive upgrade tree where you can pick up to 80 smaller upgrades at first, and after you beat the game you can pick even more.\n\nThe Okay:\n- Music is fine.\n\nThe Bad:\n- Very few of the available upgrades make sense, nothing is explained. I still don't know what happened and why. There are stats such as \"area\" and \"waviness\" that are not explained, and there's some sort of cycle thing going on during the gameplay that's never explained either.\n\nCASE 30 BYTEPATH 52509439 \n cool and short game\n\nCASE 31 BYTEPATH 53696937 \n If you're a fan of games like Path of Exile that feature endless build customization, I wholeheartedly recommend Bytepath. The depth of meaningful customization is everything I want from a game. If you want to enjoy this game for as long as possible, don't search up any guides or builds at first, and just kludge your way through it until you find your perfect build.\n\nCASE 32 BYTEPATH 54076185 \n A deceptively simple asteroids-like game, with a skill system nearly as deep as path of exile. Highly recommended, especially for the price\n\nCASE 33 BYTEPATH 55530255 \n really fun game for the price. you may not get a ton of playtime but you'll enjoy it, even if you don't like these sorts of games.\n\nCASE 34 BYTEPATH 60377825 \n path of exile but fun and made in 1971\n\nCASE 35 BYTEPATH 60644313 \n A very good game for the price. Bytepath has a nice aesthetic, with the music, terminal-esque menu, and CRT glitchy screen. Absolutely worth the price point at full price, and if you find it on sale, you are doing yourself a disservice by not buying it.\n\nPros:\n- Good aesthetic\n- Solid gameplay\n- Nice soundtrack\n- Overexpansive passive system that is fun to crawl through.\n- Simple class progression.\n\nCons:\n- \n- Nothing glaring came to mind after an hour of play, especially considering it is $2 max price. A bit of QoL that could be polished, but I'm not going to complain.\n\nA much better game than usually found for this cheap.\n\nCASE 36 BYTEPATH 63518294 \n Bizarrely addictive game. Simple idea, you can pick it up and play for 2 minutes or 2 hours and every time is different!\n\nCASE 37 BYTEPATH 79253453 \n Fun modern asteroids clone with a ton of power ups, upgrades and a branching skill tree. After you unlock a bunch of stuff it`s such complete carnage on screen visually you won`t even know where you are most of the time. This can be both highly enjoyable and also a little frustrating but the game seems aimed at a shorter playthtough so it`s easily forgivable.\nI`m not sure about any real longevity but I likely got it in a bundle so for the price it was a blast.\n\nCASE 38 BYTEPATH 84963485 \n A very nostalgic game, definitely reminds me of the asteroid days.\n\nCASE 39 BYTEPATH 94038450 \n Short game, but lets you do some really fun broken builds.\n\nCASE 40 BYTEPATH 94773147 \n its a good game, ruined by an extremely bad UI/UX\n\ni recomend for the cheap price, but if it is not on sale , it is probably not worth it\n\nCASE 41 BYTEPATH 95979041 \n Now an open-source game that needs some patching love.  It still is fun for the value of the game.\n\nHowever, this game committed a cardinal sin I can't forgive.  While literally 5SP away from finishing my first real OP class, I accidentally hit \"enter\" on \"escape\".  Without warning it reset my classes and passives and restarted the game.  All my work down the drain.\n\nCASE 42 BYTEPATH 115729972 \n Stuck in a faulty computer interface from 1980. Where's the damn game ? Taking the P. Those reviews must be the Devs. Cheapness is no bonus when the product don't work. Bah ! Owned this for two years  before discovering it's unplayability\n\nCASE 43 BYTEPATH 121586517 \n Too small anduncomfortable interface. And waaaay to many options for building the ship, you can get the feeling that you have to study the game before playing.\n\nCASE 44 BYTEPATH 122124620 \n Worth the price of admission\n\nCASE 45 BYTEPATH 141282832 \n BYTEPATH is an unusual specimen.  At face value this is basically an Asteroids ripoff, but the focus of the game isn't on the gameplay.  They've taken a shallow, ancient game and bolted a gigantic skill tree, upgrade and progression system onto it.  Architecturally, that's good game design, but in terms of big picture, it's a huge waste of time and effort, because at the end of the day, this is an extensive body kit for a Lada.  Lipstick on a pig.  It's just Asteroids and isn't all that much fun to play, no matter how much upgrading you do.  Developers really need to consider if it's worth doing what they are doing before they waste all this time doing it.  \n\nNow if they took the same skill tree and applied it to a real game, we might be onto something.  But we're not.  It's just Asteroids, you can play that for free, so there's no need to buy this.  From a technical perspective, the game doesn't meet basic minimum requirements that most PC gamers expect as standard.\n\nA choice was made to use obsolete, decades old retro pixel \"art\" as a substitute for contemporary PC graphics.  It's unclear if this is due to lack of budget or talent, regardless, the overall visual quality of the game is extremely low as a result.\n\nThere's an ugly screen filter meant to make the game look even more obsolete... like the game looks bad enough already, but they actually went out of their way to make it look worse.  There's no way to turn this off.  There's also ugly glitch effects and so on, but you can reduce those... it's unclear why the developer tried so hard to make the game look ugly.  \n\nWhile there are options to change the resolution for the game, all this does is scale up the simplistic 2D art assets used to make the game, which makes little or no difference to the graphics quality.  Without any other substantial graphics tweaks, it's not possible for gamers to improve the lacklustre 2D visuals.\n\nSome of the defects in the game can be attributed to the choice of using the video game construction kit/toolset, LOVE (a set of libraries so you can make crap 2D games with Lua).  This is a very poor quality toolset that unfortunately has very limited capabilities.  Just as you can't make a silk purse out of a sows ear, you can't make a great video game if you use a terrible engine.\n\nThese technical defects push this game below acceptable standards for any modern PC game.\n\nBYTEPATH is relatively cheap at $2 USD, but it's not worth it. Given the defects and quality issues with the game, coupled with the unrealistic price, this is impossible to recommend.\n\nCASE 46 BYTEPATH 150779168 \n Don't expect a path of exile replacement, but at 72% off, this is a pretty insane value for roguelite fans\n\nCASE 47 BYTEPATH 152900464 \n This is a fun little game. For 2 bucks, try it.\n\nCASE 48 BYTEPATH 214428866 \n if you dont like me at my fighter you dont deserve me at my striker\n\nCASE 49 BYTEPATH 215194662 \n A good roguelite with a massive skill tree to unluck but with limited resources to allocate throughout it. The music, vector graphics and terminal inputs are my kind of sugar-free jam, and I hope to beat it some day.\n"}

**Reviewing feature labels**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"ddc108","wall_time_seconds":0.0000032,"exit_code":0,"original_token_count":1127,"output":"\nCASE 50 SNKRX 92351640 \n An excellent and cheap entry into my surprisingly beloved build-simulator-stapled-onto-a-simple-game genre\n\nCASE 51 SNKRX 93454024 \n SNEK\n\nCASE 52 SNKRX 93591171 \n Great game, you are danger noodle\n\nCASE 53 SNKRX 93611317 \n Don't let the low price fool you, this game is absolutely amazing.\nPrice could at least be doubled and it would still be worth it.\n\nCASE 54 SNKRX 93814271 \n If you are a fan of roguelikes and/or auto-battlers this game is 100% worth the cost. I think it would be worth a few times the cost even. Highly recommended.\n\nCASE 55 SNKRX 93977056 \n fun\n\nCASE 56 SNKRX 94023061 \n Awesome Snake sequel :D\n\nCASE 57 SNKRX 94223863 \n Great soundtrack, great price, great gameplay and replayable. What more do you want from me egg?!\n\nCASE 58 SNKRX 94240942 \n Addicting, great music, and a much more complex strategic depth than you'll initially realize. For just $3, SNKRX is well worth it.\n\nCASE 59 SNKRX 94328695 \n this is the most fun I had with 3 bucks\n\nCASE 60 SNKRX 94342993 \n best $3 I've ever spent.\n\nCASE 61 SNKRX 94607532 \n perfect arcade game, just needs to be less vague with the descriptions.\n\nCASE 62 SNKRX 94693334 \n for such a simple game, SNKRX is addicting and has so much potential to grow!\n\nCASE 63 SNKRX 94818551 \n Great game! 9/10 a lot more replay value than you'd expect, and enough of a challenge to enjoy it.\n\nCASE 64 SNKRX 94897708 \n many yes\n\nCASE 65 SNKRX 95069015 \n The best way to say it. I judge games based on how much I payed for it. And let me promise you. I got way more than $3 worth of enjoyment out of it. I am confident you will love this game. If not. its 3 dollars...\n\nCASE 66 SNKRX 95124208 \n SNKRX has quite a bit more depth and tactics than it has any right to have. Which is to say, it breaks the mold in all the best ways. And by all accounts, the developer is a great dude. Never before have I enjoyed reading a developer blog before. Made me want to learn how to code and even checkout some Skillshare vids.\n\nCASE 67 SNKRX 95131276 \n Better than the most of AAA's on store.\n\nCASE 68 SNKRX 95141865 \n cool but if u beat the game once wtf is the point in playing more\n\nCASE 69 SNKRX 95176244 \n Pure action-arcade fun, with customizable builds.\n\nCASE 70 SNKRX 95438012 \n funny ball game\n\nCASE 71 SNKRX 95615529 \n For it's price it's an amazing game. It's easy enough to learn the basics of the game strategically and mechanically, but it gets difficult once you start getting into the different levels of new game +. There is enough diversity in units and items that you will want to play the game a decent amount to try out different synergies/builds.\n\nTL;DR: Easily worth it for the price.\n\nCASE 72 SNKRX 95745264 \n snek\n\nCASE 73 SNKRX 96146994 \n SNKRX is a good game, but far from great. I feel like with some more updates it could really shine.\n\nthe game is good, but it has major problems with tutorials, proper balancing, UI, and information hidden from the player.\n\nfirst of all, this game needs a tutorial. there is a guide in-game, but this only covers so little. when i first played the game, I was lost and had no idea what I was doing.\n\nbalancing is also a big issue. some heroes and items are ridicilously overpowered or underpowered, not even subjectivly, but statistically. after every few rounds, you can choose one of 4 items wich will give a passive bonus to your heroes. they should all be just as good then and be a relatively hard choice, but one item, for example, gives your 1st and last hero a damage buff. however, another item gives ALL your heroes a damage buff. and its not like its a higher tier item: they can appear on the same selection screen. besides: some of lower tier heroes are the exact same, and some of the lower tier heroes are objectivly better than higher tier heroes: the infestor is infinitely better than the corruptor, for example.\n\nspeaking of wich: the heroes and enemies need some proper character. currently all the heroes are just differently colored blobs: sometimes the same color, wich makes it hard to know wich hero died if you didnt pay attention. \n\none boss in particular is also just insanely unfair.\n\nsomething i really need in the game is an encyclopedia. you cant really plan ahead and create a strategy before you are going in, so you just have to guess sometimes. an encyclopedia with useful information about both heroes and enemies would be extremely helpful.\n\nits a good game: but it could be a lot better.\n\n\n\nCASE 74 SNKRX 96158471 \n very good roguelike\n"}

**Comparing review counts**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"fe1d12","wall_time_seconds":0.000002431,"exit_code":0,"original_token_count":608,"output":"\nCASE 75 SNKRX 96282286 \n I like never write reviews for games. This one i am its fire, this proves it doesn't matter what a game looks like as long at it has a  good gameplay behind it its  fun. 10/10\n\nCASE 76 SNKRX 97890466 \n good game make good fun\n\nCASE 77 SNKRX 98181948 \n fun game, love it.\n\nCASE 78 SNKRX 103417665 \n Jelly Beans\n\nCASE 79 SNKRX 103572759 \n Not that great at the game but pretty fun\n\nCASE 80 SNKRX 103601317 \n Its good\n\nCASE 81 SNKRX 104502554 \n One of the best chill roguelikes out right now. game play can get stale after a while, hopefully more content will be added soon\n\nCASE 82 SNKRX 108220737 \n gud gaym\n\nCASE 83 SNKRX 109002665 \n Didn't take very long for my first victory, but it does have NG+ which makes it harder and gives you more character slots.  For $3 there's a lot of replayability.\n\nCASE 84 SNKRX 109019358 \n Super addicting. Definitely recommend.\n\nCASE 85 SNKRX 110446735 \n Awesome game with a great soundtrack. I see myself playing a lot more in the future.\n\nCASE 86 SNKRX 111219015 \n :)\n\nCASE 87 SNKRX 112125623 \n Got to be one of the best Rouglites I've ever had the experience of Playing! Such a simple art style and premise, but yet so in depth and fleshed out. 10/10 W+God+Lot'sofMaidens\n\nCASE 88 SNKRX 113057180 \n Magic how this game can be as ♥♥♥♥♥♥♥♥ as pvp autochess when the ai aint even real man ♥♥♥♥ these tictacs\n\nCASE 89 SNKRX 117340212 \n fun and cheap rogue-LIKE\nand it's even on mobile\n\nCASE 90 SNKRX 118849702 \n swdefrtgyuhijokp\n\nCASE 91 SNKRX 118895785 \n addictive and simple\n\nCASE 92 SNKRX 122399971 \n I love this game, its really fun.\n\nCASE 93 SNKRX 137806840 \n super fun and addicting, but the orange ones are the worst\n\n\nCASE 94 SNKRX 157560678 \n fun and easy to learn\n\n\nCASE 95 SNKRX 160423431 \n The dev said it best: just go get Brotato\n\nCASE 96 SNKRX 168168149 \n This roguellike set your military squares in a train against the pill shaped enemy. Turn left or right as your troops run forward to avoid and position you troops for the offensive which is automatic.\n\nCASE 97 SNKRX 174688385 \n it's fun\n\nCASE 98 SNKRX 182388452 \n An amazingly addictive game, great for when you want to pass some time; build possibilities are extensive, NG+ opens up for new possibilities and the more you explore the more you'll be wowed by the options. Considering the price, this is a MUST BUY if you like roguelikes!\n\nCASE 99 SNKRX 216450779 \n yes bro yes\n"}

**Comparing random review lengths**

**Comparing random review lengths**

**Checking age-matched counts**

**Comparing random review lengths**

**Checking age-matched counts**

**Comparing random review lengths**

**Checking age-matched counts**

**Comparing random review lengths**

**Checking age-matched counts**

**Comparing matched-age counts**

**Comparing random review lengths**

**Checking age-matched counts**

**Comparing matched-age counts**

**Comparing random review lengths**

**Checking age-matched counts**

**Comparing matched-age counts**

Script completed
Wall time 0.8 seconds
Output:

{}
{"chunk_id":"5aafba","wall_time_seconds":0.261142985,"exit_code":0,"original_token_count":39,"output":" appid     game  n  median_words  B  E  P  F  R  U\n760330 BYTEPATH 50          35.5 27  4 18 20 12  0\n915310    SNKRX 50           9.0  6  1  6  2  8  2\n"}
{"chunk_id":"47102b","wall_time_seconds":0.244694515,"exit_code":0,"original_token_count":3258,"output":"\n Eliza 190599746 False \n I like it when developers go out of their way and do different projects. In that way, I'm glad Eliza was made, but it just didn't land for me. I've played and enjoyed (and often failed at) other Zachtronics games, and I knew what I was getting into when I bought this one on sale. It's not a puzzle game by any stretch (solitaire notwithstanding), it's a pure visual novel.\n\nVisual novels are already not my style of game, but I decided to try going into it with an open mind. Unfortunately, what I found here was... just okay. Nothing I can really recommend spending the time on. The concept is sort of neat, I thought the voice acting was good, but not enough really happens with the story. The game plays with the same themes throughout, and too little seems to develop or change through the story until an ultimate scene where [spoiler] the main character makes a final decision (you quite literally pick from a list of endings), in which the character is suddenly resolute in an epilogue about their choice despite an entire story full of responses like \"maybe\" and \"could be\" and \"possibly\". I get that part of the resolution is the main character finally deciding what to do, but the sudden change is jarring. [/spoiler]\n\nIf you're unsure about the game itself but are really curious to try out the solitaire variant, I recommend instead buying \"The Zachtronics Solitaire Collection\".\n\n Eliza 145812937 True \n I started playing this because I love the Zachtronics puzzles games, and found myself unexpectedly invested in their stories. I don't usually play graphic novels, but I enjoyed this one. I liked the writing, and the story of the protagonist struck a chord with me personally.\n\nThe game explores the automation of mental healthcare alongside the protagonist's personal struggles with disillusionment and figuring out what she wants to do in life. The mental healthcare automation story is an interesting what-if scenario with enough of the tech rooted in reality to make it feel plausible, and the writing of the protagonist's personal struggles felt reassuringly recognizable to me. For a story about big AI taking over something as human as mental healthcare counseling, it had a remarkably down-to-earth feeling, almost slice-of-life, similar to the stories from the Zachtronics puzzle games.\n\nIt's not a ground-breaking piece of literature, but certainly an interesting experience if you're interested in the story's themes, and I'm glad I played it.\n\n Eliza 54860514 True \n \"Eliza\" hits the mark of things I enjoy in a visual novel: charming, thought-provoking, and a stellar presentation. Despite it's linearity, I found the story to be compelling, presenting ideas of technology merging into the field of mental health while also tackling the main character's issues. As a fan of Zachronic games, I'm amazing how they are able to incorporate their common themes of technology and development and keep that intact into this narrative experience. I highly recommend this visual novel if you are at all interested in its modern tech sci-fi themes, or even if you are new in the world of visual novels.\n\nYou can read my full, in-depth review here: https://kupowrites.wordpress.com/2019/08/28/visual-novel-review-eliza/\n\n Eliza 54688217 False \n A visual novel from zachtronics.. I actually quite like the story elements of their games, especially in Opus Magnum. But Eliza is not good.\n\nIt starts off well, with a little sinister twist in the first Eliza session. You think it's going somewhere interesting. But it isn't, it doesn't, not at all.\n\nThe whole story is weighed down by the relentless emotional weight of the dialogue. With every character acting like they bear the weight of the whole world on their shoulders, that everything they do has enormous consequences. But at no point is all of this pompous arrogance actually justified, at no point are any real or meaningful consequences realised. So if you're not already on board with a very particular worldview, the story and its characters just leave you feeling annoyed. Half way through the game i decided to pretend that Eliza was one of those 99p fart apps that were all the rage during the early days of the app store. \n\nThe solitaire game slaps though.\n\n Eliza 54604246 True \n I bought Eliza without any prior knowledge of the game and without any reviews being posted.  I would never do that for most developers (and especially for a VN), but given Zachtronics' record and how I hold their games in such high esteem I didn't deliberate.\n\nEliza has mostly been enjoyable (I've played it through once which took six hours).  High notes that I would draw attention to is the excellent voice acting, nice art, music and initial concept.  However I'd stop short of saying that this was a great game, or even one fully deserving of it's price point.  I am recommending it, but there are some points which I felt it fell short:\n\n1. It (at least felt like it) was highly linear, except for the final chapter.  The decision about [spoiler]what I wanted to do with my career seemed like the only one that actually felt like it would affect the story line (I may be wrong).  But it also gave me choices of which I only really felt that two were viable or made any sense given what I know about Evelyn[/spoiler].\n2. The idea of being a proxy to an AI is excellent but wasn't really fully explored as much as I imagined.  It gave the story on an arc which lead me to [spoiler] make one of the two logical choices at the end, but otherwise just illustrated how everyone has their own problems (I knew that) and how crap and ineffective Eliza is in the first place.[/spoiler]  Which leads me to...\n3. [spoiler] I don't know if it was intentional to make Eliza such a bad service and be barely any more useful than the original from the 70's/80's.  The metrics on the left (positive/negative) would seem to suggest that this was intentional, but if that's true do you really expect me to think that the \"generalised\" AI is going to cut the mustard?  The notion that Evelyn had any part of herself, or any connection with the system fell completely flat because it would just ask the same scripted messages.  I saw Eliza as a system purely made to pedal more apps and finance drug companies.[/spoiler]\n4. I never really got on board with Soran's [spoiler] startup product, womanising or bondage side-story.  That felt a bit weird honestly and didn't add anything remotely to the story.  The whole dream machine concept doesn't really make any sense to me in terms of how it improves anyone's happiness when they're \"awake\" again and that's why I didn't deem working for him a viable end result [/spoiler]\n5. The music was great until it stopped and there was just silence with me reading messages on a static screen.  Perhaps some background environment noise that fades in after the music ends would help.\n6. Similarly the art was great but there wasn't much of it.  I know I'm being highly critical here but even just fading from one expression to another one and back again would add so much more to it, especially during the first consultation.\n\nI know I've gone into way more detail on the negative than positive here which might be unfair.  It's a bold move for Zachtronics and does show how far they've come that they can \"pull off\" a VN, but judged as a game or as entertainment on it's own merit I don't find it as satisfying as their other stuff.\n\n6/10\n\n Eliza 54592158 True \n As a long time Zachtronics fan, Eliza to me just cements what I already felt about the team.... It's not just their puzzle creation ability and superb presentation that makes their games so engaging. It's their choice of underlining subject matter in each game and their ability to delve into how different areas of technology can impact the people around it (Yes even Opus Magnum) that makes each game shine so damn bright.\n\nWith each game they prove that a outstanding game isn't about just taking some popular mechanic that sold well in another game and then just building them around some resurrected generic tropes or playing to current trends. Instead they find a subject they truly want to delve into and then build a game around that, and the difference in the end result as the end user really can be felt.\n\nI don't think there has been a Zachtronics game yet that felt like the game itself wasn't a purely natural extension of the subject matter (Infinifactory came kind of close, but lets forget about that one for now hehe), and Eliza still doesn't disappoint.\n\n BYTEPATH 116097298 True \n I finished the game (once, of course) and went to read some reviews, maybe write another myself. Someone pointed out at the developer. I clicked the name and went into the page, and saw they have two games. I already owned the other one, SNKRX which is FANTASTIC even if it was more expensive. And I loved this one, so I think I'm a big fan of this guy now. \n\nThe game is unique, at least I didn't play anything like it. Sure it has similar ideas here and there, but the visuals and feel of the game was unmatched for me. I really liked the style of this game, and despite being short, it's cheap as hell and even if I never come back to it, It's money well spent. Not that I mind short games, in fact I like them more. They respect our time and the value it has. Of course they have to make for it in the quality, and I'm happy to say that this game does!\n\nBuy it, it's short. If you didn't like it, refund it. You'll probably can even if you finished it once (like me).\n\n BYTEPATH 95034038 True \n Though less popular and less polished than a327ex's latest game SNKRX (which I enjoy a lot), I found lots of fun with the various build options that BYTEPATH has to offer. From ships which send homing projectiles towards basically every rock, ship, and big floating square it sees, to nuclear cores that create several screen-wide explosions every second, to exploding tornadoes full of self-replicating projectiles which, depending on the powerup weapon, can bring the framerate and even the fabric of the game itself to its knees, there are many ways to play BYTEPATH. These descriptions are not hyperbole, and are actual builds I have played with. It's fun, and only like two dollars.\n\n SNKRX 95224877 True \n Another great and addictive game by a developer that obviously knows how to string together several mechanics to make a compelling gaming experience in a small, cheap package;\n\nBytepath - A top-down shooter with several customization options through a skill tree and many ship options;\nSNKRX - A top-down shooter with centipede like controls and auto-chess type units for each segment of the centipede.  There are plenty of builds and ways to play while you mix and match upgradeable units and items.  I easily blew through 4 hours in 1 sitting, well worth the price and then some  This is a developer to keep your eye out for more content!\n\n SNKRX 93920127 True \n Honestly one of the best games I've played so far this year. It immediately eclipsed everything else and I love it. The music, the style. It´s like Autochess, but Arcade Shooter. Fast. Chaotic. Fun. Someone should do this but with Tower Defense. If you do it a327ex, I will definitively buy it.\n\nUpdates are constant but I do have one QoL request: \n\nINFORMATION. I would really like to know how my units are doing and what are they doing. Is one of them useless? Is the whole team useless? I have no way of knowing until I get destroyed and have to start all over again. Maybe you should get lives? 3 lives that cannot be replenished, so that if you die, the run doesn't stop so abruptly and you have a chance to see what is going on. But this would make the game easier, so... maybe more rounds? more bosses? Idk im just a player. I do board games, not video games. \n\nSo what information am I talking about? \n\n- Overall damage/healing output of a unit during a fight \n- Cooldown, how often do they attack?\n- Defense.\n- Some units, like healers, don't say/do anything when leveling from 1 to 2. Did something change? Or is it useless until level 3? I do not know.\n- Others have a rather confusing description (Maybe its just me). Like the lvl. 3 Gambler\n\nOther than that, some units do almost the same things but somehow share no class. Once you get past round 6-7, money becomes a non issue and have no use for it other that rerolling, which is important of course, but by lategame money really doesn't matter. maybe give the option of spending like 30/50/70 gold to add more units to the snake? Cause you get to max units really fast in Regular Game, NG+1 and NG+2 and you can't play around with synergies anymore. \n\nMaybe there could be a store for items that are for a single unit instead of global. This could also bring more depth to the position of the units on the snake. Something like \"adjecent units have +10% attack speed of whatever\", \"if this unit is at the front it gets +10 defense\", stuff like that.\n\nI can't wait to see what comes next. GG a327ex <3\n"}

**Checking sample coding**

**Checking sample coding**

**Checking review dataset**

**Checking sample coding**

**Checking review dataset**

Script completed
Wall time 0.4 seconds
Output:

{"chunk_id":"2a2dc0","wall_time_seconds":0.226118232,"exit_code":0,"original_token_count":2447,"output":"2276830 129 / 279\n\nID 199086507 True\nactical RPG. The gameplay mechanics are addicting and there is plenty of room for customizing your builds. \n\n It is not as big as Horizon's Gate, but it does do a great job if you have finished Horizon's Gate and crave for more. | lenty of room for customizing your builds. \n\n It is not as big as Horizon's Gate, but it does do a great job if you have finished Horizon's Gate and crave for more.\n\nID 177419675 False\n customisation are the bread and butter of tactical RPGs like this, but it seems like the developers took all the good ideas from Horizon's Gate out. The wait system is gone, and you no longer even choose facing (the sprites were only drawn in left and right versions...). Seems like the combat took a back seat to the story, but the story isn't enough to carry the game.\n\nIt also feels like it's missing elements to make it special. Horizon's Gate had a great theme, sailin | ck seat to the story, but the story isn't enough to carry the game.\n\nIt also feels like it's missing elements to make it special. Horizon's Gate had a great theme, sailing, trading, crew management, naval battles, city building, etc., whereas this one simply has story and walking around. \n\nOn a positive note, the UI/UX got some quality of life enhancements, which was very much needed.\n\nID 209710833 True\nIf you're like me and every other TRPG cannot stand up to Final Fantasy Tactic's combat and class systems, Rad Codex's tactics games are for you. They all have a hybrid class system just like FFT, with some changes that I like and there aren't any useless meta classes like the mediator or thief (instead you just get loot off of enemies when combat is over). Each game has a different type of campaign, but they all share the same class and comb |  not incentivized to drag out fights to maximize XP gain.\n\nas for Kingsvein in particular, This one has the most story out of all Rad Codex games. 2 of the characters are set in their personality and looks (you can build them however you like in combat minus 1 special passive for each) and then you have 2 generic mute party members you build, and your main character who has a special ability of your choice.\nThis is the first Rad Codex game that I tried and it was m |  have 2 generic mute party members you build, and your main character who has a special ability of your choice.\nThis is the first Rad Codex game that I tried and it was my most played game in 2024 on all of steam. The first few games art was a little rough so I wasn't sure about the quality of them, but this game's art style sold me and I'm glad to say I was wrong and am going through all his older games now. the exploration is fun and has some zelda/metroid like q\n\nID 207339259 True\nThis is a killer game. Among the Rad Codex tactical RPGs I'm torn between this and Horizon's Gate as the best, but I think as a 'curated' game experience this one takes the cake! It'll play a bit like a cross between a 'western' dungeon crawl and a tactical JRPG, with a soundtrack that really says \"JRPG\" in all the best ways.\r\n\r\nNow the \"Metroidvania\" tag you might see  | This is a killer game. Among the Rad Codex tactical RPGs I'm torn between this and Horizon's Gate as the best, but I think as a 'curated' game experience this one takes the cake! It'll play a bit like a cross between a 'western' dungeon crawl and a tactical JRPG, with a soundtrack that really says \"JRPG\" in all the best ways.\r\n\r\nNow the \"Metroidvania\" tag you might see on this one isn't really accurate. But the game is stil |  do I get over there\" environmental puzzles, just in a way that fits its own dungeon-crawl genre. And like previous titles in the Rad Codex series you often have some flexibility in how you choose to get \"over there.\"\r\n\r\nThe whole experience--tightly balanced combat, great battle setpieces, and rewarding exploration--comes to a head at what is probably one of my favorite final bosses in a game ever, and easily the best final boss encounter in any Rad Codex tactics \n\nID 181724811 True\nKingsvein’s biggest strength, in comparison to Horizon’s Gate, is refinement. The aesthetics are more polished across the board - the soundtrack by Kevin Phetsomphu is fantastic and the more limited colour palette leaves a striking impression. Lots of smaller changes contribute to this visual improvement as well, like gargoyle on the player’s skill bar or the grey border around character’ | on to combine any two classes together. \n\n  Encounter design shines through Kingsvein. The terrain-effecting elements of previous Rad Codex games still play a significant part here, but a return to smaller battlemaps and a glut of holes and environmental features make combat wild and fluid. Seeing a plan form around the specific advantages of the battlemap is a joy that never gets old - and is rarely the same from encounter to encounter. Enemy variety plays a subst\n\nID 195250917 True\nthe dev spent 10 years refining his idea of a tactical game and it shines through here. the scope is tight, tighter than voidspire but much much more engaging, both in terms of mechanics and story. there's just enough of a split between field exploration and combat, and the music bangs the whole time through.\n\ncombat classes and mechanics were refined immensely, though most of it centers around movement - blind restricts movements, spells require some lock\n\nID 170233853 True\nthan some of the others, but it's still very light\n                       - the new sprites and revamped art style is better than horizon's gate imo.\n                       - The difficulty is fair. My advice: use the terrain manipulation abilities, they seem weak at first but when you learn how careful the AI can be, a sea of fire or spike can act as wall. As for ice, it makes push EXTREMELY strong. Pushing enemies with ice or using desecrate is so fun. the ice mage is | rate is so fun. the ice mage is the best class, BY FAR\n\nThe negatives:- I miss some of the classes and elements of other games by Rad Codex. Namely I really liked the status effect meta of Horizon's gate, and i wish we got that ALONGSIDE the new terrain stuff. The toned down effects are a bit lame, especially since I do find the classes decent, except marksman, it's not as good this time. The changes to the combat system are a mixed bags. I liked the environment st |  negatives:- I miss some of the classes and elements of other games by Rad Codex. Namely I really liked the status effect meta of Horizon's gate, and i wish we got that ALONGSIDE the new terrain stuff. The toned down effects are a bit lame, especially since I do find the classes decent, except marksman, it's not as good this time. The changes to the combat system are a mixed bags. I liked the environment stuff, but not the delayed actions being so hard to play around.\n\n\n\nID 232012715 True\nJust beat this and I'm very positive on it. This is another tactical RPG in the Rad Codex oeuvre, and it has the sauce. There have been some adjustments to the mechanics, but fans of the series will find a lot of the same ideas coming back. Open, class based character progression, a wide world to explore, abilities to be used in and out of combat, elements that affect the environment for that immersive sim experienc |  the environment for that immersive sim experience. \n\nThis game's tight scope feels like a response to the high seas adventure in Horizon's Gate. You control a special agent sent to investigate a land no longer responding to communications. You quickly gain allies and some basic equipment, and you're left to poke around. This is a refreshingly alien world that helps the player parse it with a ubiquitous Look function, and the related journal entries and narrative paint  |  party in fun ways, and by end game you really feel like this group of soldiers have evolved into elite specialists. \n\nI find the Rad Codex games to be highly inspiring and I hope the developer is able to comfortably continue to work on more games. I highly recommend this game to fans of tactical RPGs, creative fantasy universes, or simplified immersive sims.\n\nID 170292577 False\nI loved Horizon's Gate. I really wanted to like this game too, but... pushing, bashing, and terrain-blockading just became too important in this installment, which I didn't like. On that note, I will talk about some more of the cons first, since I know that's what many readers are here for.\n\nThe bad:\n\nThe combat options are just too wombo-combo focus | ttack that you had no chance of avoiding, but I'll definitely take that over this. In fact, I do miss the Time-to-Act system from Horizon's Gate in general.\n\nNow, for the good...\n\nI liked the mount system, as others have said. The music was pretty good, and the world and aesthetic felt appealing. Actually, I pretty much liked everything except for the combat and lack of classes.\n\nAll in all, creative expression, roleplaying, and problem-solving options are harshly limit |  classes.\n\nAll in all, creative expression, roleplaying, and problem-solving options are harshly limited in this game compared to Horizon's Gate, and ultimately, that is my reason for the thumbs down.\n\nID 156159442 True\nGreat little game, I think just pipped by Horizon's Gate, but it's also not that much like the previous ones. Love the focus on special characters (Wisp Slayer plus bandolier plus 4 weapons really feels like a badass monster slayer). Delayed attacks are very hard to get working, but then push and pull are king in this game.\n\nGood going, that's 4 Rad Codex games I've played and 4 I've | onster slayer). Delayed attacks are very hard to get working, but then push and pull are king in this game.\n\nGood going, that's 4 Rad Codex games I've played and 4 I've liked!\n1224290 91 / 1048\n1366100 2 / 30\n"}

**Analyzing sampled reviews**

**Analyzing sampled reviews**

**Analyzing sampled reviews**

**Computing portfolio bands**

**Analyzing sampled reviews**

**Computing portfolio bands**

**Analyzing sampled reviews**

**Computing portfolio bands**

**Plotting review-age curves**

**Analyzing sampled reviews**

**Computing portfolio bands**

**Plotting review-age curves**

Script completed
Wall time 0.5 seconds
Output:

{}
{"chunk_id":"cc8254","wall_time_seconds":0.287908547,"exit_code":0,"original_token_count":487,"output":"OWN WINDOWS\n     game  days  reviews\nBYTEPATH     1       24\nBYTEPATH     7       61\nBYTEPATH    14       66\nBYTEPATH    30       71\nBYTEPATH    90       85\nBYTEPATH   365      134\n   SNKRX     1       13\n   SNKRX     7       52\n   SNKRX    14       99\n   SNKRX    30      590\n   SNKRX    90     2298\n   SNKRX   365     3322\nSEQUENCES\n     developer          anchor                    quiet              return_game  anchor_current  quiet_current  return_current  anchor_day30  quiet_day30  return_day30  anchor_day90  quiet_day90  return_day90  current_return_to_anchor  day90_return_to_anchor  anchor_reviews_before_return  quiet_reviews_before_return\n    Rad Codex  Horizon's Gate                   Azalea                Kingsvein            1212             40             303            88           16           141           166           19           165                     0.250                   0.994                           809                           33\nStudio Fizbin    Say No! More              Lost At Sea             Reignbreaker            1757             43             440           174            4           201           235            4           244                     0.250                   1.038                          1226                           24\nTuatara Games   Let Them Come         Bare Butt Boxing Let Them Come: Onslaught            1082             13             216            58           11           152            88           12           178                     0.200                   2.023                          1049                           13\n   Desert Fox Bad Dream: Coma Faded Stories: Full Moon     Bad Dream: Afterlife            1966             41              31            74            6            12            91           10            17                     0.016                   0.187                          1874                           30\nBAND CHANGES 7 of 18\n"}

**Verifying historical estimates**

**Verifying historical estimates**

**Assessing build-taste portability**

**Verifying historical estimates**

**Assessing build-taste portability**

**Verifying historical estimates**

**Assessing build-taste portability**

**Checking catalog completeness**

Script completed
Wall time 1.4 seconds
Output:

{"chunk_id":"1a40bb","wall_time_seconds":1.232355402,"exit_code":0,"original_token_count":165,"output":"  appid       role  ordinal            name_x    clan_id           name_y  reviews  valid_released  main         developer  is_free\n 769970 developers        0        Desert Fox 45172409.0 Bad Dream: Fever     1343            True False        Desert Fox        1\n1673600 developers        0        Desert Fox 45172409.0    Game For Anna      131            True False        Desert Fox        1\n1746370 developers        0        Desert Fox 45172409.0   DUMB: Treasure       41            True False        Desert Fox        1\n4240790 developers        0 Tuatara Tea Party        NaN  Cooking For MA!        0           False False Tuatara Tea Party        0\n"}

**Reviewing Desert Fox titles**

**Re-running collector**

**Re-running collector**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"a7fe14","wall_time_seconds":1.000958753,"session_id":88125,"original_token_count":0,"output":""}

**Checking cached review records**

**Checking cached review records**

**Checking collection timestamps**

**Checking cached review records**

**Checking collection timestamps**

Script completed
Wall time 0.0 seconds
Output:

{}
{"chunk_id":"4614bb","wall_time_seconds":0.000001614,"exit_code":0,"original_token_count":377,"output":"Bad Dream: Coma complete 1966 summary 1967 requests 193\nBad Dream: Fever complete 908 summary 909 requests 204\nDarkness Under My Bed complete 85 summary 85 requests 204\nGame For Anna complete 0 summary 0 requests 205\nDUMB: Treasure complete 0 summary 0 requests 206\nBad Dream: Stories complete 116 summary 116 requests 206\nFaded Stories: Greenberg complete 54 summary 54 requests 206\nBad Dream: Purgatory complete 41 summary 41 requests 206\nFaded Stories: Full Moon complete 41 summary 41 requests 206\nBad Dream: Afterlife complete 31 summary 31 requests 206\nVoidspire Tactics complete 271 summary 271 requests 206\nAlvora Tactics complete 135 summary 135 requests 206\nHorizon's Gate complete 1212 summary 1212 requests 206\nAzalea complete 40 summary 40 requests 206\nKingsvein complete 303 summary 304 requests 206\nThe Inner World complete 920 summary 920 requests 206\nThe Inner World - The Last Wind Monk complete 455 summary 455 requests 206\nSay No! More complete 1757 summary 1757 requests 206\nMinute of Islands complete 714 summary 714 requests 206\nLost At Sea complete 43 summary 43 requests 206\nReignbreaker complete 440 summary 440 requests 206\nLet Them Come complete 1082 summary 1082 requests 206\nBare Butt Boxing complete 13 summary 13 requests 206\nLet Them Come: Onslaught complete 216 summary 216 requests 206\nEliza complete 1007 summary 1007 requests 206\nBYTEPATH complete 306 summary 306 requests 206\nSNKRX complete 4197 summary 4197 requests 206\nCollection closed: 206 requests, 16353 reviews.\n"}

**Re-expanding free-game API data**

**Re-expanding free-game API data**

**Collecting supplementary free-game data**

**Re-expanding free-game API data**

**Collecting supplementary free-game data**

**Re-expanding free-game API data**

**Collecting supplementary free-game data**

**Re-expanding free-game API data**

**Collecting supplementary free-game data**

**Splitting paid/free histories**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"ecfca0","wall_time_seconds":1.001065646,"session_id":25850,"original_token_count":0,"output":""}

**Fixing history analysis**

**Fixing history analysis**

**Checking combined review counts**

Script completed
Wall time 0.0 seconds
Output:

{}
{"chunk_id":"ed0481","wall_time_seconds":0.000001181,"exit_code":0,"original_token_count":49,"output":"Bad Dream: Fever 1343 all-acquisition reviews; shared requests 221\nGame For Anna 131 all-acquisition reviews; shared requests 224\nDUMB: Treasure 41 all-acquisition reviews; shared requests 226\n"}

**Reviewing validation totals**

**Reviewing validation totals**

**Reviewing validation totals**

**Reviewing validation totals**

**Building analysis script**

Script completed
Wall time 0.9 seconds
Output:

{"chunk_id":"ae62f4","wall_time_seconds":0.686483501,"exit_code":0,"original_token_count":2057,"output":"Warning: truncated output (original token count: 2057)\nTotal output lines: 49\n\nHISTORY\n                                 name    n  english  summary_total  earliest_lag_days  clock_usable  count_by_day30  count_by_day90  count_by_day365\n                     Bad Dream: Coma 1966      565           1967               0.08          True            74.0            91.0            182.0\n                    Bad Dream: Fever  908      318            909               0.09          True            99.0           123.0            183.0\n               Darkness Under My Bed   85       46             85               0.16          True            10.0            13.0             40.0\n                       Game For Anna    0        0              0                NaN         False             NaN             NaN              NaN\n                      DUMB: Treasure    0        0              0                NaN         False             NaN             NaN              NaN\n                  Bad Dream: Stories  116       60            116               0.29          True            12.0            22.0             36.0\n            Faded Stories: Greenberg   54       31             54               0.11          True            12.0            17.0             29.0\n                Bad Dream: Purgatory   41       24             41               0.13          True            11.0            12.0             20.0\n            Faded Stories: Full Moon   41       25             41               0.93          True             6.0            10.0             19.0\n                Bad Dream: Afterlife   31       18             31               0.83          True            12.0            17.0              NaN\n                   Voidspire Tactics  271      244            271               0.10          True            13.0            21.0             36.0\n                      Alvora Tactics  135      123            135               3.19          True             8.0            13.0             24.0\n                      Horizon's Gate 1212     1048           1212  …1057 tokens truncated…                           448                       1618            19.0\n    Rad Codex                            Kingsvein                               Azalea                          33                    40                             1179                       1658           165.0\nStudio Fizbin The Inner World - The Last Wind Monk                      The Inner World                         199                   920                              199                        920            34.0\nStudio Fizbin                         Say No! More The Inner World - The Last Wind Monk                         174                   455                              585                       1375           235.0\nStudio Fizbin                    Minute of Islands                         Say No! More                         217                  1757                              820                       3132           158.0\nStudio Fizbin                          Lost At Sea                    Minute of Islands                          98                   714                              952                       3846             4.0\nStudio Fizbin                         Reignbreaker                          Lost At Sea                          24                    43                             3014                       3889           244.0\nTuatara Games                     Bare Butt Boxing                        Let Them Come                         866                  1082                              866                       1082            12.0\nTuatara Games             Let Them Come: Onslaught                     Bare Butt Boxing                          13                    13                             1062                       1095           178.0\n       a327ex                                SNKRX                             BYTEPATH                         219                   306                              219                        306          2298.0\n"}
{}

**Patching free-status fields**

**Patching free-status fields**

**Checking collection timestamps**

**Patching free-status fields**

**Checking collection timestamps**

Script completed
Wall time 0.9 seconds
Output:

{}
{"chunk_id":"b39c32","wall_time_seconds":0.740751594,"exit_code":0,"original_token_count":2057,"output":"Warning: truncated output (original token count: 2057)\nTotal output lines: 49\n\nHISTORY\n                                 name    n  english  summary_total  earliest_lag_days  clock_usable  count_by_day30  count_by_day90  count_by_day365\n                     Bad Dream: Coma 1966      565           1967               0.08          True            74.0            91.0            182.0\n                    Bad Dream: Fever  908      318            909               0.09          True            99.0           123.0            183.0\n               Darkness Under My Bed   85       46             85               0.16          True            10.0            13.0             40.0\n                       Game For Anna    0        0              0                NaN         False             NaN             NaN              NaN\n                      DUMB: Treasure    0        0              0                NaN         False             NaN             NaN              NaN\n                  Bad Dream: Stories  116       60            116               0.29          True     …1557 tokens truncated…Minute of Islands                          98                   714                              952                       3846             4.0\nStudio Fizbin                         Reignbreaker                          Lost At Sea                          24                    43                             3014                       3889           244.0\nTuatara Games                     Bare Butt Boxing                        Let Them Come                         866                  1082                              866                       1082            12.0\nTuatara Games             Let Them Come: Onslaught                     Bare Butt Boxing                          13                    13                             1062                       1095           178.0\n       a327ex                                SNKRX                             BYTEPATH                         219                   306                              219                        306          2298.0\n"}

**Preparing evidence snippets**

Script completed
Wall time 2.0 seconds
Output:

{"chunk_id":"40509e","wall_time_seconds":0.209519229,"exit_code":0,"original_token_count":39,"output":" appid     game  n  median_words  B  E  P  F  R  U\n760330 BYTEPATH 50          35.5 27  4 18 20 12  0\n915310    SNKRX 50           9.0  6  1  6  2  8  2\n"}
{"chunk_id":"e632d2","wall_time_seconds":0.295159254,"exit_code":0,"original_token_count":487,"output":"OWN WINDOWS\n     game  days  reviews\nBYTEPATH     1       24\nBYTEPATH     7       61\nBYTEPATH    14       66\nBYTEPATH    30       71\nBYTEPATH    90       85\nBYTEPATH   365      134\n   SNKRX     1       13\n   SNKRX     7       52\n   SNKRX    14       99\n   SNKRX    30      590\n   SNKRX    90     2298\n   SNKRX   365     3322\nSEQUENCES\n     developer          anchor                    quiet              return_game  anchor_current  quiet_current  return_current  anchor_day30  quiet_day30  return_day30  anchor_day90  quiet_day90  return_day90  current_return_to_anchor  day90_return_to_anchor  anchor_reviews_before_return  quiet_reviews_before_return\n    Rad Codex  Horizon's Gate                   Azalea                Kingsvein            1212             40             303            88           16           141           166           19           165                     0.250                   0.994                           809                           33\nStudio Fizbin    Say No! More              Lost At Sea             Reignbreaker            1757             43             440           174            4           201           235            4           244                     0.250                   1.038                          1226                           24\nTuatara Games   Let Them Come         Bare Butt Boxing Let Them Come: Onslaught            1082             13             216            58           11           152            88           12           178                     0.200                   2.023                          1049                           13\n   Desert Fox Bad Dream: Coma Faded Stories: Full Moon     Bad Dream: Afterlife            1966             41              31            74            6            12            91           10            17                     0.016                   0.187                          1874                           30\nBAND CHANGES 7 of 18\n"}
{"chunk_id":"9f2519","wall_time_seconds":1.001129448,"session_id":84360,"original_token_count":0,"output":""}

**Checking validation alignment**

**Checking validation alignment**

**Checking collection validation**

**Checking validation alignment**

**Checking collection validation**

**Checking validation alignment**

**Checking collection validation**

Script completed
Wall time 0.6 seconds
Output:

{}
{"chunk_id":"610b6c","wall_time_seconds":0.383759768,"exit_code":1,"original_token_count":127,"output":"Traceback (most recent call last):\n  File \"/home/adn/a327ex/steam-research/career-history-2026-09-06/validate.py\", line 27, in <module>\n    check('request_pacing',len(at)<2 or np.diff(at.asi8).min()/1e9>=1.9)\n    ~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n  File \"/home/adn/a327ex/steam-research/career-history-2026-09-06/validate.py\", line 11, in check\n    def check(k,b):checks[k]=bool(b);assert b,k\n                                            ^\nAssertionError: request_pacing\n"}
{"chunk_id":"87b4a0","wall_time_seconds":0.000001309,"exit_code":0,"original_token_count":7,"output":"Saved two PNG/SVG figures.\n"}

[
  {
    "type": "text",
    "text": "Script completed\nWall time 1.1 seconds\nOutput:\n"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"ab7f76\",\"wall_time_seconds\":0.18358241,\"exit_code\":0,\"original_token_count\":8,\"output\":\"datetime64[us, UTC] 2.000023\\n\"}"
  },
  {
    "type": "text",
    "text": "{}"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"86cb0d\",\"wall_time_seconds\":0.551934259,\"exit_code\":0,\"original_token_count\":120,\"output\":\"{\\n  \\\"passed\\\": 446,\\n  \\\"review_collection_requests\\\": 226,\\n  \\\"feasibility_requests\\\": 10,\\n  \\\"unique_reviews\\\": 16960,\\n  \\\"primary_paid_reviews\\\": 15445,\\n  \\\"free_context_reviews\\\": 1515,\\n  \\\"main_summary_differences\\\": [\\n    {\\n      \\\"appid\\\": 538070,\\n      \\\"collected\\\": 1966,\\n      \\\"summary_total\\\": 1967\\n    },\\n    {\\n      \\\"appid\\\": 769970,\\n      \\\"collected\\\": 908,\\n      \\\"summary_total\\\": 909\\n    },\\n    {\\n      \\\"appid\\\": 2276830,\\n      \\\"collected\\\": 303,\\n      \\\"summary_total\\\": 304\\n    }\\n  ]\\n}\\n\"}"
  },
  {
    "type": "image",
    "note": "[image: shown below]"
  },
  {
    "type": "image",
    "note": "[image: shown below]"
  }
]

**Adjusting chart y-limits**

Script completed
Wall time 1.6 seconds
Output:

{}
{"chunk_id":"3747eb","wall_time_seconds":1.001108343,"session_id":50623,"original_token_count":0,"output":""}
{"chunk_id":"ca1193","wall_time_seconds":0.237763851,"exit_code":0,"original_token_count":807,"output":"                    name  n  positive  english\nFaded Stories: Full Moon 41        33       25\n                  Azalea 40        39       30\n             Lost At Sea 43        20       24\n        Bare Butt Boxing 13         9        9\n                    name      previous_name  previous_reviews_at_launch  previous_reviews_now\nFaded Stories: Greenberg Bad Dream: Stories                          29                   116\n          Alvora Tactics  Voidspire Tactics                          58                   271\n          Horizon's Gate     Alvora Tactics                          53                   135\n                  Azalea     Horizon's Gate                         231                  1212\n       Minute of Islands       Say No! More                         217                  1757\n             Lost At Sea  Minute of Islands                          98                   714\n        Bare Butt Boxing      Let Them Come                         866                  1082\n                    game  selected_paid_reviews_before  additional_currently_free_game_reviews_before                              currently_free_games                                                                                                                                                                    note\n   Darkness Under My Bed                          1116                                            527                                  Bad Dream: Fever Paid-game channel uses Steam purchases; free-game supplement uses all acquisition types. These are selected review records, not buyers or a comparable paid-only total.\n      Bad Dream: Stories                          1298                                            691 Bad Dream: Fever | Game For Anna | DUMB: Treasure Paid-game channel uses Steam purchases; free-game supplement uses all acquisition types. These are selected review records, not buyers or a comparable paid-only total.\nFaded Stories: Greenberg                          1520                                            831 Bad Dream: Fever | Game For Anna | DUMB: Treasure Paid-game channel uses Steam purchases; free-game supplement uses all acquisition types. These are selected review records, not buyers or a comparable paid-only total.\n    Bad Dream: Purgatory                          1736                                            954 Bad Dream: Fever | Game For Anna | DUMB: Treasure Paid-game channel uses Steam purchases; free-game supplement uses all acquisition types. These are selected review records, not buyers or a comparable paid-only total.\nFaded Stories: Full Moon                          1817                                            999 Bad Dream: Fever | Game For Anna | DUMB: Treasure Paid-game channel uses Steam purchases; free-game supplement uses all acquisition types. These are selected review records, not buyers or a comparable paid-only total.\n    Bad Dream: Afterlife                          2157                                           1389 Bad Dream: Fever | Game For Anna | DUMB: Treasure Paid-game channel uses Steam purchases; free-game supplement uses all acquisition types. These are selected review records, not buyers or a comparable paid-only total.\n"}

**Comparing review patterns**

**Comparing review patterns**

**Comparing review patterns**

**Appending source citations**

[
  {
    "type": "text",
    "text": "Script completed\nWall time 0.9 seconds\nOutput:\n"
  },
  {
    "type": "text",
    "text": "{}"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"b2845d\",\"wall_time_seconds\":0.654158775,\"exit_code\":0,\"original_token_count\":120,\"output\":\"{\\n  \\\"passed\\\": 446,\\n  \\\"review_collection_requests\\\": 226,\\n  \\\"feasibility_requests\\\": 10,\\n  \\\"unique_reviews\\\": 16960,\\n  \\\"primary_paid_reviews\\\": 15445,\\n  \\\"free_context_reviews\\\": 1515,\\n  \\\"main_summary_differences\\\": [\\n    {\\n      \\\"appid\\\": 538070,\\n      \\\"collected\\\": 1966,\\n      \\\"summary_total\\\": 1967\\n    },\\n    {\\n      \\\"appid\\\": 769970,\\n      \\\"collected\\\": 908,\\n      \\\"summary_total\\\": 909\\n    },\\n    {\\n      \\\"appid\\\": 2276830,\\n      \\\"collected\\\": 303,\\n      \\\"summary_total\\\": 304\\n    }\\n  ]\\n}\\n\"}"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"0e958b\",\"wall_time_seconds\":0.000001676,\"exit_code\":0,\"original_token_count\":7,\"output\":\"Saved two PNG/SVG figures.\\n\"}"
  },
  {
    "type": "image",
    "note": "[image: shown below]"
  }
]

**Reviewing captured source counts**

**Reviewing captured source counts**

**Reviewing captured source counts**

**Correcting tag-distance analysis**

**Reviewing captured source counts**

**Correcting tag-distance analysis**

Script completed
Wall time 0.4 seconds
Output:

{"chunk_id":"0266fd","wall_time_seconds":0.249287102,"exit_code":0,"original_token_count":7,"output":"915310 3620 7\n760330 259 6\n"}

**Checking game-count wording**

Script completed
Wall time 0.0 seconds
Output:

{}

**Checking existing classifier misses**

**Checking existing classifier misses**

**Defining host-genre strata**

**Checking existing classifier misses**

**Defining host-genre strata**

**Condensing research options**

[sealed content withheld: memory]

[sealed content withheld: memory]

**Choosing genre-host tags**

**Choosing genre-host tags**

**Refining build criteria**

**Choosing genre-host tags**

**Refining build criteria**

**Choosing genre-host tags**

**Refining build criteria**

**Defining strict build criteria**

**Choosing genre-host tags**

**Refining build criteria**

**Defining strict build criteria**

**Choosing genre-host tags**

**Refining build criteria**

**Defining strict build criteria**

**Grouping genre hosts**

**Choosing genre-host tags**

**Refining build criteria**

**Defining strict build criteria**

**Grouping genre hosts**

**Choosing genre-host tags**

**Refining build criteria**

**Defining strict build criteria**

**Grouping genre hosts**

**Freezing eligible cohorts**

**Choosing genre-host tags**

**Refining build criteria**

**Defining strict build criteria**

**Grouping genre hosts**

**Freezing eligible cohorts**

**Freezing audit sample**

**Choosing genre-host tags**

**Refining build criteria**

**Defining strict build criteria**

**Grouping genre hosts**

**Freezing eligible cohorts**

**Freezing audit sample**

**Choosing genre-host tags**

**Refining build criteria**

**Defining strict build criteria**

**Grouping genre hosts**

**Freezing eligible cohorts**

**Freezing audit sample**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"1d4917","wall_time_seconds":1.000971395,"session_id":14572,"original_token_count":0,"output":""}

**Planning sample review**

**Planning sample review**

**Auditing sampled match labels**

Script completed
Wall time 5.1 seconds
Output:

{"chunk_id":"e35358","wall_time_seconds":5.000741821,"session_id":14572,"original_token_count":22,"output":"Classified 10000\nClassified 20000\nClassified 30000\nClassified 40000\nClassified 50000\n"}
{"chunk_id":"270d56","wall_time_seconds":0.000002383,"exit_code":0,"original_token_count":40,"output":"total 12K\ndrwxr-xr-x 1 adn adn   44 Sep  6 11:53 __pycache__\n-rw-r--r-- 1 adn adn 6.9K Sep  6 11:53 design.py\n-rw-r--r-- 1 adn adn 3.0K Sep  6 11:53 prepare.py\n"}

**Preparing audit-ready analysis**

**Preparing audit-ready analysis**

**Defining control fallbacks**

**Preparing audit-ready analysis**

**Defining control fallbacks**

**Preparing developer-weighted analysis**

**Preparing audit-ready analysis**

**Defining control fallbacks**

**Preparing developer-weighted analysis**

**Preparing audit-ready analysis**

**Defining control fallbacks**

**Preparing developer-weighted analysis**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"29689b","wall_time_seconds":0.000001598,"exit_code":0,"original_token_count":988,"output":"Classified 60000\nClassified 70000\n/home/adn/a327ex/steam-research/build-destinations-2026-09-06/prepare.py:10: PerformanceWarning: DataFrame is highly fragmented.  This is usually the result of calling `frame.insert` many times, which has poor performance.  Consider joining all columns at once using pd.concat(axis=1) instead. To get a de-fragmented frame, use `newframe = frame.copy()`\n  for label,tags in HOSTS.items():g[f'host{rank}_{label}']=g.appid.isin(t.loc[t.tag.isin(tags)&t['rank'].le(rank),'appid'])\n/home/adn/a327ex/steam-research/build-destinations-2026-09-06/prepare.py:10: PerformanceWarning: DataFrame is highly fragmented.  This is usually the result of calling `frame.insert` many times, which has poor performance.  Consider joining all columns at once using pd.concat(axis=1) instead. To get a de-fragmented frame, use `newframe = frame.copy()`\n  for label,tags in HOSTS.items():g[f'host{rank}_{label}']=g.appid.isin(t.loc[t.tag.isin(tags)&t['rank'].le(rank),'appid'])\n/home/adn/a327ex/steam-research/build-destinations-2026-09-06/prepare.py:10: PerformanceWarning: DataFrame is highly fragmented.  This is usually the result of calling `frame.insert` many times, which has poor performance.  Consider joining all columns at once using pd.concat(axis=1) instead. To get a de-fragmented frame, use `newframe = frame.copy()`\n  for label,tags in HOSTS.items():g[f'host{rank}_{label}']=g.appid.isin(t.loc[t.tag.isin(tags)&t['rank'].le(rank),'appid'])\n/home/adn/a327ex/steam-research/build-destinations-2026-09-06/prepare.py:10: PerformanceWarning: DataFrame is highly fragmented.  This is usually the result of calling `frame.insert` many times, which has poor performance.  Consider joining all columns at once using pd.concat(axis=1) instead. To get a de-fragmented frame, use `newframe = frame.copy()`\n  for label,tags in HOSTS.items():g[f'host{rank}_{label}']=g.appid.isin(t.loc[t.tag.isin(tags)&t['rank'].le(rank),'appid'])\n/home/adn/a327ex/steam-research/build-destinations-2026-09-06/prepare.py:10: PerformanceWarning: DataFrame is highly fragmented.  This is usually the result of calling `frame.insert` many times, which has poor performance.  Consider joining all columns at once using pd.concat(axis=1) instead. To get a de-fragmented frame, use `newframe = frame.copy()`\n  for label,tags in HOSTS.items():g[f'host{rank}_{label}']=g.appid.isin(t.loc[t.tag.isin(tags)&t['rank'].le(rank),'appid'])\n/home/adn/a327ex/steam-research/build-destinations-2026-09-06/prepare.py:10: PerformanceWarning: DataFrame is highly fragmented.  This is usually the result of calling `frame.insert` many times, which has poor performance.  Consider joining all columns at once using pd.concat(axis=1) instead. To get a de-fragmented frame, use `newframe = frame.copy()`\n  for label,tags in HOSTS.items():g[f'host{rank}_{label}']=g.appid.isin(t.loc[t.tag.isin(tags)&t['rank'].le(rank),'appid'])\n/home/adn/a327ex/steam-research/build-destinations-2026-09-06/prepare.py:11: PerformanceWarning: DataFrame is highly fragmented.  This is usually the result of calling `frame.insert` many times, which has poor performance.  Consider joining all columns at once using pd.concat(axis=1) instead. To get a de-fragmented frame, use `newframe = frame.copy()`\n  for label,tags in [('core',CORE),('rogue',ROGUE)]:g[label]=g.appid.isin(t.loc[t.tag.isin(tags)&t['rank'].le(10),'appid'])\n/home/adn/a327ex/steam-research/build-destinations-2026-09-06/prepare.py:11: PerformanceWarning: DataFrame is highly fragmented.  This is usually the result of calling `frame.insert` many times, which has poor performance.  Consider joining all columns at once using pd.concat(axis=1) instead. To get a de-fragmented frame, use `newframe = frame.copy()`\n  for label,tags in [('core',CORE),('rogue',ROGUE)]:g[label]=g.appid.isin(t.loc[t.tag.isin(tags)&t['rank'].le(10),'appid'])\nPrepared 75860 games; 96 unique audit cases.\nStrict 4946 extended 7144 additional strict 758\n"}
{"chunk_id":"64fc1f","wall_time_seconds":0.000002272,"exit_code":0,"original_token_count":3150,"output":"\nCASE 0 4168960 Bad Goys HOST Platforming strict\nBad Goys is a fast-paced political parody where you play greedy elite caricatures racing to collect coins before time runs out. Featuring 75 levels, 3 maps, trinkets to collect, star rankings, character customisation, and no forced ads.\nEVIDENCE: {\"synergy\": [\"sed level scoring - master each level like in Angry Birds 8 hilarious character customizations to mix and match No Ads Fast-paced arcade action that rewards greed and speed 👃 Coming Soon Local & online multiplayer - out-greed your friends New characte\"]}\n\nCASE 1 4539000 SoulSwap HOST Platforming strict\nDefeat enemies and take over their bodies to fight on. A dark fantasy 2D action roguelite where you transform into a new being with every monster you slay.\nEVIDENCE: {\"loadout\": [\"nto that creature and continue the fight. Each monster carries unique abilities and combat styles — the body you choose completely changes your strategy and playstyle Ever-Changing Combat and Strategy Hold on with a weaker body, or take the risk and switch to a more powerful foe? Every battle demands cons\"], \"parts\": [\" and absorb their souls. The moment you slay a monster, you instantly swap into that creature and continue the fight. Each monster carries unique abilities and combat styles — the body you choose completely changes your strategy and playstyle Ever-Changing Combat and Strategy Hold on with a wea\"]}\n\nCASE 2 2431080 Blood Dungeon HOST Platforming strict\nFrom the creators of Nidhogg, murder hordes of monsters in this 2D platforming autoshooter with a killer soundtrack. Collect Blood, gold, and bones to increase your powers. Can you survive the Blood Dungeon?\nEVIDENCE: {\"build\": [\"t them, and harvest their blood. Increase your blood level to upgrade your build, or find chests hidden throughout the arenas. Circulate around the various drop sites to supplement your arsenal and master each arena. Vis\"]}\n\nCASE 3 3006670 BomberHit™ HOST Platforming extended_only\nBomberHit™ is an action-packed multiplayer game where you defeat your opponents by detonating bombs. Think strategically, develop tactics, and conquer your adversaries in labyrinthine battles! Engage in combat with friends or online players and fight to become the ultimate bomber!\nEVIDENCE: {\"parts\": [\"perience! Set up rooms, matchmaking settings, map selection, and item adjustments to suit your preferences. 3. Strategic Abilities : Gain more bomb abilities, control your speed, and utilize other strategic skills during the game. But remember, your opponents will have the same abilities! 4. Char\"]}\n\nCASE 4 3396900 DevTycoon HOST Platforming extended_only\nDevTycoon is a software development business simulation with a charming pixel art style, which challenges you to build and manage your own company in a world where competition is fierce and the challenges reflect real life.\nEVIDENCE: {\"parts\": [\"re, you can create your own website or service platform. Building and Customizing: Design and build the platform from scratch, choosing servers, modules, and key features that will appeal to users. Functional Modules: Adding functional modules to your website can enhance its appeal. Some mod\"]}\n\nCASE 5 2356350 Grand War: Rome HOST Platforming extended_only\nA turn-based strategy war chess game.\r\nYou will play a commander in the Roman period, and command the generals and troops to complete the specified mission objectives and occupy important strongholds in the level.\r\nRome, Samnite, Epirus, Carthage. Different choices will open different histories!\nEVIDENCE: {}\n\nCASE 6 3114440 War of Charge HOST Platforming no_signal\nBuy soldiers, assemble troops, and simply command them to fight. Each branch has its own strengths, rationally match your lineup, break through the other side's defense line, defeat all the other soldiers, or go to the other side's base camp. Support multiplayer online,and lobby matching\nFULL: · A casual online strategy game where you buy soldiers, assemble troop, and simply command them to fight. Together with your teammates, let your soldiers defeat the enemy. Each branch has its own strengths, rationally match your lineup, break through the other side's defense line, defeat all the other soldiers, or go to the other side's base camp. · There are three arms in the game, respectively, melee, ranged, cavalry, each has its own strengths, such as cavalry mass, can break through the line of defense, forward charge. Remote units can cause a lot of output in the rear. There are many types of melee troops, and there may be a strategy for every situation. · This is a lightweight game, characterized as casual and easy to use, with almost no learning costs, all you have to do is buy soldiers and click, move or attack. The game is divided into multiple rounds, round interval to obtain a certain gold, you can buy soldiers, according to the opponent's lineup with their own lineup, the victory of the first to reach a certain number of victories. · Support local games or multiplayer online.And lobby matching is supported.\n\nCASE 7 2617190 Soulshard HOST Platforming no_signal\nIn a realm between life and death, a confused soul, haunted by a life filled with misguided choices, seeks redemption and freedom.\nFULL: In a realm between life and death, a confused soul, haunted by a life filled with misguided choices, seeks redemption and freedom. In Soulshard, you command a soul in search of an exit from a dark and desolate realm, using the environment to your advantage and avoiding treacherous traps alone or in cooperative mode! Single-Player! Local CO-OP! (Offline) 30 levels with more than 6 mechanics! Minimalistic visual in pixel art.\n\nCASE 8 3251690 Anti Terrorist Shooting Game HOST First-person shooting strict\nJoin elite forces in this fast-paced anti-terrorist FPS. Engage in tactical gunfights, master a variety of weapons, and outsmart enemy squads in immersive multiplayer battles. Your skills, reflexes, and strategy will be crucial for survival and victory!\nEVIDENCE: {\"loadout\": [\" and skill. Weapons & Arsenal: Choose from a wide selection of modern weapons, including assault rifles, sniper rifles, shotguns, grenades, and more. Upgrade your loadout to better withstand enemy attacks and dominate the battlefield. Gameplay Features: Auto-Shooting: Focus on strategy while the game handles \"]}\n\nCASE 9 4109600 Requisition VR: Hunt & Extract HOST First-person shooting strict\nExperience an intense multiplayer PvPvE extraction shooter set in a vast post-apocalyptic world. Scavenge, duct-tape random objects into deadly weapons, fight both players and AI enemies, extract valuable loot, upgrade your base and return stronger each raid.\nEVIDENCE: {\"loadout\": [\"igh-stakes mission, so get ready to extract your loot or lose it all. Upgrade your hideout, enhance your weapons and build the ultimate loadout to dominate the wasteland. The challenge awaits! Innovative crafting mechanics Combine almost any items you come across in the environment \"]}\n\nCASE 10 570950 Devilated HOST First-person shooting strict\nDevilated is an ultraviolent RPG old-school shooter where you brutalize and cannibalize your demonic foes in equal measure. Eat their face before they eat yours.\nEVIDENCE: {\"loadout\": [\"age. Headbang along to our incredible Metal soundtrack. Equip various weapons and form your own playstyle. Save your game with checkpoints throughout each level. Find secret easter eggs that give a glimpse into the madman behind Devilated. Multi\"]}\n\nCASE 11 3286930 Heretic + Hexen HOST First-person shooting extended_only\nStep into the definitive re-release of two dark spell-casting shooters that shaped the genre. This enhanced restoration makes Heretic™ + Hexen™ available to new audiences and long-time fans with expanded accessibility across more platforms and languages than ever before.\nEVIDENCE: {\"parts\": [\"casting shooters that shaped the genre. Forged by Raven Software on a modified DOOM engine, Heretic and Hexen were released in 1994 and 1995, pushing first person shooters into bold new territory. These innovative cult classics in\"]}\n\nCASE 12 3729780 Doom Drive HOST First-person shooting extended_only\n**Doom Drive – No Rules. No Mercy**: Drive cars, buses & motorcycles to crush mutant and city zombies. Use powerful weapons and customize your hero as you survive the ultimate apocalypse!\nEVIDENCE: {}\n\nCASE 13 3224950 Emoji Puzzle Action - Codename: ManKun HOST First-person shooting extended_only\nIn a future transformed by tech, social media and emojis dominate communication. But a disaster strikes when the \"Emoji Virus\" turns emojis into aggressive beings, overrunning the world. You, \"ManKun\", the sole survivor and ex-security expert, must now save humanity.\nEVIDENCE: {}\n\nCASE 14 2597570 Chasmal Fear HOST First-person shooting no_signal\nChasmal Fear is a next-gen FPS action horror game built in Unreal Engine 5 from the perspective of a body camera. Manage limited resources, make smart choices that affect playthrough, and survive monsters that mutate to get stronger. Played as singleplayer or Co-Op with 2 players.\nFULL: ABOUT THIS GAME Chasmal Fear is a first-person tactical shooter survival horror game that puts players in a mysteriously empty sci-fi underwater facility where something has gone terribly wrong. Ammo is scarce. The odds are stacked against you. Choose wisely between fighting or sneaking past the monstrosities which once killed might mutate into something even worse. Unlock hidden areas and solve puzzles to uncover the truth of what happened in this top-secret facility. Every decision will impact your gameplay experience, some swarming relentless monsters against you while others offer the chance to employ stealth and cunning to survive. Wishlist Chasmal Fear now to stay up to date with the latest news! KEY FEATURES: LIMITED RESOURCES Limited player ammo and scarce ammo boxes will encourage players to think strategically and be cautious while traversing the abandoned steel hallways. IMMERSIVE BODYCAM Artistic direction that utilizes the perspective of a bodycam footage. Set in the near future, players are also equipped with an AR Interface to help you through the dark hallways of the Magnus Facility. MONSTER MUTATION SYSTEM Some monsters will randomly start to mutate when killed into a stronger form. Do not be surprised when you encounter more dangerous foes in an area you thought you cleaned out. Make sure to end them before the mutation starts. CHOICES AFFECT YOUR PLAYTHROUGH Your choices will affect your gameplay experience. Choosing to open one way will close other paths to your next goal. Depending on your choice, you can dive guns blazing or sneak stealthily above a swarm of enemies; hide in the shadows or use your environment to squash the monstrosities. FULLY RANDOMIZED EVENTS Randomized events occurring throughout the game provide replayability from varying sounds, crumbling ceilings, floors, and unexpected scares. LORE & INTERACTION Find out clues, pieces of lore, emails, audio logs, and video logs, and find out how Magnus Underwater City came to be, what happened in the meantime, and learn more about the dangers you're facing. REALISTIC ENVIRONMENTS Graphically realistic, sci-fi underwater station built underneath the waves and inside a massive ocean cliff. Its hallways seem abandoned. It is up to you to find your way out and survive the things that lurk within. CO-OP POSSIBILITY Bring a friend to Chasmal Fear and survive this nightmare together. The difficulty changes procedurally based on whether you play singleplayer or multiplayer co-op.\n\nCASE 15 3332660 SICKO SANCTUM 2 HOST First-person shooting no_signal\nEnemies are weapons. Style is power. Your only goal, total violence. SICKO SANCTUM 2 is an all new synergistic score-attack FPS made for veterans of the movement-shooter subgenre.\nFULL: is an innovative score-attack movement-shooter crafted from the design lessons of modern classics such as ULTRAKILL and HYPER DEMON . Featuring a simple player move-set in combination with synergistic enemy design, SICKO SANCTUM 2 represents the cutting-edge of digitized violence. ENEMIES ARE WEAPONS Powerful special moves can be done through interacting with enemies in specific ways. Sometimes the strongest attacks are not your own. STYLE IS POWER Stylish play is rewarded with level ups that improve speed and damage. Creative moment to moment strategy is essential to survival. !?!?!? Get to the core of a mysterious and surreal post-apocalyptic world. What's at the bottom and beyond? Only through violence may answers be found.\n"}

**Auditing primary classification rules**

**Auditing primary classification rules**

**Auditing build criteria**

**Auditing primary classification rules**

**Auditing build criteria**

**Auditing primary classification rules**

**Auditing build criteria**

**Refining genre-tag criteria**

**Auditing primary classification rules**

**Auditing build criteria**

**Refining genre-tag criteria**

Script completed
Wall time 0.7 seconds
Output:

{"chunk_id":"d49316","wall_time_seconds":0.527105882,"exit_code":0,"original_token_count":516,"output":"3396900 [{'rank': 1, 'tag': 'Choose Your Own Adventure'}, {'rank': 2, 'tag': '2D Platformer'}, {'rank': 3, 'tag': 'Immersive Sim'}, {'rank': 4, 'tag': 'Sandbox'}, {'rank': 5, 'tag': 'Building'}, {'rank': 6, 'tag': 'Management'}, {'rank': 7, 'tag': 'Game Development'}, {'rank': 8, 'tag': 'Strategy'}, {'rank': 9, 'tag': 'Life Sim'}, {'rank': 10, 'tag': 'Character Customization'}, {'rank': 11, 'tag': '2D'}, {'rank': 12, 'tag': 'Cute'}, {'rank': 13, 'tag': 'Top-Down'}, {'rank': 14, 'tag': 'Simulation'}, {'rank': 15, 'tag': 'Choices Matter'}, {'rank': 16, 'tag': 'Adventure'}, {'rank': 17, 'tag': 'Pixel Graphics'}, {'rank': 18, 'tag': 'Singleplayer'}, {'rank': 19, 'tag': 'Indie'}]\n2356350 [{'rank': 1, 'tag': '4X'}, {'rank': 2, 'tag': 'Turn-Based Tactics'}, {'rank': 3, 'tag': 'Wargame'}, {'rank': 4, 'tag': '2D Platformer'}, {'rank': 5, 'tag': 'Medieval'}, {'rank': 6, 'tag': 'Rome'}, {'rank': 7, 'tag': 'Military'}, {'rank': 8, 'tag': 'Building'}, {'rank': 9, 'tag': 'War'}, {'rank': 10, 'tag': 'Turn-Based Strategy'}, {'rank': 11, 'tag': 'Board Game'}, {'rank': 12, 'tag': 'Strategy RPG'}, {'rank': 13, 'tag': 'Strategy'}, {'rank': 14, 'tag': 'Grand Strategy'}, {'rank': 15, 'tag': '2D'}, {'rank': 16, 'tag': 'Top-Down'}, {'rank': 17, 'tag': 'Third Person'}, {'rank': 18, 'tag': 'Historical'}, {'rank': 19, 'tag': 'Turn-Based Combat'}, {'rank': 20, 'tag': 'Singleplayer'}]\n3114440 [{'rank': 1, 'tag': 'Casual'}, {'rank': 2, 'tag': 'Strategy'}, {'rank': 3, 'tag': 'Multiplayer'}, {'rank': 4, 'tag': 'Medieval'}, {'rank': 5, 'tag': 'PvP'}, {'rank': 6, 'tag': 'RTS'}, {'rank': 7, 'tag': '2D Platformer'}, {'rank': 8, 'tag': 'Incremental'}, {'rank': 9, 'tag': 'Co-op'}, {'rank': 10, 'tag': 'Idler'}, {'rank': 11, 'tag': 'Real Time Tactics'}, {'rank': 12, 'tag': '3D'}, {'rank': 13, 'tag': 'MOBA'}, {'rank': 14, 'tag': 'Simulation'}, {'rank': 15, 'tag': 'Turn-Based Combat'}, {'rank': 16, 'tag': 'Turn-Based Tactics'}, {'rank': 17, 'tag': 'Combat'}, {'rank': 18, 'tag': 'Singleplayer'}, {'rank': 19, 'tag': '3D Platformer'}, {'rank': 20, 'tag': 'Online Co-Op'}]\n"}
{"chunk_id":"4c5255","wall_time_seconds":0.000002366,"exit_code":0,"original_token_count":3546,"output":"\nCASE 16 4982590 Train Plus Japan Drive Western Japan Trains! Rapid Service Edition HOST Racing strict\nDrive real JR-West trains! An authentic railway driving game — Approved by West Japan Railway Company. Three fictional lines, every section drivable in both directions. A series-first scoring system and card-deck strategy turn precise driving into high scores. Over 40 hours of content.\nEVIDENCE: {\"deck\": [\"the same station for a station score bonus, and upgrade your cards to build the ultimate deck. New cards are obtained through an in-game gacha — there are no paid purchases. Struggling with the controls? A clever deck can still carry\"]}\n\nCASE 17 3646970 The last throne HOST Racing strict\nRoll the dice to shape your fate.  Each throw triggers events as you explore and leap across tiles.  Battle with cards, build your deck, and manage scarce resources.  Time is limited—plan wisely and defeat the Demon King before darkness consumes all.\nEVIDENCE: {\"deck\": [\"gers events as you explore and leap across tiles.  Battle with cards, build your deck, and manage scarce resources.  Time is limited—plan wisely and defeat the Demon King before darkness consumes all. This game world is drive\"]}\n\nCASE 18 2541900 BattleDrive.io HOST Racing strict\nSurvive, Upgrade, Conquer. Race to victory in a relentless arena battle. Drift, unlock new cars, and defeat unique bosses!\nEVIDENCE: {\"loadout\": [\" the chance to buy new cars and customize your vehicles with powerful upgrades. Tailor your ride to your playstyle and dominate the competition with your unique, souped-up machine. Master the Art Of Drift Experience the thrill of precision drifting like \"]}\n\nCASE 19 4300530 Simu V3 HOST Racing extended_only\nStep into the most brutal touring cars in Argentine motorsport and experience what it truly means to race on the edge. Massive power, unforgiving tracks, and zero room for mistakes. Every corner demands precision. Every braking zone punishes hesitation.\nEVIDENCE: {\"parts\": [\" deliver an aggressive and immersive soundscape. Detailed Car Setup & Tuning Adjust suspension, gearing, aerodynamics, and balance like a real race engineer. Challenging AI Opponents Competitive rivals that punish mistakes and force c\"]}\n\nCASE 20 3700840 Car Tuning Garage Simulator HOST Racing extended_only\nBuild your dream garage, customize every detail, and tune high-performance machines. Restore classics, fulfill client orders, and race your creations on streets and tracks. Car Tuning Simulator is your ultimate playground for creativity and car culture.\nEVIDENCE: {\"parts\": [\"pe. 🛠️ Total Control, Infinite Customization Welcome to your personal tuning workshop - where every bolt matters. Dive deep into visual customization: swap out body parts, install widebody kits, spoilers, wheels, lights, and more. Shape the car’s look down to the finest detail. Go under the hood with performa\"]}\n\nCASE 21 3917090 Assetto Corsa Rally HOST Racing extended_only\nAssetto Corsa Rally is a high-fidelity rally sim built for precision and challenge. With 3D laser-scanned stages and cars, professional co-driver support, and dynamic conditions, it delivers a demanding, immersive experience where every second behind the wheel counts.\nEVIDENCE: {\"parts\": [\"cifically refined for the needs of this motorsports discipline, and a customized build of Unreal Engine 5. AC Rally introduces a rally-specific chapter to the Assetto Corsa franchise, designed to deliver a demanding and immersive experience wh\"]}\n\nCASE 22 2512080 DDI Rally Championship HOST Racing no_signal\nStrap in and get ready to experience the thrill of high-speed racing, as you navigate through dynamic terrains, including rugged mountains, dense forests, and treacherous deserts.\nFULL: Inspired by the racing games found in real arcades in the 90s like SEGA Rally Championship and SEGA Rally 2, DDI RALLY CHAMPIONSHIP aims to recreate a game that you can pick up and play without feeling like it's a job. No career modes or complicated mechanics, just choose a course, choose a car and drive. Fully Compatible With The STEAM DECK Current Features: Full Controller Support Force feedback Wheel Support Controller Rumble Controller Remapper 5 Selectable Vehicles 5 Selectable Courses Playground Island Dynamic Time of Day Dynamic Weather Championship Mode Time Trial Mode Practice Mode CRT Screen Effect Championship Mode Leaderboards Time Trial Mode Leaderboards Upcoming Planned Features: Fastest Lap Leaderboards Replays Ghost Vehicles Attract Mode Mirrored Courses Reversed Courses Mirrored & Reversed Courses Another 11 Unlockable Vehicles/Variations DDI Rally Championship is an electrifying arcade video game that puts you behind the wheel of powerful rally cars, ready to conquer the most challenging off-road tracks across the globe. Strap in and get ready to experience the thrill of high-speed racing, as you navigate through dynamic terrains, including rugged mountains, dense forests, and treacherous deserts. Whether you're a seasoned rally racer or a rookie looking for excitement, DDI Rally Championship guarantees heart-pounding arcade racing action, stunning visuals, and an authentic rally racing atmosphere that will keep you coming back for more. It's time to rev your engines, conquer the world's most demanding tracks, and claim your place in rally racing history!\n\nCASE 23 2278120 Break Everything - Park HOST Racing no_signal\nThis game is made on Unreal Engine 5.\r\nYou need to use the toy rifle to shoot the balls to destroy objects.\nFULL: This game is made on Unreal Engine 5. You need to use the toy rifle to shoot the balls to destroy objects. In the game, you control the player on the WASD keys. and aim with the mouse, when you aim at the object, shoot, if you hit the right point, the object will be destroyed, you need to destroy objects as quickly as possible. Objects have displaced centers of impact, so you need to understand exactly which point of the object to shoot at. Break objects the fastest! Be the first on the Steam leaderboard! The game has achievements.\n\nCASE 24 3629050 Punch to Glory HOST Sports strict\nStart your journey as a boxer from the ground up, fighting your way to glory. Develop your character, collect fight cards, and face the challenges life throws your way. Train hard, manage stress, and give everything you've got to turn your championship dreams into reality. \nEVIDENCE: {\"deck\": [\"ng attention to the strengths and weaknesses of your enemies, you can build a deck of cards and use different strategies against each of them. The fighting moves you need to use will be different for each opponent. This bo\"]}\n\nCASE 25 3621330 Pro Jank Footy HOST Sports strict\nPro Jank Footy is an over-the-top arcade comedy-sports game. Score goals to unleash game-changing power-ups, build a deck of absurd abilities, and tackle a season where no two matches are ever the same—solo or with your mates.\nEVIDENCE: {\"deck\": [\"e comedy-sports game. Score goals to unleash game-changing power-ups, build a deck of absurd abilities, and tackle a season where no two matches are ever the same—solo or with your mates. AUSTRALIAN RULES FOOTBALL LIKE YOU\"]}\n\nCASE 26 4638590 Goaldeck HOST Sports strict\nFootball meets roguelike deck-building. Be the coach: draft legends, stack Manager Cards, and manage player egos with Reigns-style choices. Place your squad across 3 zones, open packs, upgrade cards, and conquer all 8 cups against escalating AI. Lose and start over — win and lift the trophy.\nEVIDENCE: {\"build\": [\"ats, unlock traits, or turn a Bronze benchwarmer into a match-winner. Build synergies between your upgrades and your Managers to craft truly broken squads. 🧠 THINK TACTICALLY Place players across three field zones — Attack, M\"], \"deck\": [\"Football meets roguelike deck-building. Be the coach: draft legends, stack Manager Cards, and manage player egos with Reigns-style choices. Place your squad across 3 zones, open \"], \"synergy\": [\"idual player cards with upgrades that permanently raise stats, unlock traits, or turn a Bronze benchwarmer into a match-winner. Build synergies between your upgrades and your Managers to craft truly broken squads. 🧠 THINK TACTICALLY Place players across three field zones — Attack, M\"]}\n\nCASE 27 2115130 Parkour Legends HOST Sports extended_only\nComplete time-trial courses in this fast-paced side-scrolling parkour platformer. Gain speed boosts and build momentum as you vault over obstacles, slide under objects, run on walls, and rebound over impossible gaps. Collect hard-to-reach gear. Style your runner. Climb leaderboards. Become a legend!\nEVIDENCE: {}\n\nCASE 28 3939990 Truck Titans HOST Sports extended_only\nTRUCK TITANS is a high-octane monster truck racing game with 9+ levels of intense action. Customize your truck, pick sponsors, and compete in score attacks and races across city arenas, deserts, and night tracks. Nail stunts, crush obstacles, and become a monster truck legend.\nEVIDENCE: {\"parts\": [\" handling, and raw power. Design your driver, pick your sponsors, and modify your rig with upgraded engines, wild paint jobs and more. Every second counts. Nail jumps, crush obstacles, and chain together combos to rack up high scores and unlock el\"]}\n\nCASE 29 2588230 Real Drift Multiplayer 2 HOST Sports extended_only\nReal Drift Multiplayer 2 offers a unique experience waiting for drift enthusiasts! Experience a real racing simulator with customizable vehicles, competitive online modes, and realistic drift dynamics.\nEVIDENCE: {\"parts\": [\"ltiplayer 2 is for you. Creating the vehicle of your dreams with many customization options, from car color to wheels, suspension, and spoiler, is possible. Remember, a personalized tool can help you score higher in-game. In the single mode of the game, your goal is to\"]}\n\nCASE 30 3774170 Retro Basketball Manager HOST Sports no_signal\nTake control of a basketball team. Choose tactics, sign players and help them develop. Guide your team to victory!\nFULL: CUSTOMIZE THE WORLD Use the editor to create or edit players, coaches and teams. You can also use downloaded databases! MAKE CAREER MOVES Start with a team, national team or unemployed and look for offers from other teams! CHANGE YOUR TACTICS Choose your lineup, playbook and defensive instructions. Watch or sim games to see if they work! MANAGE YOUR ROSTER Build your roster by trading or signing players. DEVELOP PLAYERS Help your players develop by assigning specific drills.\n\nCASE 31 3104610 MEGARACER QUANTUMRUSH HOST Sports no_signal\nIn the futuristic city of Mega City, thrilling races await. Customize unique vehicles and characters to reflect your personal style and join the race. Compete with friends and players from around the world in fun yet fierce contests, all in pursuit of victory. Available on both PC and VR platforms.\nFULL: MegaRacer QuantumRush - The Ultimate Open World Racing Experience MegaRacer QuantumRush is an immersive open-world kart racing game that sets a new standard for racing experiences. Set in the sprawling futuristic city of Mega City , players can enjoy a variety of game modes while driving unique karts and controlling animated, cartoon-style characters. Designed for players of all ages, the game offers rich content that satisfies both beginners and seasoned racers. Diverse Game Modes MegaRacer QuantumRush delivers more than just traditional racing. Experience endless fun with a wide range of modes, including. Street Racing : Speed through the city streets of Mega City in high-intensity races that test your reflexes and precision. Obstacle Racing : Overcome complex barriers and traps in this strategic racing mode where careful planning leads to victory. Crown Battle : Capture the crown within the time limit and hold onto it until the end to win. Soccer Mode : An inventive mode that fuses racing with soccer—use your kart to drive the ball and score goals with your team. Other Single and Multiplayer Modes : Explore Mega City to discover over 10 unique game modes and special events awaiting your challenge. Explore Mega City and discover over ten unique modes and special events waiting to be challenged. Immersion in VR with the Comfort of PC Playable on both PC and VR platforms, MegaRacer QuantumRush is designed for comfortable, extended play sessions. Advanced Visual Processing : Our specially designed VR camera and gameplay systems minimize motion sickness, ensuring a smooth and enjoyable experience. Flexible Controls : Choose between immersive steering wheel-based VR control or relaxed gameplay using joysticks or keyboards. Dual Input Support : Whether you prefer keyboard, mouse, gamepad, or VR controllers, all input types are fully supported and switch seamlessly depending on your setup. Even without VR, the game offers a fully satisfying experience with smooth controls and dynamic visuals on standard PC environments. Kart and Character Customization In MegaRacer QuantumRush , players can freely customize their karts and characters. Kart Customization : Choose your favorite colors, parts, and design elements to create a kart that reflects your personality. Character Styling : Equip adorable characters with various outfits and gear to express your unique flair. A Racing Experience for Everyone MegaRacer QuantumRush is built to be accessible and fun for everyone—from casual beginners to racing veterans. Variety of Races : Compete in lively multiplayer races or enjoy solo mini-games. Explore every corner of Mega City to discover hidden items and secrets. Ongoing Updates with the Community MegaRacer QuantumRush will continue to evolve through close communication with the community. Regular content updates will introduce new tracks, modes, and customization options. Beta tests and surveys will gather feedback to shape a better and more engaging game environment. Start Your Race in the World of Mega City! Chase the dream of becoming the ultimate racer in a world of speed and strategy. Race through the streets of Mega City, conquer unique challenges, and define your own racing style. MegaRacer QuantumRush offers a new level of casual racing fun that anyone can enjoy—regardless of platform.\n"}

**Defining host-text criteria**

**Defining host-text criteria**

**Defining host-text criteria**

**Refining activity-match rules**

**Defining host-text criteria**

**Refining activity-match rules**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"2b76b7","wall_time_seconds":0.000003226,"exit_code":0,"original_token_count":3120,"output":"\nCASE 32 3091340 Metal Thunder HOST Flight/space simulation strict\nCommand the legendary gunship in a high-octane action-roguelite! Engage insurgents with devastating firepower in ever-changing missions. Upgrade your arsenal, adapt to dynamic challenges, and survive relentless waves of enemies. Every run offers new tactics and chaos as you dominate the skies.\nEVIDENCE: {\"loadout\": [\"ed to think on your feet and make split-second decisions. Choose your weapons wisely, customize your loadout, and deploy tactical maneuvers to turn the tide in your favor. No two missions are the same. Encounter a variety of insurgent strategies an\"]}\n\nCASE 33 2094610 Magocracy: Experimentation HOST Flight/space simulation strict\nThe first Flying Third Person Shooter where you can customize everything.\nEVIDENCE: {\"loadout\": [\"ter a world where mages fly and fight in the sky, unleashing powerful spells and devastating the landscape. Customize your mage with dozens of attributes to suit your playstyle and strategy. Whether you prefer speed, size, or sheer firepower, you can fine-tune every aspect of your character and become the ultimate \"], \"parts\": [\"ith 9 Unit Variables and 3 Shot Modes, each with 33 variables to fine-tune your mages’ abilities. Skirmish Mode is a sandbox for your creativity and strategy as the power level has no restraint here. Recreate your favorite aerial superp\"]}\n\nCASE 34 2513490 HELL GALAXY HOST Flight/space simulation strict\nA thrilling singleplayer space shooter that blends weird fiction, dark gritty sci-fi, and transhumanism. Fly your spaceship and fight your way through a galaxy where your mind is at risk just as much as your \"body\": nothing here is human. Not even you. \nEVIDENCE: {\"build\": [\"eaten your sanity as much as your armor. Read each enemy, reconfigure your build, and choose which part of yourself you're willing to risk: the ship that keeps you alive, or the mind that keeps you you . Forget Your Huma\"], \"parts\": [\"keeps you you . Forget Your Humanity Your survival depends on a fully customizable ship. Gather resources and install modules, starting from your cockpit, to reshape yourself into an ever-deadlier raider. Experiment with countlessc ombinations and adapt to the dang\"]}\n\nCASE 35 5068020 KLAST HOST Flight/space simulation extended_only\nKLAST is a fast twin-stick bullet-hell roguelite where every run transforms your ship. Chain kills, rescue survivors, collect powerful upgrades, and conquer three distinct campaigns filled with shifting arenas, relentless swarms, and colossal guardians.\nEVIDENCE: {}\n\nCASE 36 4748330 Luna Meridian HOST Flight/space simulation extended_only\nLuna Meridian is a near-future narrative lunar colony game. A card-based drafted crew runs a 5-phase mission on the Moon's surface. Build mission critical infrastructure, manage resources, respond to hazards, follow and shape their stories and make decisions that determine if the colony survives.\nEVIDENCE: {}\n\nCASE 37 2729440 VERZEUS HOST Flight/space simulation extended_only\nBlast and weave your way through enemy fire while managing a wide variety of ship power-ups in this retro style horizontal shoot-'em-up!\nEVIDENCE: {}\n\nCASE 38 3304990 Starship Destroyer HOST Flight/space simulation no_signal\nA sci-fi space vehicle combat game, where the player must defend their capital ship from being destroyed, while attempting to destroy the alien force's own capital ship!\nFULL: Earth is under attack by a deadly alien force, and you are one of the last remaining starship pilots to defend it. Take to Earth's orbit to beat back the alien force and destroy their capital ship! -Choose from a suite of different armaments for your ship, such as the powerful bombing system, or the rapid fire laser system, to help you combat your enemies. -Collect scrap from destroyed enemy ships and successful missions to upgrade your capital ship's capabilities and features. -Keep an eye out for powerful boosters hidden in asteroids to give you the edge in combat. -Allocate your defense force with fighter, bomber, and defender ships to help you cover the battlefield. -Destroying all the marked 'weak points' on a capital ship will bring victory to you and your forces. You are Earth's last line of defense. Can you save your world?\n\nCASE 39 2766050 Lightspeed HOST Flight/space simulation no_signal\nYou're the only human in a huge, robotic space ship, flying at incredible speeds through deep space. Your mission: find and settle a new planet where humanity can thrive. Combat is inevitable, but you'll need more than flying skills to succeed.\nFULL: ATARI PRESENTS is a series of titles geared toward the preservation of cult classics, forgotten favorites, and titles that would otherwise be lost to time. In the spirit of that, each ATARI PRESENTS title is presented unchanged from its original version other than providing the means to play them on modern hardware. Even the store description below is the very same language used to describe the title at the time of its release. Please enjoy! You're the only human in a huge, robotic space ship, flying at incredible speeds through deep space. Your mission: find and settle a new planet where humanity can thrive. Combat is inevitable, but you'll need more than flying skills to succeed. You'll also need resourcefulness, charm and political skill to help you negotiate with the bizarre, intelligent - sometimes hostile - beings who correctly see you as the alien. 3D Combat Flight Simulation Like You've Never Seen Earth-bound simulations spend most of their graphic power on landmarks and other terrain. But since there's no terrain in the black void of out space, Lightspeed devotes its Super 3D Graphics power to showing space ships in breathtaking detail and animating them with unprecedented smoothness. - Pilot a starship 15,000 feet long - Customize your ship for the performance you need - Control remote fightercraft in strategic battles with intelligent enemies Fascinating Role-Playing Depth Multiple star clusters await your exploration, each with unique challenges. A dozen solar systems spin through each cluster; scores of planets revolve around each sun. Most important, 25 species of aliens require your constant attention. - The Broodsmasters - sinister, manipulative and eager for universal domination - The Lutin - quasi-intelligent, with no memory and no concept of the future - The Cicisbeo - totally self-absorbed, vain and arrogant\n\nCASE 40 4565780 COSMIC WARFRONT HOST Rhythm strict\nTop-down space combat shooter. 6 chassis, 8 weapons, 3 enemy factions, 3 bosses. Craft your build, fight through 12 sectors, survive endless waves, and compete in daily challenges. Global leaderboards.\nEVIDENCE: {\"build\": [\"mbat shooter. 6 chassis, 8 weapons, 3 enemy factions, 3 bosses. Craft your build, fight through 12 sectors, survive endless waves, and compete in daily challenges. Global leaderboards. YOU ARE THE LAST LINE OF DEFENSE Th\"], \"loadout\": [\"one gun. It's a combat machine constructor. Your chassis defines your playstyle. Your weapon defines your rhythm. Your drone covers your weaknesses. Modules and artifacts turn a good build into an unstoppable one. 6 CHASSIS: Interceptor — speed, combo multipliers, time slowdown Assault — balanced DPS, orbital\"]}\n\nCASE 41 2025320 Estencel HOST Rhythm strict\nAtmospheric souls-like. Fight enemies and bosses and explore the gloomy city.\nEVIDENCE: {\"loadout\": [\" to your secondary to overwhelm your opponent. Use a large arsenal of weapons, from a ripping whip to a stunning shotgun, to customize your play style and mow down your enemies with efficiency. About the developer — the game has been developed by one person for more than 3 years. The durat\"]}\n\nCASE 42 3149370 ROGUE LIGHT DECK BUILDER HOST Rhythm strict\nYour house needs a deck. You have a hammer. Freedom is $1,000,000.\nEVIDENCE: {\"deck\": [\"e needs a deck. You have a hammer. Freedom is $1,000,000. ROGUE LIGHT DECK BUILDER is a digital wooden board game. Your house needs a deck. You have a hammer. You have wood and you have nails. The light outside is faulty. \"]}\n\nCASE 43 2766880 Meowingtons Simulator HOST Rhythm extended_only\nThis isn’t your typical dance game...Explore club Oberhasli and Control meowingtons' movements, every dance is an opportunity to create something unique, funny, or downright absurd. Whether you’re bouncing around on the decks, or find a better place to throw down, there’s no wrong way to dance. neat\nEVIDENCE: {}\n\nCASE 44 4453880 Neon Hole HOST Rhythm extended_only\nDive into the infinite, procedurally generated neon void. Dodge obstacles at hypersonic speeds, upgrade your ship via the skill tree, and survive to the rhythm of pulsing electronic music. How deep into the system can you hack?\nEVIDENCE: {\"tree\": [\"void. Dodge obstacles at hypersonic speeds, upgrade your ship via the skill tree, and survive to the rhythm of pulsing electronic music. How deep into the system can you hack? Neon Hole is a dynamic, cyberpunk tunnel run\"]}\n\nCASE 45 3028720 Plaza Security #neverforget7/13 HOST Rhythm extended_only\n#neverforget7/13\r\nThis game is a blend of 4 line rhythm mechanics and tower defense.\r\nYou are a plaza security guard trying to defend waves of looters.\r\nAs you go further you begin to discover some interesting events.\nEVIDENCE: {}\n\nCASE 46 1990940 The Impossible Game 2 HOST Rhythm no_signal\nThe original rhythm platform game is back! Test your skills with all-new levels, online battle royale and more.\nFULL: The platformer that no one could beat is back , this time with online battle royale, a level editor and more. ALL-NEW LEVELS featuring awesome music and crazy new mechanics like guns and portals. Tap to jump, fly and shoot your way through four worlds, each with a unique theme and an intense boss battle. ONLINE BATTLE ROYALE with up to sixty players at once. Will you come out on top? DESIGN YOUR OWN LEVELS with an easy-to-use level editor containing (almost) every mechanic seen in the rest of the game. Infinite possibilities await! AN AWESOME SOUNDTRACK featuring music from Panda Eyes, Nitro Fun, MDK and many more. There are no in-app purchases. All levels and items come included with the game!\n\nCASE 47 3302150 Karaoke Love Song HOST Rhythm no_signal\nShare your best beautiful voice with a karaoke game themed around love songs.\nFULL: karaoke Love Song is a unique karaoke game where you can sing your favorite tracks, create personalized music videos, and share your performances live with the world. With a wide selection of love songs, from timeless classics to modern hits, this game is perfect for music lovers and those who want to express their emotions through singing. Key Features: A diverse catalog of love songs : Sing heartwarming ballads or the latest romantic hits with a constantly growing playlist. Create music videos : Record and customize your performances by adding unique visual effects, then share them directly on your favorite social media platforms or streaming services. Streamer Mode : Easily connect to Twitch or YouTube and captivate your audience with live performances. Dance to the songs : Based on user votes and wishlist requests, discover new dances added to the game as part of our roadmap. Players can directly influence future content updates with our wishlist system. Harder, Better, Faster, Stronger : Push your karaoke skills to the limit with new challenge modes designed to test your singing abilities. Complete more difficult songs, sing with greater precision, and perform under pressure to earn special rewards. Track your progress and get stronger as you go, unlocking faster-paced challenges and better scoring systems. Become a Legend : Stand out from the crowd by turning your best performances into iconic concert experiences . Earn the \"Star of the Concert\" badge by impressing the community, and rise to legend status within the game. Show off your star power and become a true karaoke icon! Play solo or duet : Whether you prefer singing alone or teaming up with a friend for a duet, get ready to create unforgettable memories. Follow the Roadmap The Love Song roadmap is shaped by the players! Depending on the songs added to the wishlist and community feedback, new dance modes, gameplay improvements, and even more songs will be added in each update. You have the power to influence the future of the game! Grab your microphone, set the mood, and become the star of your own romantic music journey—maybe even a legend !\n"}

**Refining deckbuilding criteria**

**Refining deckbuilding criteria**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"f580fb","wall_time_seconds":0.000002388,"exit_code":0,"original_token_count":2827,"output":"\nCASE 48 4895520 Backrooms Manager - Together HOST Stealth/heist strict\nA 1–6 player co-op survival logistics game where a routine delivery traps your crew inside the Backrooms. Scavenge, recycle, deliver cargo, manage a shared budget, and survive 49 days to earn your freedom.\nEVIDENCE: {\"loadout\": [\" in ever-shifting, claustrophobic liminal environments. Progression & Loadout System — Purchase high-tier gear to boost your survival odds on subsequent shifts. You Don't Respawn. Your Debt Does. Backrooms Manager - Together is a tactical co-op survi\"]}\n\nCASE 49 2784740 Shinobi HOST Stealth/heist strict\nA social stealth party game for 2–4 players. Blend in with AI ninjas, find the real threats, and strike before they strike you.\nEVIDENCE: {\"loadout\": [\" up with friends to outwit opponents. Casual & Economic Modes — Fixed loadouts or earn credits to buy gear, adding strategic depth. Multiple Game Modes – Keep gameplay fresh with varied and unpredictable rounds. Customizable Loadouts & Rules – Sh\"]}\n\nCASE 50 2009620 Harvest Hunt HOST Stealth/heist strict\nYou are the Warden. Enter the dense cornfields and gather the resources your village desperately needs. But beware: you are not alone. Build a deck of abilities and choose your weapons to face the growing dangers of the night. With a relentless monster after you, will you make it out alive?\nEVIDENCE: {\"deck\": [\"ources your village desperately needs. But beware: you are not alone. Build a deck of abilities and choose your weapons to face the growing dangers of the night. With a relentless monster after you, will you make it out al\"], \"loadout\": [\" weapons and abilities for your Warden and village and build your own deck to suit your needs. Will you sacrifice your own health to equip them? Choose your playstyle and survive. An aristocrat far removed from wealth and ease, a priest left armed with his sacred verses; what led them here, and why do the\"]}\n\nCASE 51 3111070 Seeker of the Ruins HOST Stealth/heist extended_only\nAs a fearless dwarf delver, explore ancient ruins, uncover hidden treasures, and solve riddles. Use items and potions to outsmart monsters & environmental hazards in this stealthy non-combat adventure. Venture into the deeps in search of luck and fortune. Will you succeed?\nEVIDENCE: {}\n\nCASE 52 4391420 Spymaster HOST Stealth/heist extended_only\nIn a world where privacy and trust are a luxury most can’t afford, the agents of NODE are the last line of defence. Combining action-packed sequences and puzzle-solving in immersive missions, command three elite agents with unique skills and manipulate time using the top secret C.A.S.S.E.T.T.E.  \r\n\nEVIDENCE: {}\n\nCASE 53 4296720 Survivors 27 HOST Stealth/heist extended_only\nScavenge-Fight-Extract, PvE/PvPvE/PvP, Enter the danger zone, explore the map, scavenge for supplies and gear, battle enemies, and successfully extract. Every mission is fraught with the unknown—will you escape safely with your loot, or lose everything?\nEVIDENCE: {\"parts\": [\"supplies, gear, ammunition, medical items, and high-value loot Weapon customization: extensive weapon attachments and performance tuning Modification location : Optic, Muzzle, Barrel, Grip, Stock, Magazine, Tactical Players create rooms to host; support\"]}\n\nCASE 54 2909560 SoulShake HOST Stealth/heist no_signal\nAfter waking up, you will find a way to escape from an abandoned hospital in the basement and figure out what happened.\nFULL: Purchase Notice This is a horror game with no saved single player experience. You will walk through a terrifying abandoned hospital from a first person perspective, exploring its past and why you appeared in this abandoned hospital. There are not many scary points in the game settings, so you can consider it as a walking simulator for an abandoned hospital. story After waking up, you were locked in the basement of a eerie abandoned hospital. You need to find a way to leave the hospital and understand why you appeared in the hospital. Game Time 1 Hour to 1.5 Hours Decrypt The game decryption content is very simple. Follow the prompts to find the corresponding door key in order to get closer to the truth\n\nCASE 55 3476240 Berghotel Heist HOST Stealth/heist no_signal\nIn this (up to) two-player stealth horror game, you are trapped in an haunted Austrian hotel inhabited by a Percht and other creatures from alpine folklore. You must work together, solve puzzles and avoid detection to escape the hotel alive.\nFULL: After the mysterious disappearance of a hotel manager, two thieves attempt what should be an easy heist in an Alpine ski hotel. But the plan unravels as they discover they are not alone. Many objects in the hotel are interactable and can be picked up, inspected, or combined in creative ways. The game relies on stealth and tension rather than combat or gore. It can be played in two-player co-op or single-player. What begins as a simple heist quickly turns into a nightmare as you discover the hotel is haunted by a Percht and other creatures from Alpine folklore. Uncover the dark secret of this forsaken place by exploring its haunted halls, collecting clues, and piecing together the truth about what was really going on in this so called ski hotel. The Percht reacts dynamically to light, sound, and movement, can open doors and actively searches rooms, and will pursue you if detected. You cannot fight back and must rely on hiding, observation and teamwork. A CCTV surveillance system helps you track his movements throughout the hotel. Survival depends on cooperation and wit as you solve intricate puzzles that guard the hotel's darkest secrets. Coordinate with your partner to unlock sealed areas, complete arcane rituals, and gather the items needed for your escape. Only by working together can you unleash the Alpine magic and make it out of the Alps alive.\n\nCASE 56 3552780 Y2Kill HOST Horror strict\nA top-down arcade zombie roguelite set in 99's New York City. Build broken combos, unlock a million-dollar arsenal, slay four monster bosses, and survive the party at the end of the world. \nEVIDENCE: {\"build\": [\"cash and XP, and level up MID-RUN by drafting upgrade cards that turn your build from office temp to walking apocalypse. When the rage meter fills, unleash a blast that deletes the entire screen. No loading bar. No mercy\"], \"deck\": [\"s into monster combos, vacuum up cash and XP, and level up MID-RUN by drafting upgrade cards that turn your build from office temp to walking apocalypse. When the rage meter fills, unleash a blast that deletes the entire screen. No \"]}\n\nCASE 57 3140620 C.A.S.T. (CHERNOZEM ANOMALY STRIKE TEAM) HOST Horror strict\nC.A.S.T is a third-person PvE extraction shooter playable solo or in co-op. Select your character and loadout to battle enemy hordes, evade anomalies, and complete objectives across handcrafted maps with dynamic procedural elements.\nEVIDENCE: {\"loadout\": [\"-person PvE extraction shooter playable solo or in co-op. Select your character and loadout to battle enemy hordes, evade anomalies, and complete objectives across handcrafted maps with dynamic procedural elements. Welcome a new me\"]}\n\nCASE 58 3031070 Mother Bunker VR HOST Horror strict\nThe zombie apocalypse is here, and you want to survive (obviously). Or at least take out as many undead as possible before you succumb. So grab your weapon, look for supplies, build your bunker, and always aim for the head.\nEVIDENCE: {\"loadout\": [\"as. Unique and Rare Weapons: New unique and rare weapons with special abilities and attributes, offering players powerful new tools for combat and survival. Customizable Loadouts: Players can customize their loadouts with a wider variety of weapons and gear, tailored to their preferred playstyle and strategic needs. \"], \"parts\": [\"ristic weaponry, each with distinct combat styles and effects. Weapon Customization: Expanded customization options for weapons, allowing players to modify appearance, stats, and abilities through crafting and upgrades. Unique Weapon Attributes: Introduction of weapons with unique attributes and abilities, such as elemental da\"]}\n\nCASE 59 3312430 Second Odium HOST Horror extended_only\nSeek purpose while managing a plethora of various afflictions and personality traits in hopes to survive an endless barrage of nightmarish threats that continuously seek to consume your mind.\nEVIDENCE: {\"parts\": [\"40+ lootable caves gather EXP to level up Collect items and equipment equip armor, weapons and equipment plethora of strange items and abilities in depth player stats and customization Over 200 Hand drawn enemy monsters that originated on pen and paper\"]}\n\nCASE 60 3306470 Into Death and Beyond HOST Horror extended_only\nInto Death and Beyond is a gripping first-person survival horror game set on the hostile planet XB32-ENT. As bounty hunter Mark Helkins, scavenge for resources, solve puzzles, and confront terrifying mutants. Unravel dark secrets and capture a rogue alien before it hunts you down. Will you survive?\nEVIDENCE: {}\n\nCASE 61 3834600 House of Empty Names HOST Horror extended_only\nYou wake up on a cold floor in an unknown house, missing a kidney, and realize you’ve been kidnapped. Your one goal is to escape at any cost.\r\n\r\nChoose where to go, where to hide, and how to build your path to freedom. Every decision brings you one step closer to escape.\nEVIDENCE: {}\n\nCASE 62 3748640 Wasteland Bites HOST Horror no_signal\nWasteland Bites is a horror-based food truck sim. Cooking up questionable cuisine in your post-apocalyptic food truck is tough. Fending off mutants, UFOs, rabid dogs, and things lurking in the dark? That's the real challenge. A chaotic mix of cooking and survival in a long gone world.\nFULL: \"HEY, YOUR TRUCK STILL OPEN?\" Run a food truck in the post-apocalypse , serving bizarre meals to even stranger customers. Can you survive the high-pressure horrors of both the apocalyptic wasteland and customer service all at once? Mess up too many orders, serve the wrong freaks, or let something slip past you, and stress will pile up fast. Hit your limit, and it’s lights out. \"YOU GOT FOOD? I GOT MOUTH.\" Hungry weirdos are lining up, and it’s your job to feed their freaky appetites . Toast, bowls, and shish kebabs - raw, cooked, rotten or radioactive - you've got to keep up! But sometimes, you'll simply need to feed them a bullet ! It’s all part of the gig. Keep your truck moving , manage your limited resources , survive each stop , and maybe—just maybe—you’ll make it to that legendary non-irradiated beach on the horizon...\n\nCASE 63 3952230 3000m to Whatever’s Up There HOST Horror no_signal\nJump. There must be something up there.\nFULL: You were discarded the moment you drew your first breath. Weak, useless, nothing but a statistical failure. Thrown into an abyss… The end of the line. But you survived. Climb through cursed paths toward the malevolent peak, where, so they say, something incredible awaits. No one has ever made it that far. And yet… you keep jumping. Features: One single action: Jump. Every leap is designed to look easy. It never is. A unique atmosphere: Sounds and visuals crafted to disguise the hell that awaits you. Relentless playtime: From 2 hours to 2 years, depending on the instability of the path… and the stubbornness of the climber.\n"}

**Refining combo matching**

**Refining combo matching**

**Narrowing configuration matches**

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"73b745","wall_time_seconds":0.000002809,"exit_code":0,"original_token_count":3316,"output":"\nCASE 64 4563450 The Sage's Seven HOST Puzzle strict\nA 7-turn logic duel. Analyze and refute villains’ arguments using draft-based card strategy.\nEVIDENCE: {\"deck\": [\"A 7-turn logic duel. Analyze and refute villains’ arguments using draft-based card strategy. Seven turns. Break their logic. ‘The Sage’s Seven’ is a logic-driven card strategy game that tests your reasoning and judgment. V\"]}\n\nCASE 65 3974240 Blackjack Trainer HOST Puzzle strict\nPractice blackjack basic strategy AND card counting in a fully customizable game. Configure any house rule (decks, S17/H17, payouts, splits) and the live strategy table adapts. Train deck counting and live-game counting. One-time purchase. No subscription, no ads, no login.\nEVIDENCE: {\"deck\": [\"Practice blackjack basic strategy AND card counting in a fully customizable game. Configure any house rule (decks, S17/H17, payouts, splits) and the live strategy table adapts. Train deck counting and live-game counting. One-time purchase. No subscripti\"]}\n\nCASE 66 2663790 TaxingTiles HOST Puzzle strict\nShape your realm. Conquer chaos. Unveil mysteries. Deck-building roguelike fun!\nEVIDENCE: {\"deck\": [\"Shape your realm. Conquer chaos. Unveil mysteries. Deck-building roguelike fun! Prepare yourself for an extraordinary gaming experience that melds roguelike strategy, deck building, and nostalgic arcade t\"]}\n\nCASE 67 3764410 Merge Match March HOST Puzzle extended_only\nAn RPG that combines merge puzzles and battles, where you become the strategist of a hero-led army to command your troops and protect the kingdom from monsters!\nEVIDENCE: {}\n\nCASE 68 3139670 H.P Supernatural Event Archives HOST Puzzle extended_only\nThis is a text-based interactive adventure game with a suspenseful horror theme. Through different plot choices, you can collect various items, survive the attacks of indescribable creatures, and finally find the truth.\nEVIDENCE: {}\n\nCASE 69 2605410 Spot the Diff 3D HOST Puzzle extended_only\nThe first ever brand new 3D Spot the difference game!\nEVIDENCE: {\"parts\": [\"ameplay in different ways. They'll find the correct answer for you or adjust the time for you. Level up to improve your abilities.\"]}\n\nCASE 70 2775510 Checkers Twist HOST Puzzle no_signal\nA turn-based strategy game, Checkers Twist is a variant of the checkers board game, where the board is irregular and procedurally generated, offering new tactics each game.\nFULL: A variant of the classic game Checkers (also called Draughts), where the grid is irregular. On a regular grid, each cell has 8 neighbors, but in this case, they can have between 6 and 10. Therefore tiles can have 3 diagonals or 5 diagonals, which offers different moves depending on where a piece is. This leads to more difficult games, as you can't build the same mental tricks to take over your opponent. This also makes your opponent harder to anticipate. Procedurally generated Grids are procedurally generated on each game, and there are hundreds of them. This makes each game unique. Pre-selected boards We've pre-selected some boards that are either similar enough to a regular one, so that can ease into this new variant, or very different, so that you can really enjoy those new strategies. Starting a game When you start a new game, you can choose one of the different modes: - 1 player against a bot - 2 players on the same device - 2 players online Online Multiplayer To play multiplayer online, you can first see if any room created by other users is available. If there are no room, you can create yours: You can decide if your room will be public or private. If you create a public one , any player can join it. If you create a private one , you will see a window with a code you can share (via whatsapp, email, etc) with the person you want to join.\n\nCASE 71 2942670 Ancient Scroll HOST Puzzle no_signal\nAncient Scroll - the game is a mixture of an adventure game and classic quests. You have to go to an abandoned ranch in search of an ancient scroll. Where an exciting adventure awaits you, hiding many mysteries!\nFULL: Ancient Scroll - the game is a mixture of an adventure game and classic quests. You have to go to an abandoned ranch in search of an ancient scroll. Where an exciting adventure awaits you, hiding many mysteries! Explore the abandoned ranch, solve puzzles to advance further in the story! Be very careful and don't miss hints that can help you! Be smart and resourceful to overcome all obstacles and achieve your goal! Are you ready for an exciting adventure? Then hit the road and find the scroll!\n\nCASE 72 3212630 Rogue Factory HOST Management strict\nThe AUTOMATION bullet heaven! Build your custom spellcasting factory and tear through endless waves of monsters. Unlock 20+ characters and 100+ golden totems to craft broken roguelite synergies and survive the swarm! Build, automate, autocast, OBLITERATE!\nEVIDENCE: {\"synergy\": [\"asting factory and tear through endless waves of monsters. Unlock 20+ characters and 100+ golden totems to craft broken roguelite synergies and survive the swarm! Build, automate, autocast, OBLITERATE! Build, Automate, Autocast! Design your factory build from scratch to autocast\"]}\n\nCASE 73 4970940 Save Cent HOST Management strict\nPlace blocks to create rectangular coupons and collect discount stickers to strengthen your deck. Combine a variety of effects to maximize your discounts and get products from the supermarket for free. A puzzle deckbuilding game where every run offers new combinations and strategies.\nEVIDENCE: {\"deck\": [\"ur discounts and get products from the supermarket for free. A puzzle deckbuilding game where every run offers new combinations and strategies. Build the Ultimate Coupon with Discount Stickers Save Cent is a puzzle deckbui\"], \"synergy\": [\"count stickers you choose will shape your strategy. Discover powerful synergies and maximize your discounts. Discover Powerful Synergies Strengthen your deck by collecting new discount stickers at the end of every round, allowing you to create even stronger coupons in the next. Collect discount s\"]}\n\nCASE 74 3629080 Sizzle & Stack HOST Management strict\nSizzle & Stack is a fast-paced restaurant management card game where you stack ingredients, cook dishes, explore recipes, and serve hungry customers. Upgrade your kitchen and master strategic cooking in a unique blend of card play and time management!\nEVIDENCE: {\"deck\": [\"t! Sizzle & Stack is a fast-paced, card-based cooking simulator where deck-building meets restaurant chaos. Inspired by Stacklands , this unique twist on the culinary genre puts you in charge of a growing kitchen where ever\"], \"loadout\": [\"rds. Spend your hard-earned coins on ingredient packs with randomized cards. Expand your setup, unlock new tools, and customize your playstyle. Impress high-profile gourmets to earn prestigious Sizzle Medals —unlocking powerful upgrades, new mechanics, and endgame challenges. Wheth\"]}\n\nCASE 75 2774040 The Boss Gangster: Criminal Empire HOST Management extended_only\nBuild a criminal empire from the streets up. Commit crimes in third-person, run and automate illegal businesses, recruit your own crime family, fight rival organizations, and take control of a seamless open-world city.\nEVIDENCE: {\"parts\": [\" dynamically to your actions. • Vehicle Theft & Customization: Steal, customize, upgrade, race, or dismantle vehicles for profit. • Deep Progression System: Unlock skills, improve your abilities, expand your businesses, and discover new ways to build your empire. • Every System is Connected: Crimes, businesse\"]}\n\nCASE 76 2884190 Fight School Simulator HOST Management extended_only\nAre you ready to become a legendary martial arts master? Build your own school, train your students, and battle rival schools to become the strongest. Join tournaments, raise your grasshoppers, and most importantly, collect tuition from your students!\nEVIDENCE: {\"parts\": [\"ting abilities and make your students even stronger. Design your Dojo Customize your dojo exactly how you want with dozens of training equipment and decoration options. Create a motivating, martial arts-inspired space for your students and make the most of every corner of your dojo! \"]}\n\nCASE 77 3748140 Syr and The Stars' Revival HOST Management extended_only\n“Syr and The Stars' Revival” follows the new Nextuber “Syr” as she sets out to restore the fallen Cyber World.\r\nStream, gain fans, and help the Cyber World flourish once again!\r\nA brand-new raising simulation where building your room is the key to streaming success!\nEVIDENCE: {}\n\nCASE 78 4601340 4 Power Marble Race Manager HOST Management no_signal\nChoose from 139 real nations, develop your athletes through training, discover young talent, and lead your country from continental competitions to the World Cup in a deep and rewarding sports management simulation.\nFULL: ABOUT THIS GAME Winning one race is easy. Keeping a nation at the top for years is the real challenge. 4 Power Marble Race Manager transforms marble racing into a deep sports management simulation where strategy matters more than reflexes. Take control of a real national team, train your athletes, discover future stars, and guide your country through years of competition in pursuit of international glory. Every season brings new rivals, new opportunities, and new decisions. Your success won't depend on quick reactions, but on your ability to build a winning team for the long term. 🌍 139 REAL NATIONS. ONE GOAL: WORLD CHAMPIONSHIP. Start your career by choosing from 139 real countries , each with its own national marble team. Every nation has unique athletes with different strengths and weaknesses. It's your job to develop them, make the right tactical decisions, and build a team capable of competing with the best in the world. Climb through continental competitions. Earn your place among the elite. Bring the World Cup home. 🏆 COMPETE ACROSS SIX CONTINENTS The world is divided into six continents , each featuring its own football-style competition structure. Compete in: National Leagues Continental Cups Qualification Tournaments The World Cup Only the strongest teams will qualify for the biggest tournaments. Every successful season opens the door to even greater challenges. 📈 TRAIN, IMPROVE AND PERFECT YOUR ATHLETES Every marble athlete is unique. Develop their abilities through training and improve attributes such as: Speed Acceleration Handling Hill Performance Power Build your team around your preferred strategy. Create balanced athletes or specialize them for specific track types. Your decisions shape every competitor's career. 🔍 DISCOVER THE NEXT GENERATION Great managers don't only win today. They build tomorrow. Scout promising young marble athletes, analyze their potential, and develop them into future champions through your youth academy. The rookie you discover today could become the superstar that wins the World Cup years later. ⏳ EVERY CAREER HAS AN END Time never stops. Athletes grow older. Their development slows down. Eventually, they retire. Long-term success depends on replacing veterans with the next generation at the right moment. Managing your squad across multiple seasons is just as important as winning your next race. 🗺️ MORE THAN 60 UNIQUE TRACKS Race across 60+ unique tracks inspired by locations around the world. Every course presents different challenges. Some reward pure speed. Others demand excellent handling, power, or hill-climbing ability. There is no single perfect athlete for every race. Success comes from preparing the right competitors for the right track. 🎥 PREPARE YOUR TEAM. WATCH THEM COMPETE. Once your preparations are complete, it's time to let your athletes perform. Send your national team onto the track and watch every race unfold. Every overtake. Every corner. Every finish. Every victory—or defeat—is the result of the choices you made as manager. Will your country become the next world champion? KEY FEATURES 🌍 Manage one of 139 real national teams . 🏆 Compete across 6 continents , featuring leagues, continental cups, qualification tournaments, and the World Cup. 📈 Develop athletes through a detailed training system. 🔍 Scout and develop young talent for long-term success. ⏳ Experience aging, slower progression, and retirement mechanics. 🗺️ Race on 60+ unique tracks inspired by locations around the world. 🎥 Watch your athletes compete after preparing them for each event. 📊 Build a dynasty and lead your nation to the top of the world rankings.\n\nCASE 79 2369730 Kaiten Sushi VR HOST Management no_signal\nKaiten Sushi VR is a game that you can experience basic operations of a conveyor belt sushi restaurant.\nFULL: ABOUT: Kaiten Sushi VR is a game that you can experience basic operations of a conveyor belt sushi restaurant. Make sushi and serve to your customers within the time limit and make them happy. There are four game modes: - Endless Mode: No time limit but if you make five customers angry, game will over. - Time Limit Mode: In 10 minutes, serve as much sushi as possible. - Ninja Mode: The ninja is coming to kitchen and attacks you. Protect yourself from the ninja. - Free Mode: There is no limit. You can practice your skills in this mode. Enjoy!\n"}

Script completed
Wall time 0.1 seconds
Output:

{"chunk_id":"c8709e","wall_time_seconds":0.000003264,"exit_code":0,"original_token_count":2821,"output":"\nCASE 80 3761740 Arrow Dungeon HOST Fighting strict\nA tactical action roguelike where you survive a dark dungeon with limited arrows, light-based exploration, and fast-paced bow combat.\nEVIDENCE: {\"loadout\": [\"lore branching paths, hidden chambers, traps, and random events. - ⚙️ Skill Unlock and Gear-Based Progression Earn new active skills and equip powerful gear that changes your playstyle. - 🎲 Challenge Modes and Replayability Conquer time trials, resource restrictions, and other special modes to test your mastery. Are you re\"]}\n\nCASE 81 4222860 Void Fighters HOST Fighting strict\nVoid Fighters is a multiplayer roguelike where you fight your way through the mysterious Void Dungeon alone or together with friends.\r\n\r\nChoose your upgrades carefully, create powerful builds, and collect Void Essence during your runs to unlock cosmetics and characters.\r\n\nEVIDENCE: {\"build\": [\"e or together with friends.\\r\\n\\r\\nChoose your upgrades carefully, create powerful builds, and collect Void Essence during your runs to unlock cosmetics and characters.\\r\\n Void Fighters is a multiplayer roguelike where you fight y\"]}\n\nCASE 82 3645920 Perk Up HOST Fighting strict\nPerk Up! is a multiplayer deck-building fighting game where every defeat makes you stronger. Each time you fall, you gain new powers, evolving into an unstoppable force. Keep fighting, adapt, and outlast everyone to become the last one standing!\nEVIDENCE: {\"deck\": [\"Perk Up! is a multiplayer deck-building fighting game where every defeat makes you stronger. Each time you fall, you gain new powers, evolving into an unstoppable force. Keep figh\"]}\n\nCASE 83 2933060 Haloo HOST Fighting extended_only\nFight the endless mass of enemies, collect loot, get stronger and Haloo!\nEVIDENCE: {}\n\nCASE 84 4817450 Bounce Arena Simulator HOST Fighting extended_only\nSet up, predict, and spectate! A stream-friendly battle simulator featuring physics auto-battles with 20 unique weapons. Enjoy YouTube, Twitch, and Discord chat integration, viewer predictions, and a 60-floor PvE Tower of Challenge. No controls needed!\nEVIDENCE: {}\n\nCASE 85 3624390 Mordeny HOST Fighting extended_only\nDive into Mordeny, a fast-paced 2D top-down shooter! Unlock abilities, customize your character, and battle through unique game modes. Whether solo or with friends, each fight demands skill and precision. Join the action today!\nEVIDENCE: {\"parts\": [\"tactical thinking are your best weapons. Explore unique gamemodes and customize your own character as you unlock new abilities and gear. Whether you're battling solo or teaming up with friends, Mordeny offers an action-packed adventure like no other where every batt\"]}\n\nCASE 86 4214230 Final 4 Survival HOST Fighting no_signal\nFinal 4 Survival is an Arcade style top-down shooter. You play as an elemental soldier and are tasked with having to defeat waves of enemies!\nFULL: Immerse yourself in this arcade‑style top‑down shooter. You’ll play as an elemental soldier tasked with defeating waves of zombies and other horrors. Each enemy type has its own unique ability, and defeating them grants you XP that can be used to upgrade your weapons, abilities, or elemental features. Elemental features include production (amount generated per second) and cap (maximum amount you can hold). Combat requires careful management of your element, as actions like running, shooting, and using specials will consume it. Without your element, you’re as good as gone...\n\nCASE 87 2146610 Reality Rash HOST Fighting no_signal\nA 3D beat 'em up inspired by the classics — fight your way through 4 different worlds.\r\nPlay solo or in 2-player co-op to gather the 4 cores and save your home in this nostalgic and modern beat 'em up!\nFULL: Resume: A corporation creates a machine capable of traveling between realities in order to save humanity in a future on the brink of imminent extinction of life on Earth. A tragedy happens when the machine fails, causing 4 distinct realities to merge and the 4 cores of the machine are dispersed among the merged territories. Amidst the chaos, whoever gathers the 4 cores can successfully restore one reality, causing the others to disappear within a period of 72 hours. Featuring: Arcade gameplay! No skill tree, all moves and combos unlocked from the start! Classic gameplay! Easy to learn, hard to master. Straight to the action! No boring, long texts Original soundtracks 20 unique stages! Play solo or co-op local this nostalgic beat 'em up!\n\nCASE 88 4939290 Shopenkraft's Quantum Curiosities HOST Job/hobby simulation strict\nTest your creativity! Develop your workshop and assemble weird alien components into a wide variety of wacky guns!\nEVIDENCE: {\"synergy\": [\"m the back of your van spaceship! The game features over 50 different modules that you can freely mix and match to create all sorts of useful (or dysfunctional) products. Some modules cancel each other out, but others create unique effects and functio\"]}\n\nCASE 89 2286780 Aeruta HOST Job/hobby simulation strict\nA bread revolution across the world now begins!\r\nAeruta is a side-scrolling action game with a shop management twist, fight monsters for materials and use them to make breads, then finally make the deserted town popular again!\nEVIDENCE: {\"loadout\": [\"style! Use your weapon and skills to create a path! There're multiple weapons to choose from, They all have a unique skilltree to make your own playstyle! Multiple areas with different theme and challenges! The locals are cute but also dangerous! Rare materials! What bread can they make? Deli\"]}\n\nCASE 90 3833800 Crafter's Hollow HOST Job/hobby simulation strict\nForge gear, run a shop, and send mercenaries into danger so you don’t have to. A cozy-but-clever crafting sim where business booms, heroes bicker, and dungeons provide all the “supply chain” problems you’ll ever need.\nEVIDENCE: {\"synergy\": [\"e from “victorious” to “oops, all dead.” Specializations & Strategy – Mix and match merc classes and playstyles to find your winning team. It’s a world where smithing meets strategy, economics meets exploration, and your greatest enemy \"], \"loadout\": [\" to “oops, all dead.” Specializations & Strategy – Mix and match merc classes and playstyles to find your winning team. It’s a world where smithing meets strategy, economics meets exploration, and your greatest enemy isn’t the dunge\"]}\n\nCASE 91 3919140 Sous Raccoon HOST Job/hobby simulation extended_only\nAn adorably hardcore cooking roguelite where no two restaurants play the same. Unlock powerful upgrades, survive increasingly chaotic shifts, and prove to Chef that hiring a raccoon as his sous chef wasn't a terrible mistake.\nEVIDENCE: {}\n\nCASE 92 3197560 What's On The Menu? HOST Job/hobby simulation extended_only\nIn What's on the Menu?, step into the shoes of a restaurant manager. From selecting ingredients and crafting menus to hiring staff and customizing your restaurant, every decision counts. Face random events, unlock new recipes, and rise through the ranks to become the ultimate restaurateur!\nEVIDENCE: {\"parts\": [\"ts and increase your restaurant’s reputation. Key Features Design and customize your restaurant with furniture, decor, and kitchen equipment. Create unique menus by selecting ingredients and crafting dishes that match your restaurant’s style. Hire and manage staff to ensure smoot\"]}\n\nCASE 93 3075060 Doctor Life Simulator HOST Job/hobby simulation extended_only\nStep into the shoes of a dedicated medical professional in Doctor Life Simulator. Experience the challenges and triumphs of running your own medical practice, from diagnosing complex illnesses to performing life-saving surgeries.\nEVIDENCE: {\"parts\": [\"xpertise. Whether it’s performing advanced medical procedures or fine-tuning your skills, Doctor Life Simulator keeps pushing you to be the best doctor you can be. Narrative World In Doctor Life Simulator , the patients aren’t j\"]}\n\nCASE 94 3771390 Carefully Stamped HOST Job/hobby simulation no_signal\nCraft and sell charming postcards to your customers in a relaxing sandbox full of color and creativity. Carefully Stamped is a cozy job simulator where you are a postcard designer. Take on new orders, craft the cutest mail, write a wholesome note and collect stamps to send your postcards 🎟️✉️\nFULL: Welcome to the world of cozy correspondence, where you are a postcard designer! Carefully Stamped is a wholesome postcard-making sandbox game where you take on the charming job of a postcard designer. Craft delightful cards, decorate with adorable stickers, write sweet notes, and match the perfect stamp—all while enjoying relaxing tunes and satisfying vibes. Whether you’re creating personalized mail for quirky customers or building your own dreamy stamp collection, there’s always a reason to send a little joy. 📮 Features Craft & Sell: Take on customer orders and craft postcards that match their unique requests and personalities. Design Freely: Decorate with a wide variety of cute stickers, washi tapes, doodles, and more! Write Notes: Pen short, wholesome messages to make every postcard feel personal and heartfelt. Collect Stamps: Unlock and collect a colorful assortment of stamps—from vintage vibes to quirky icons. Wholesome Sandbox Mode: No timers. No pressure. Just cozy creativity and peaceful play. Relaxing Soundtrack: Enjoy calming background tunes while you let your creativity flow. Mailing Magic: Drop your finished postcards into the mailbox and watch them go on their way! 💌 A Game for Stationery Lovers If you’ve ever fallen down a sticker organization rabbit hole or lost time browsing through cute stationery, this is the game for you. Carefully Stamped turns the art of mailing into a joyful experience. So grab your favorite pen, pick out a stamp, and let the postcard magic begin!\n\nCASE 95 4039900 ZOMBUTCHER: Monster Business Simulator HOST Job/hobby simulation no_signal\nMeat shop simulator where you play as a ZOMBIE! Sell meat to humans by day, butcher them by night, and manage a dark business through stealth and resource management all while eating BRA-A-A-INS!\nFULL: ZOMBUTCHER – is a stealth action simulator in which you take on the role of a zombie butcher. Hiding in the shadows of a Louisiana town and avoiding night patrols, you will gather food and resources for other monsters. Every day begins according to plan: I pick up meat from the farmers, bring it to my shop, cut it up with a knife, and pack everything into convenient bags. Zombies aren't the strongest monsters, but you have the shadows on your side, so getting brains and new consumables for goods isn't a difficult task. A stranger in a dark alley is prey. But remember, you're not the only hunter on the streets. During the day – for people, at night – from people. The main thing is not to get confused! A piece of human flesh may taste similar to chicken, but not everyone will appreciate such a delicacy. And it is unlikely that any of the nighttime shoppers will be happy with a berry smoothie instead of a bottle of fresh blood. Consumer appetites are growing! If you want to satisfy their demands, upgrade your equipment, expand your store, and seek out the highest quality ingredients so that EVERY customer leaves satisfied.\n"}

**Refining cosmetic matching**

**Refining cosmetic matching**

**Refining genre filters**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"e50b60","wall_time_seconds":1.001349708,"session_id":56403,"original_token_count":0,"output":""}

**Classifying audit results**

**Classifying audit results**

**Refining audit classifications**

**Classifying audit results**

**Refining audit classifications**

**Classifying audit results**

**Refining audit classifications**

**Finalizing audit labels**

**Classifying audit results**

**Refining audit classifications**

**Finalizing audit labels**

**Classifying audit results**

**Refining audit classifications**

**Finalizing audit labels**

**Computing comparison metrics**

Script completed
Wall time 0.0 seconds
Output:

{}
{"chunk_id":"a6babc","wall_time_seconds":0.000001825,"exit_code":0,"original_token_count":53,"output":"Classified 10000\nClassified 20000\nClassified 30000\nClassified 40000\nClassified 50000\nClassified 60000\nClassified 70000\nPrepared 75860 games; 96 unique audit cases.\nStrict 3977 extended 6185 additional strict 525\n"}

**Running original audit**

**Running original audit**

**Reviewing extended audit rules**

**Running original audit**

**Reviewing extended audit rules**

Script completed
Wall time 1.6 seconds
Output:

{}
{"chunk_id":"5a3f40","wall_time_seconds":1.000574687,"session_id":12460,"original_token_count":0,"output":""}
{}
{"chunk_id":"484bf1","wall_time_seconds":0.252266807,"exit_code":0,"original_token_count":113,"output":"judgment          broader_configuration  clear_setup  no_confirmed_configuration  unclear\noriginal_stratum                                                                         \nextended_only                        12            0                          11       13\nno_signal                             5            0                          18        1\nstrict                                1           29                           3        3\n"}

**Refining the signal audit**

**Refining the signal audit**

**Optimizing expected computations**

**Refining the signal audit**

**Optimizing expected computations**

Script completed
Wall time 5.0 seconds
Output:

{}
{"chunk_id":"a78bfb","wall_time_seconds":5.00317769,"session_id":12460,"original_token_count":30,"output":"Classified 10000\nClassified 20000\nClassified 30000\nClassified 40000\nClassified 50000\nClassified 60000\nClassified 70000\n"}

**Awaiting numeric completion**

Script completed
Wall time 0.7 seconds
Output:

{"chunk_id":"b00db6","wall_time_seconds":0.503291074,"exit_code":0,"original_token_count":273,"output":"                             name                    host    judgment\n                    Blood Dungeon             Platforming clear_setup\n     Anti Terrorist Shooting Game   First-person shooting clear_setup\n                   Punch to Glory                  Sports clear_setup\n                   Pro Jank Footy                  Sports clear_setup\n                         Goaldeck                  Sports clear_setup\n                    Metal Thunder Flight/space simulation clear_setup\n                      HELL GALAXY Flight/space simulation clear_setup\n                          Shinobi           Stealth/heist clear_setup\n                           Y2Kill                  Horror clear_setup\n                 The Sage's Seven                  Puzzle clear_setup\n                   Sizzle & Stack              Management clear_setup\n                          Perk Up                Fighting clear_setup\nShopenkraft's Quantum Curiosities    Job/hobby simulation clear_setup\n                 Crafter's Hollow    Job/hobby simulation clear_setup\nAll revised strict: {'clear_setup': 26}\n"}
{"chunk_id":"338fb6","wall_time_seconds":0.000001294,"exit_code":0,"original_token_count":24,"output":"Prepared 75860 games; 96 unique audit cases.\nStrict 3977 extended 6155 additional strict 525\n"}

**Checking genre filter coverage**

Script completed
Wall time 1.2 seconds
Output:

{"chunk_id":"380e48","wall_time_seconds":1.001145231,"session_id":37436,"original_token_count":0,"output":""}

**Preparing prevalence chart**

**Preparing prevalence chart**

**Defining chart thresholds**

**Preparing prevalence chart**

**Defining chart thresholds**

Script completed
Wall time 5.0 seconds
Output:

{"chunk_id":"67a08a","wall_time_seconds":5.000553876,"session_id":37436,"original_token_count":0,"output":""}

**Preparing local validation**

**Preparing local validation**

**Building validation checks**

Script completed
Wall time 5.0 seconds
Output:

{"chunk_id":"c70739","wall_time_seconds":5.000917307,"session_id":37436,"original_token_count":5,"output":"Finished 2019_2022\n"}

**Preparing earliest developer credits**

**Preparing earliest developer credits**

**Mapping earliest developer credits**

**Preparing earliest developer credits**

**Mapping earliest developer credits**

**Preparing earliest developer credits**

**Mapping earliest developer credits**

Script completed
Wall time 5.0 seconds
Output:

{"chunk_id":"d0e9f0","wall_time_seconds":5.000763728,"session_id":37436,"original_token_count":0,"output":""}
{"chunk_id":"171ad3","wall_time_seconds":0.000002329,"exit_code":0,"original_token_count":86,"output":"1050717       05:06 99.7 steam-research/.venv/bin/python steam-research/build-destinations-2026-09-06/analyze.py\n1063636       00:00  0.0 /usr/bin/bash -c ps -eo pid,etime,pcpu,args | rg 'python steam-research/build-destinations-2026-09-06/analyze.py'\n1063638       00:00  0.0 rg python steam-research/build-destinations-2026-09-06/analyze.py\n"}

**Analysis is still running**

**Analysis is still running**

Script completed
Wall time 1.0 seconds
Output:

{"chunk_id":"ad7735","wall_time_seconds":0.848155906,"exit_code":0,"original_token_count":595,"output":"                     host    n  builds  share  extended  reached100\n         Action roguelike 1696     461   27.2       606         111\n             Deckbuilding  724     589   81.4       603         160\n             Auto battler  201      88   43.8       112          31\n               Action RPG  739      71    9.6       161          31\n             Tactical RPG  895     132   14.7       238          43\n            Tower defense  699      86   12.3       136          15\n              Platforming 2735      32    1.2        71           6\n             Metroidvania  253       6    2.4        26           3\n    First-person shooting 1005      57    5.7        93          24\n  Top-down/arena shooting  825     135   16.4       194          27\n              Shoot-em-up  651     122   18.7       155          22\n                   Racing  608       5    0.8        63           2\n       Driving simulation  743       6    0.8        70           1\n  Flight/space simulation  410      11    2.7        31           2\n                   Sports  666      17    2.6        90           6\n                   Rhythm  238      11    4.6        17           3\n            Stealth/heist  291       7    2.4        22           6\n                   Horror 3607      25    0.7        56           6\n        Survival/crafting 1176      74    6.3       137          17\n                   Puzzle 5099      45    0.9        72           7\n                  Match-3  185       7    3.8         9           1\n                  Pinball   45       1    2.2         2           0\n      Board/dice/tabletop  530      61   11.5        74          15\n     City/colony building  386      16    4.1        21          10\n               Automation  168       8    4.8        13           2\n               Management 1228      49    4.0       103          15\n                  Farming  319       9    2.8        15           5\n                  Fishing  122       5    4.1         9           2\n                 Fighting  241       6    2.5        28           4\n               Beat-em-up   93       5    5.4         6           1\n       Real-time strategy  313      24    7.7        55           8\n      Turn-based strategy  998     168   16.8       253          53\nAdventure/point-and-click  857       1    0.1         2           0\n     Job/hobby simulation  402       5    1.2        12           4\n"}

**Checking build-screen counts**

**Checking build-screen counts**

**Refining build detection**

**Checking build-screen counts**

**Refining build detection**

Script completed
Wall time 0.0 seconds
Output:

{"chunk_id":"b29dcb","wall_time_seconds":0.000001375,"exit_code":0,"original_token_count":1609,"output":"Warning: truncated output (original token count: 1609)\nTotal output lines: 47\n\nFinished 2023_2025\nFinished 2026_JanAug\nMAIN\n                      host    n  b_n  share  b_ge100  c_ge100  b_median  c_median  ratio100  text_ratio100  b_ge1000  price_date_support  host_text_coverage\n         Action roguelike 1696  461  0.272      111      229      18.0      10.0     1.663          1.546        29               0.790               0.660\n             Deckbuilding  724  589  0.814      160       16      22.0      11.0     2.214          2.214        41               0.000               0.823\n             Auto battler  201   88  0.438       31       22      30.5       9.0     1.892          1.892         7               0.000               0.452\n               Action RPG  739   71  0.096       31      120      47.0       6.0     2.770          2.770         9               0.268               0.373\n             Tactical RPG  895  132  0.147       43      147      24.5       9.0     1.814          1.815        12               0.280               0.574\n            Tower defense  699   86  0.123       15       62      17.0       3.0     2.060          2.060         3               0.453               0.654\n              Platforming 2735   32  0.012        6      249      23.0       4.0     1.271          1.282         2               0.969               0.407\n             Metroidvania  253    6  0.024        3       67      71.5      19.0     1.760          1.760         0               0.000               0.460\n    First-person shooting 1005   57  0.057       24      171      34.0       7.0     2.430          2.424        13               0.509               0.495\n  Top-down/arena shooting  825  135  0.164       27       63      19.0       4.0     2.746          2.729         6               0.674               0.414\n              Shoot-em-up  651  122  0.187       22       67      17.5       7.0     2.153          2.153         7               0.508               0.238\n                   Racing  608    5  0.008        2       79      84.0       5.0     3.606          3.606         1               0.000               0.447\n       Driving simulation  743    6  0.008        1      154      44.0      11.0     0.956          0.956         1               0.167               0.876\n  Flight/space simulation  410   11  0.027        2       49       7.0       5.0     1.508          1.508         0               0.091               0.568\n                   Sports  666   17  0.026        6       98      42.0       9.0     2.416          2.416         1               0.294               0.456\n                   Rhythm  238   11  0.046        3       42      70.0       8.0     1.528          1.528         0               0.000               0.460\n            Stealth/heist  291    7  0.024        6       75    3092.0      16.0     3.317          3.317         5               0.000               0.490\n                   Horror 3607   25  0.007        6      653      12.0      14.0     1.102          1.083         3               0.760           …109 tokens truncated…   0               0.000               0.464\n                  Pinball   45    1  0.022        0        4      32.0       8.5     0.000          0.000         0               0.000               0.900\n      Board/dice/tabletop  530   61  0.115       15       48      17.0       4.0     2.460          2.460         7               0.148               0.392\n     City/colony building  386   16  0.041       10      130     167.5      28.0     1.966          1.966         3               0.000               0.457\n               Automation  168    8  0.048        2       68      20.0      67.0     0.719          0.719         0               0.000               0.571\n               Management 1228   49  0.040       15      385      42.0      26.0     1.026          1.006         5               0.857               0.723\n                  Farming  319    9  0.028        5      105     409.0      22.5     1.680          1.680         1               0.000               0.660\n                  Fishing  122    5  0.041        2       29      35.0      15.0     1.708          1.708         0               0.000               0.824\n                 Fighting  241    6  0.025        4       46     492.0       8.0     2.646          2.646         2               0.000               0.215\n               Beat-em-up   93    5  0.054        1       17      57.0       8.5     1.043          1.043         1               0.000               0.110\n       Real-time strategy  313   24  0.077        8       69      20.0      11.0     1.384          1.384         4               0.000               0.234\n      Turn-based strategy  998  168  0.168       53      147      27.0       9.0     1.949          1.953        16               0.446               0.436\nAdventure/point-and-click  857    1  0.001        0      149      10.0      16.0     0.000          0.000         0               0.000               0.219\n     Job/hobby simulation  402    5  0.012        4      120     263.0      20.0     2.920          2.920         1               0.000               0.794\nSPARSE PASSES\n                  host  b_n  b_ge100  responding_developers  ratio100  without_pub_ratio100  without_top3_ratio100\n           Action RPG   71       31                     30     2.770                 2.719                  2.610\n         Tactical RPG  132       43                     43     1.814                 1.786                  1.728\n        Tower defense   86       15                     15     2.060                 1.972                  1.740\nFirst-person shooting   57       24                     23     2.430                 2.396                  2.293\n               Sports   17        6                      6     2.416                 2.224                  1.567\n  Board/dice/tabletop   61       15                     14     2.460                 2.336                  2.094\n City/colony building   16       10                     10     1.966                 1.962                  1.772\n"}

**Checking date-price support**

**Checking date-price support**

**Matching controls by date-price**

**Checking date-price support**

**Matching controls by date-price**

**Checking date-price support**

**Matching controls by date-price**

**Setting date-price filters**

Script completed
Wall time 1.5 seconds
Output:

{}
{"chunk_id":"6ccba8","wall_time_seconds":1.001249085,"session_id":80984,"original_token_count":0,"output":""}
{"chunk_id":"995246","wall_time_seconds":0.177444434,"exit_code":0,"original_token_count":884,"output":"                     host    n  b_n  share  b_ge100  c_ge100  b_median  c_median  ratio100  b_ge1000\n         Action roguelike  675  326  0.483       39       39       4.0       3.0     1.144         7\n             Deckbuilding  414  363  0.877       56        4       7.0       2.0     1.967         9\n             Auto battler  117   61  0.521       10        4       8.0       2.5     2.295         2\n               Action RPG  226   39  0.173       12       25      24.0       4.0     2.380         7\n             Tactical RPG  310   70  0.226       19       27       9.0       2.0     2.212         3\n            Tower defense  260   71  0.273        6        8       5.0       2.0     2.086         2\n              Platforming  838   26  0.031        6       44       4.0       2.0     4.591         2\n             Metroidvania   79   12  0.152        4       12      14.0      13.0     1.861         1\n    First-person shooting  322   35  0.109        7       37      11.0       4.0     1.604         3\n  Top-down/arena shooting  297   88  0.296        8        6       2.0       1.0     2.425         3\n              Shoot-em-up  267  103  0.386       13       13       4.0       2.5     1.669         2\n                   Racing  167    4  0.024        1       13      22.5       2.0     3.135         0\n       Driving simulation  260    6  0.023        2       38      23.5       5.0     1.807         0\n  Flight/space simulation  134    5  0.037        0       15      13.0       3.0     0.000         0\n                   Sports  254   18  0.071        3       23       6.0       3.0     1.740         0\n                   Rhythm   78    4  0.051        0        8      44.5       4.0     0.000         0\n            Stealth/heist   86    2  0.023        1       13     548.5       5.0     3.231         1\n                   Horror 1224   21  0.017        7      157      55.0       7.0     2.845         1\n        Survival/crafting  461   73  0.158       13       73       8.0       6.0     0.930         4\n                   Puzzle 1503   37  0.025        6       98       5.0       2.0     3.140         1\n                  Match-3   54    7  0.130        2        0       4.0       1.0       NaN         1\n                  Pinball   21    8  0.381        1        0       1.5       4.0       NaN         0\n      Board/dice/tabletop  255   52  0.204       13        6       4.5       3.0     9.085         3\n     City/colony building  133   12  0.090        1       25       6.0      13.0     0.403         0\n               Automation   89    7  0.079        2       12       8.0      10.5     1.952         0\n               Management  703   39  0.055        9      120       3.0      12.0     1.426         0\n                  Farming  102    7  0.069        1       15      21.0       8.0     0.905         0\n                  Fishing   62    3  0.048        1       11       1.0       8.0     1.788         0\n                 Fighting   75    3  0.040        0        7       1.0       3.0     0.000         0\n               Beat-em-up   32    3  0.094        2        3     230.0       4.0     6.444         1\n       Real-time strategy   96   12  0.125        3       20      20.0       6.5     1.050         1\n      Turn-based strategy  426  109  0.256       20       28       2.0       2.0     2.551         3\nAdventure/point-and-click  249    1  0.004        0       34       1.0       7.0     0.000         0\n     Job/hobby simulation  363   11  0.030        4       72      10.0      11.0     1.657         0\n"}

**Reading 2023 puzzle data**

**Reading 2023 puzzle data**

**Waiting for analysis results**

**Reading 2023 puzzle data**

**Waiting for analysis results**

**Reading 2023 puzzle data**

**Waiting for analysis results**

**Reading 2023 puzzle data**

**Waiting for analysis results**

**Checking sports records**

Script completed
Wall time 0.4 seconds
Output:

Warning: truncated output (original token count: 10318)
Total output lines: 2

{"chunk_id":"882b5c","wall_time_seconds":0.244077115,"exit_code":0,"original_token_count":9102,"output":"Warning: truncated output (original token count: 9102)\nTotal output lines: 181\n\n\nHOST Sports N 6\n1543960 Squiggle Football reviews 143 developer Max Barry core True rogue True\nSquiggle Football is a deck-building football management roguelike, where you create a team and form powerful combinations of players, coaches, tactics and culture. Recruit, trade & innovate through nail-biting matches to reach & conquer the National League. Inspired by Australian Rules Football.\nSquiggle Football is a deck-building football management roguelike, where you create a team and form powerful combinations of players, coaches, tactics and culture. Recruit, tr\n1566200 Tape to Tape reviews 3461 developer Excellent Rectangle core False rogue True\nTape to Tape is a roguelite hockey game where every choice matters on your journey to build the best hockey team. Choose your path, hire unique players, discover synergies between them, pick unique abilities, bribe the referees or visit the Blademaster on your way to restore hockey's former glory.\nthe best hockey team. Choose your path, hire unique players, discover synergies between them, pick unique abilities, bribe the referees or visit the Blademaster on your way to restore hockey's former glory. Explore a gorgeous roguelite hockey world where \n1742360 Esports History reviews 383 developer Quantum Gear Studios core False rogue False\nEsports History is an RPG sim where you start as a regular guy, and work your way up to the very top of esports. Defeat bosses in tournaments, make a name for yourself within the esports community, and become a real pro!\nsure you don't fall behind in the race to be the best. PUT TOGETHER A UNIQUE BUILD Use a flexible talent system and choose which stats to upgrade through practice to become an unbeatable player. JOIN A COMMUNITY OF GAMERS \n1894720 Motordoom reviews 173 developer Hobo Cat Games core True rogue True\nMotördoom is a freestyle-sports roguelite horde shooter that combines high octane, visceral combat with fluid movement and trick mechanics. Perform gnarly tricks while defeating grotesque monsters and upgrade your character with demonic offerings to become ever more powerful.\nin the form of cards. Build your deck by choosing the right cards for your build. Unlock new cards for later runs to be prepared for the horrors Motördoom has to offer. Ride your chainsaw driven pit bike through a monste | se demonic abominations! Demonic Offerings come in the form of cards. Build your deck by choosing the right cards for your build. Unlock new cards for later runs to be prepared for the horrors Motördoom has to offer. Ride you | your own style and do the tricks that you want. Choose from different characters and bikes that suite your play style. Unlock new characters with scrap you collect in runs. Each character has it's unique weapon and play style. Choose which one suits your st\n2457580 Fight Crab 2 reviews 128 developer Calappa Games core False rogue False\nClash claws and flip opponents belly-up in physics-based crab battles! Combine dual-wielded weaponry with multiple fighting stances to create a custom fighting style. Bash, blast, scoop and pummel enemies onto their backs to achieve victory!\nr backs to achieve victory! Crab battles, more epic than ever before! Armor-clad humans mount immortal crabs in the world’s newest combat sport! Select a weapon loadout (one weapon for each claw, of course!), assume an advantageous fighting stance, and unleash crab-fueled chaos on opponents. Select from 20+\n2995450 Clutchtime™: Basketball Deckbuilder reviews 181 developer Bigosaur core True rogue True\nBuild your perfect deck and become a Champion.\nBuild your perfect deck and become a Champion. Clutchtime is a deck builder centered around basketball plays and team synergies . Double Team your opponents and fo | Build your perfect deck and become a Champion. Clutchtime is a deck builder centered around basketball plays and team synergies . Double Team your opponents and force Turnovers or create a dream team of Superstar shooters who Splash three pointers even while getting \n\nHOST City/colony building N 10\n1011590 Night is Coming reviews 383 developer Wild Forest Studio core False rogue False\nSurvival strategy game set in a world inspired by Slavic mythology and the nature of Eastern Europe. Build your settlement, protect its dwellers, and help them prepare for the coming Darkness.\nver, you'll be able to unlock new spells, magic boosters, new warrior subclasses, advanced armor, and weaponry. Experience a dynamic and ever-changing world: sunrises and sunsets, sounds of nature, elemental spirits, sea\n1270580 Mind Over Magic reviews 4304 developer Sparkypants core True rogue False\nMind Over Magic is a colony sim set at the site of a cursed magic school. Build and optimize your own magical academy, train powerful Mages, and master mighty relics in order to battle through the Underschool, a mysterious underground dungeon that holds the key to liberation from a deadly curse.\ngic by customizing your students' education. Apprenticed students can multiclass to wield a diverse array of spells, or strive towards single discipline mastery to unlock devastating ultimate abilities, powerful enough t\n1840710 These Doomed Isles reviews 179 developer Triplevision Games Limited core True rogue True\nHarvest cards to raise land from the sea, build settlements and rain fire upon invaders in this survival city-builder with roguelike deckbuilding mechanics. Choose from 4 gods with unique cards, playstyles and abilities to lead their civilisations to victory in strategic run-based gameplay!\n rain fire upon invaders in this survival city-builder with roguelike deckbuilding mechanics. Choose from 4 gods with unique cards, playstyles and abilities to lead their civilisations to victory in strategic run-based gam | and win the run! Harvest cards to craft the deck you want. Manage the cards in your deck to ensure you can survive until the next harvest. Find powerful synergies with over 300 cards to play with to overcome the challenges in your way. Will you survive these doomed isles? You'll face increasing challe | with roguelike deckbuilding mechanics. Choose from 4 gods with unique cards, playstyles and abilities to lead their civilisations to victory in strategic run-based gameplay! Play god in this survival city-builder with roguelike\n2071430 Roots of Yggdrasil reviews 300 developer ManaVoid Entertainment core True rogue True\nBuild settlements, explore the Nine Realms, collect artifacts and upgrade your deck to reach the top of Yggdrasil in this post-Ragnarok roguelike city-builder! Can you survive the end of the world?\nd Tree and attempt to reach the top of Yggdrasil. Create settlements, build your deck and collect artifacts as you explore the nine realms of Norse Mythology in search of answers in this roguelike city-building game. Can you  |  acts as the main hub of the game. Unlock new mechanics and permanent upgrades that will help you rebuild your tribe! Befriend a diverse cast of Scions that will bring unique playstyles, bonus and modifiers to each loop. Procedurally generated islands, encounters, and events will punctuate your experience, guaranteeing that\n2290000 TerraScape reviews 2052 developer Bitfall Studios core True rogue False\nBuild gorgeous kingdoms on floating islands by strategically placing buildings in this cozy, medieval City-Puzzler. Discover secret combinations, take on weekly challenges and explore randomly generated maps in various single- and multiplayer game modes.\noffer: Resource Management 4x Strategy Citizen Simulation Combat Free Deck building Kingdom Mode – Our main mode: Play until you run out of points and buildings Scenarios – Play with special conditions and goals to achieve  | ace your buildings to unlock more buildings, more draws, and new Keep upgrades to choose additional Decks that synergize well with what you already have. Each building benefits from different influences, be it other buildings and natural resources nearby, or t\n2346410 Border Pioneer reviews 1665 developer Yahzj Games core False rogue False\nAs the kingdom's most promising pioneer, embark on a legendary journey to tame the frontier! Build thriving towns across various landscapes, fending off waves of diverse enemies until you establish a firm foothold and have explored everything the eye can see!\new buildings and units as you progress. Collect And Expand Your Deck! Build Your Deck! Choose from over 200 unique cards to construct your perfect deck. Use your cards with skill and timing to overcome challenges and build yo\n2438260 Tiny Kingdom reviews 149 developer neltile core True rogue True\nTiny Kingdom is a minimalist city-building strategy game where every placement matters.\r\nStart with a single hut and grow a thriving kingdom by combining buildings, unlocking synergies, and creating efficient layouts.\ng your kingdom’s potential. Unlock new cards and opportunities as you build. Maximize your placements to ensure your deck grows and evolves. | d every card holds the promise of new adventures. Strategically place cards. Each card interacts with adjacent tiles, creating charming synergies and expanding your kingdom’s potential. Unlock new cards and opportunities as you build. Maximize your placements to ensure your deck grows\n2655590 Kingdom's Deck reviews 488 developer Ivan Aco core True rogue True\nBuild by day, defend by night! Kingdom's Deck is a card-based city builder where you build and protect your Kingdom from the enemy siege. Build your deck, deploy your cards wisely, defend your walls, manage your resources, and respond to various events that will shape your Kingdom!\nBuild by day, defend by night! Kingdom's Deck is a card-based city builder where you build and protect your Kingdom from the enemy siege. Build your deck, deploy your cards wisely, defe\n2749100 Dawnmaker reviews 156 developer Arpentor Studio core True rogue False\nDawnmaker is a turn-based city building and deckbuilding game, made by 2 strategy fans. It feels like playing a solo board game. Build luminous cities, optimize your deck and create powerful combinations between buildings and cards to bring life back to a dead world. Play the demo now!\nDawnmaker is a turn-based city building and deckbuilding game, made by 2 strategy fans. It feels like playing a solo board game. Build luminous cities, optimize your deck and create powerful combi\n2825110 Technotopia reviews 592 developer Yustas core True rogue False\nWelcome to Technotopia, a city-builder card game. Take on the role of an AI designed to build a perfect city. Build districts, meet the needs of diverse communities and maintain the delicate balance between factions vying for control over you!\n intelligence designed for building a utopia. Welcome to Technotopia! Build a perfect city You will have a deck of cards with buildings at your disposal. Build skyscrapers, parks, residential districts and transport highways. It's up to you to decide \n\nHOST First-person shooting N 24\n671860 BattleBit Remastered reviews 127675 developer SgtOkiDoki core False rogue False\nBattleBit Remastered is a low-poly, massive multiplayer FPS, supporting 254 players per server. Battle on a near-fully destructible map with various vehicles!\nn leaves a lasting impact on the world around you! Fully customizable weapons (WIP) Tailor your arsenal to your playstyle in BattleBit Remastered ! Modify everything from scopes to grips or barrels to mags. Create the perfect weapon loadout to dominate every ba\n704000 GUNHEAD reviews 155 developer Alientrap core False rogue True\nGunhead is a fast paced strategic roguelike FPS where you take the role of a pirate mech with a gun for a head, looting derelict spaceships filled with monstrous drones made out of bone and steel.\r\nSequel to the 2d roguelite Cryptark.\nled FPS gameplay with 4 gun slots ready to be configured with over 50 weapon options Each enemy ship is its own unique challenge. Set your loadout strategize a plan of attack to take out the complex interconnected enemy security systems. Blood-pumping soundtrack perfect for intense fir\n789940 UNYIELDER reviews 116 developer TrueWorld Studios core True rogue True\nUNYIELDER is an intense Boss-rush Movement-FPS. Assemble your arsenal of destruction, smash your way through waves of bosses, face your own Final-Boss.\n hidden incremental effects. Begin each run by selecting from curated Perk \"Suit\" loadouts. Discover synergies, experiment with builds, and break the game wide open. THE TEMPORAL VAULT Send weapons and items back in time to assist your future self. A |  hidden incremental effects. Begin each run by selecting from curated Perk \"Suit\" loadouts. Discover synergies, experiment with builds, and break the game wide open. THE TEMPORAL VAULT Send weapons and items back in time to assist\n860020 EXFIL reviews 301 developer Tiny Leviathan core False rogue False\nEXFIL delivers raw tactical FPS intensity with no distractions. No loot grinding, no gimmicks, just pure realism. EXFIL tests your strategy, comms, and teamwork in high-stakes tactical engagements. Will your squad rise to the challenge?\nters - from AI presence on the map to time of day, weather, available weapons, and respawn rules - allowing them to create unique challenges, and playstyles to suit any preference. Additional Features NPC opposition forces Global stats tracking Loadout customization Early Access and Beyond We ar\n1272080 PAYDAY 3 reviews 49117 developer Starbreeze Studios core False rogue False\nPAYDAY 3 is the ultimate co-op FPS crime fantasy. Choose between stealth or loud playstyles, customize your character, and get ready to pull off the perfect heist. \n the ultimate co-op FPS crime fantasy. Choose between stealth or loud playstyles, customize your character, and get ready to pull off the perfect heist.  PAYDAY 3 is a 1-to-4 player FPS where you live out the ultimate heist fantasy. Tackle jobs i\n1294810 Redfall reviews 2459 developer Arkane Austin core False rogue False\nThe town of Redfall is under siege by a legion of vampires who have cut the island off from the outside world. Trapped with a handful of survivors, slay alone or squad up in open-world FPS action.\nou will collect an arsenal of specialized weaponry and customize your character with unique upgrades and abilities to suit your playstyle for guns blazing, stealth combat, and everything in-between. Unconventional Heroes Vs. the Night Created by science experiments gone wrong \n1605250 Moros Protocol reviews 728 developer Pixel Reign core False rogue True\nFace the darkness in Moros Protocol, a pixel-painted sci-fi FPS roguelite. Survive the nightmare alone or with allies as you navigate the decaying warship, stalked by horrors and guided by a voice you don’t recognize. Adapt, upgrade, and push forward. You may be outnumbered, but not outmatched.\nnes relentless action with adaptability and customization. Shape your loadout and let the augments and weapons you find dictate your fighting style. And yes, you can kick your enemies. 🔺ROGUELIKE PROGRESSION Death is only the beginning. Learn, upgrad\n1721110 Abyssus reviews 4968 developer DoubleMoose Games core True rogue True\nChoose your loadout, blast through the corrupted denizens of a sunken civilization, and wield godly powers in Abyssus – a fast-paced 1-to-4 players FPS and roguelite brinepunk adventure.\nmeans survival. Team up with fellow Brinehunters and carefully choose synergic loadouts, skills, suits, and weapon mods. Work together and fight through a merciless roguelite gauntlet, where no run is the same as the last. Wield highly customizable, brin | vival. Team up with fellow Brinehunters and carefully choose synergic loadouts, skills, suits, and weapon mods. Work together and fight through a merciless roguelite gauntlet, where no run is the same as the last. Wield highly customizable, brin\n2072660 Voidborn reviews 350 developer Xekvera core True rogue True\nSlay demons and demigods in this fast-paced roguelite FPS. Create devastating weapon builds as you journey through massive dark fantasy environments on your quest to avenge god. Challenge the demigods, face inevitable death, and rise stronger with every rebirth.\nons and demigods in this fast-paced roguelite FPS. Create devastating weapon builds as you journey through massive dark fantasy environments on your quest to avenge god. Challenge the demigods, face inevitable death, and ri\n2116120 Incursion Red River reviews 8205 developer Games Of Tomorrow GmbH core False rogue False\nIRR is a tactical PvE Co-Op extract shooter set in a de-stabilized and immersive modern-day Vietnam. As a PMC for a mysterious organization, deal with 3 factions that stand to gain from the chaos. Engage in ruthless combat, plan your deployment, and explore the world as your journey unfolds.\nide a balanced first-person experience that feels realistic, yet fun. Equipment and weapons influence ergonomics and combat playstyle, so you will need to pick your loadout carefully. Access the Marketplace to purchase limited quantities of weapons, medical equipment, gear, and ammunition. Whether you prefer to\n2124120 SULFUR reviews 9402 developer Perfect Random core False rogue True\nSULFUR is a modern old-school action-adventure: Outsmart enemies, find treasure, improve weapons, harness powers. Delve deeper, find answers, make it end.\ne playful and tactile weaponry in visceral and intense combat. Create unique builds by combining weapons, gear, tools, and enchantments to suit your playstyle. Whether you’re a gun-blazing risk-taker or a careful strategist | nry in visceral and intense combat. Create unique builds by combining weapons, gear, tools, and enchantments to suit your playstyle. Whether you’re a gun-blazing risk-taker or a careful strategist, the choice is yours. Adapt your approach, experiment, and survive the dan\n2272540 FBC: Firebreak reviews 1977 developer Remedy Entertainment core False rogue False\nPurchase FBC: Firebreak and invite your friends to play the full game for free on a…102 tokens truncated…d of Con |  Kit with specialized tools, grenades, support items, and paranatural augments... then modify them to suit your strategy and change the way you play. Experiment with different loadouts to perfect your playstyle and synergize with your team, giving you the edge to succeed in every crisis, no matter the difficulty. Return to\n2312280 7 Days Blood Moons reviews 417 developer iLLOGIKA core False rogue False\n7 Days Blood Moons is a 4v1 fight to the death. Survivors fortify bases and scavenge for materials while the Zombie Master commands and possesses the horde to crush them. Play solo or with friends – up to 4 survivors can team up against a human or AI Zombie Master. Who will prevail?\nd come for you in this destructible voxel world. Progress : Pick your loadout from a large arsenal of unlockable unique weapons and perks to prepare for your next match. Craft : Get more firepower and hit harder by crafting and unlocking new weapons. Craft and unlock new const\n2369390 Far Cry® 6 reviews 33152 developer Ubisoft Toronto core False rogue False\nLead the revolution in this first-person open world guerrilla shooter. Fight with improvised weapons, travel however you can, and build your rebel network to take down Castillo's regime.\ne island nation of Yara. SCRAPPY AND EXPLOSIVE FIRST-PERSON GUERRILLA COMBAT Build improvised weapons like missile backpacks or motorbike engine miniguns and fight alongside unique animal companions who distract, scout, or\n2406770 Bodycam reviews 50189 developer Reissad Studio core False rogue False\nBodycam is the first tactical multiplayer FPS with a true body-camera view in Unreal Engine 5. Close-quarters combats feel raw and…318 tokens truncated…eveloper Stubby Games core False rogue True\nFast paced Bullet hell Roguelite FPS 🔴 Fight relentless machines driven by a hostile AI. Escape an endless cycle.\r\nCraft infinite weapon variants with deep gun modding. Forge deadly synergies. Carve your way out through enviro destruction\r\n/BREAK THE SYSTEM /BREAK FREE\nines driven by a hostile AI. Escape an endless cycle.\r\nCraft infinite weapon variants with deep gun modding. Forge deadly synergies. Carve your way out through enviro destruction\r\n/BREAK THE SYSTEM /BREAK FREE You are trapped in an endless sequence by the all-seeing, hos\n2687970 TRIBES 3: Rivals reviews 1322 developer Prophecy Games core False rogue False\nTRIBES 3: Rivals is a team-based first person shooter that elevates the genre with jetpacks, skiing, and class-based warfare. Teamwork, tactics, high speed, and intense combat are supreme in this rebirth of the world’s fastest shooter.\nwhile ferociously defending their own. Players can choose from unique classes to customize their abilities, perks, and weapon loadouts, while moving across epic terrains from snow covered peaks to haunting deserts. Teamwork, tactics, high speed, and intense combat reign sup\n2705210 Good Kill! reviews 105 developer gamegine core False rogue False\nGood Kill is an immersive FPS game where players take control of a door gunner in a military helicopter during wartime. Engage in intense combat across destructible environments, complete missions, and upgrade your helicopter and weapons to dominate the battlefield.\nd with destructible objects and buildings. Adaptability is key as you upgrade your helicopter and weapons between missions, customizing your loadout to suit the ever-evolving demands of the battlefield. The game is under development. Visuals does not represent end-quality of the game. Br\n2851700 Nasty Zombies reviews 206 developer Con No 1 core False rogue False\nTeam up with up to 8 players and take on endless waves of Nasty Zombies in a high-intensity, movement-focused boomer shooter. Use agility and smart positioning to survive semi-procedural maps and unlock permanent upgrades like Blessings, Miracles, Invocations, and cosmetics.\n are 10 Miracles . Perks that provide permanent upgrades and powerful passive bonuses to enhance your abilities. There are 13 Blessings . Powerful abilities that define your playstyle, allowing you to activate unique effects at will and build around your chosen path. There are 8 Invocations . Unlock new camos for your wea\n2881660 Holy Shoot reviews 266 developer Tale Era Interactive core True rogue True\nHoly Shoot is a fast-paced roguelite action FPS set in a satirical version of Hell, packed with chaos, blessed heroes, and demonic bosses. Pick your hero, dive into hell, upgrade your build, and shoot your way through sin.\nd heroes, and demonic bosses. Pick your hero, dive into hell, upgrade your build, and shoot your way through sin. Holy Shoot is a fast-paced roguelike FPS packed with chaos, style, and comedy. Start as a Sanctum warrior  | hat you find along the way. Build, Upgrade, and Adapt Unlock powerful perks mid-run and earn currencies to buy permanent upgrades back at HUB. Mix and match upgrades to suit your style, and build your own flavor of divine destruction. A HUB Full of Sanctum Oddballs After each chaotic run, you re\n2933080 Crime Boss: Rockay City reviews 8483 developer INGAME STUDIOS core True rogue True\nCrime Boss: Rockay City is a 1-4 player co-op FPS heist game with enjoyable stealth and epic killing spree combat. Featuring a solo roguelike campaign that combines FPS and strategy elements with the most hilariously cheesy voice acting that make you laugh. \n and co-operative game modes! Each level up allows players unlock new weapons, heisters, equipment and skins. Kit yourself for success by customising your loadouts and crew. Keep an eye out for our double XP events to climb the ladder even faster. Operate Solo in the Single-player Campaign! Experienced\n3043760 HELLBREAK reviews 272 developer Double Barrel Games core True rogue True\nHELLBREAK is a fast-paced roguelite, survivor-like FPS where you fight through escalating waves in brutal arenas in Hell. Build powerful loadouts, earn favor from demon lords to unlock new gear, and climb the leaderboards for your salvation.\nght through escalating waves in brutal arenas in Hell. Build powerful loadouts, earn favor from demon lords to unlock new gear, and climb the leaderboards for your salvation. HELLBREAK thrusts you into the infernal depths as a soldier from the World Wars, fighting t\n3228590 Deadzone: Rogue reviews 14928 developer Prophecy Games core False rogue True\nDeadzone: Rogue is a roguelite FPS set in space. Blast through swarms of hostile machines, combine powerful augments and elements to forge your build, and grow stronger with each run — solo or in co-op.\n of hostile machines, combine powerful augments and elements to forge your build, and grow stronger with each run — solo or in co-op. Adventure on your own, or conquer the ship in a co-op squad of up to 3 players. Mark e\n\nHOST Board/dice/tabletop N 15\n1652250 Tales & Tactics reviews 1727 developer Table 9 Studio core True rogue False\nRoguelike Squad-based Autobattler in a Tabletop RPG setting. Draft your army, carefully equip and position your units, make important choices, and meet a colorful cast of characters as you make ready for the Grand Tournament.\nect a team of Tanks, Healers, Mages, Engineers and more - using their synergies to turn the tide of battle in your favor. Find powerful Equipment to give to your units, expand your Army Size, unlock new and more powerful tiers of Units, and acquire unique Level-Up Perks to help you on your journey. No two \n1676840 For The King II reviews 12670 developer IronOak Games core True rogue True\nBattle against Fahrul's tyrannical Queen alone or as a party of four players in the sequel to For The King, the massively popular turn-based roguelite tabletop RPG. Will you unravel the once beloved Queen's terrible secret?\ntart stronger and wiser on your next playthrough with better starting items to draft into your loadout. Roll up, roll up! Put your skills to the ultimate test with the infinite dungeon mode - Dark Carnival. Led by the devious Ringmaster, your\n1689500 Dune: Imperium reviews 3792 developer Dire Wolf core True rogue False\nThe award-winning board game of strategy and intrigue in the iconic universe of Dune. The Spice must flow.\nyours, but war is not the only way to claim it. Dune: Imperium blends deck-building and worker placement in a deeply thematic new strategy game where the fate of the Empire hangs on your decisions. Will you seek political a\n1722870 Clank! reviews 537 developer Dire Wolf core True rogue False\nBurgle your way to adventure in the award-winning board game!\nway to adventure in the award-winning board game! Clank! , the iconic deck-building board game adventure, challenges daring thieves to sneak into a dragon’s lair, steal one of her precious artifacts, and escape – before you\n1755830 Astrea: Six-Sided Oracles reviews 2986 developer Little Leo Games core True rogue True\nAstrea is a DICE-deck-building roguelike that flips the script on deckbuilders by using dice instead of cards and a unique dual “damage” system: Purification vs Corruption. Build a dice pool strong enough to purify Astrea's out-of-control corruption and save the Star System.\nAstrea is a DICE-deck-building roguelike that flips the script on deckbuilders by using dice instead of cards and a unique dual “damage” system: Purification vs Corruptio | ul, if you take too much Corruption you will be consumed by it. • Not cards, but dice! - Build a dice pool that fits your playstyle. Choose from over 350 dice and three dice types; reliably safe, perfectly balanced, or powerfully risky. A dice type system designed with h\n1990110 SpellRogue reviews 1559 developer Guidelight Games core True rogue True\nSpellRogue is the ultimate tactical turn-based roguelike! Draft from over 350 powerful spells to assemble the perfect combination and crush your enemies in 70+ unique encounters. Face brutal choices, adapt to impossible odds, and strike back at the horrors that destroyed the world.\nities to challenge your strategy. 💰145+ artifacts to shape and refine your build. 🎲 Prove your mastery by boosting the difficulty with the Descent system. 🏆 Take on Daily Challenges with ever-changing twists. 🏹 Battle th | e manipulation and wield powerful spells in SpellRogue , a turn-based deckbuilding roguelike where every roll shapes your fate. Control the fickle Mana Dice by rerolling, duplicating, splitting, flipping, and enchanting th\n2096510 The Ouroboros King reviews 464 developer Oriol Cosp Games core False rogue True\nThe Ouroboros King combines the strategic depth of chess with the build variety and replayability of roguelikes. Assemble a formidable army, discover powerful relics, and buy surprising gadgets to defeat the Coven.\nThe Ouroboros King combines the strategic depth of chess with the build variety and replayability of roguelikes. Assemble a formidable army, discover powerful relics, and buy surprising gadgets to defeat the Coven. The \n2347080 Frosthaven reviews 807 developer Snapshot Games Inc. core True rogue False\nFrom the creative mind of Isaac Childres, the highly acclaimed board game moves to the digital realm in the video game adaptation of FROSTHAVEN! The dark fantasy tactical RPG features deep mechanics and challenging yet satisfying gameplay both in Single Player and Co-Op Online Multiplayer.\ntters. Players must balance risk and reward, working towards the best synergies between the classes comprising their party. Carefully managing resources is paramount to progress through a gripping campaign, set in a vast world filled with  | res will ensure all players can enjoy their experience. The brand-new character classes, each featuring unique playstyles and abilities, makes it so that every decision matters. Players must balance risk and reward, working towards the best synergies between th\n2693930 Dice & Fold reviews 3486 developer Tinymice Entertainment core True rogue True\nA turn-based dungeon crawler where your only way of attacking the enemy is by rolling dice. Roll dice to damage enemy cards to FOLD. Inspired by classic Solitaire, in a top-down table and enemies made out of cards, try to beat the dungeon with 35 unique classes. Build your deck, kill monsters.\nes made out of cards, try to beat the dungeon with 35 unique classes. Build your deck, kill monsters. Dice & Fold is a game that combines the strategic depth of turn-based dungeon crawlers with the thrill of chance inherent i | enges and rewards. As the player you will choose the enemy to fold or companions you like to combine with your hero skill synergy. Between stages, players have the opportunity to choose items. These upgrades range from enhancing dice rolls and bolstering class abilitie | reservation. Adding to the complexity, the game features 35+ distinct classes, each with unique abilities and playstyles. This diversity allows for varied approaches to beating the dungeon, encouraging replayability and experimentation. ROGUELIKE & DECKBUILDIN\n2749100 Dawnmaker reviews 156 developer Arpentor Studio core True rogue False\nDawnmaker is a turn-based city building and deckbuilding game, made by 2 strategy fans. It feels like playing a solo board game. Build luminous cities, optimize your deck and create powerful combinations between buildings and cards to bring life back to a dead world. Play the demo now!\nDawnmaker is a turn-based city building and deckbuilding game, made by 2 strategy fans. It feels like playing a solo board game. Build luminous cities, optimize your deck and create powerful combi\n2797650 Pip My Dice reviews 135 developer Diving Swan Games core True rogue True\nPip My Dice is a Yahtzee Roguelike experience. Score enough to make it to the next level, customize your dice to score more and collect relics to create unlimited combos! Collect money through the rounds to buy new dice, relics, and customizers to upgrade your dice. Enjoy a different run every time!\na different run every time! About the game Pip My Dice is a roguelike deck builder inspired by Balatro and the Yahtzee game. Customize different dice and use them together with powerful relics to break high scores and clim | ee game. Customize different dice and use them together with powerful relics to break high scores and climb the leaderboard. Collect powerful relics to create synergies with your dice and upgrade your combination slots. Take your time and make wise decisions about what you buy in the shop and how you use it\n2909580 Viractal: Will You Trust Your Party? reviews 441 developer Sting core True rogue False\nForm a party of up to three characters in this board game-style RPG, where every member joins the fight! Level up each character, develop their abilities, collect items, and build individual decks. Combine skill cards as dice rolls and strategy create a new adventure every time you play!\n Level up each character, develop their abilities, collect items, and build individual decks. Combine skill cards as dice rolls and strategy create a new adventure every time you play! Deep in a throbbing headache, you awaken in a s\n3415570 Die For The Lich reviews 273 developer Monovoid core True rogue True\nGather magical dice, adapt your gear and venture through a land haunted by folklore creatures.\nie for the Lich, we set out to recreate the magic of playing a unique deck-builder for the first time. We replaced cards with dice that you need to pair with your gear and the game took on a life of its own, a curious blen\n3444020 Demonic Mahjong reviews 3857 developer Boxed Lighting Games core True rogue True\nDemonic Mahjong is a Roguelike deckbuilding game inspired by the timeless mechanics of Mahjong. Masterfully use the tiles in your hand to craft powerful combinations, reach staggering high scores, and develop strategies that are uniquely your own!\nDemonic Mahjong is a Roguelike deckbuilding game inspired by the timeless mechanics of Mahjong. Masterfully use the tiles in your hand to craft powerful combinations, reach staggering\n3548520 DOG WITCH reviews 322 developer Heckmouse core True rogue True\nDog Witch is a roguelike deckbuilder where your spells are fueled by enchanted dice. Combine over 150 unique items to forge ever-evolving strategies, perfect synergy combos, summon minions, and tackle an endless range of foes so bizarre, they’ll leave your tail wagging in disbelief!\r\n\nDog Witch is a roguelike deckbuilder where your spells are fueled by enchanted dice. Combine over 150 unique items to forge ever-evolving strategies, perfect synergy combos, su | ere your spells are fueled by enchanted dice. Combine over 150 unique items to forge ever-evolving strategies, perfect synergy combos, summon minions, and tackle an endless range of foes so bizarre, they’ll leave your tail wagging in disbelief!\r\n DOG WITCH combines \n"}
{"chunk_id":"548874","wall_time_seconds":0.000001221,"exit_code":0,"original_token_count":1178,"output":"MATCHED MAIN\n                      host  build_n  supported_n  observed100  expected100  ratio100\n         Action roguelike      461          460          110       93.750     1.173\n             Deckbuilding      589          496          117       60.700     1.928\n             Auto battler       88           78           25       14.250     1.754\n               Action RPG       71           67           29       15.850     1.830\n             Tactical RPG      132          130           42       33.350     1.259\n            Tower defense       86           86           15       13.400     1.119\n              Platforming       32           32            6        6.000     1.000\n             Metroidvania        6            6            3        2.600     1.154\n    First-person shooting       57           56           23       16.467     1.397\n  Top-down/arena shooting      135          134           27       14.467     1.866\n              Shoot-em-up      122          120           21       16.400     1.280\n                   Racing        5            4            1        1.600     0.625\n       Driving simulation        6            6            1        1.600     0.625\n  Flight/space simulation       11           11            2        0.800     2.500\n                   Sports       17           17            6        3.200     1.875\n                   Rhythm       11           10            3        1.800     1.667\n            Stealth/heist        7            3            2        1.200     1.667\n                   Horror       25           25            6       10.000     0.600\n        Survival/crafting       74           73           17       14.800     1.149\n                   Puzzle       45           45            7        6.600     1.061\n                  Match-3        7            7            1        0.200     5.000\n                  Pinball        1            0            0        0.000       NaN\n      Board/dice/tabletop       61           55           10        5.600     1.786\n     City/colony building       16           16           10        5.400     1.852\n               Automation        8            7            2        1.400     1.429\n               Management       49           49           15       15.000     1.000\n                  Farming        9            9            5        2.600     1.923\n                  Fishing        5            3            1        1.667     0.600\n                 Fighting        6            6            4        1.000     4.000\n               Beat-em-up        5            2            0        0.000       NaN\n       Real-time strategy       24           23            8        4.533     1.765\n      Turn-based strategy      168          166           52       37.850     1.374\nAdventure/point-and-click        1            1            0        0.000       NaN\n     Job/hobby simulation        5            4            3        1.467     2.045\nCANDIDATES\n                  host  b_n  b_ge100  supported_n  support  matched_ratio100  no_core_build_n  no_core_build_ge100  no_rogue_build_n  no_rogue_build_ge100  no_prior_current1000_build_n  no_prior_current1000_build_ge100\n           Action RPG   71       31           67    0.944             1.830                0                    0                47                    23                            61                                21\n         Tactical RPG  132       43          130    0.985             1.259                0                    0                80                    23                           112                                26\n        Tower defense   86       15           86    1.000             1.119               42                    8                31                     7                            83                                14\nFirst-person shooting   57       24           56    0.982             1.397               43                   17                37                    12                            48                                17\n               Sports   17        6           17    1.000             1.875                9                    3                11                     2                            15                                 5\n  Board/dice/tabletop   61       15           55    0.902             1.786               21                    1                29                     6                            54                                10\n City/colony building   16       10           16    1.000             1.852                4                    2                 9                     6                            16                                10\n"}

**Checking a few ratios**

**Checking a few ratios**

**Refining genre labels**

**Checking a few ratios**

**Refining genre labels**

**Checking a few ratios**

**Refining genre labels**

**Checking genre counts**

**Checking a few ratios**

**Refining genre labels**

**Checking genre counts**

**Checking a few ratios**

**Refining genre labels**

**Checking genre counts**

**Checking a few ratios**

**Refining genre labels**

**Checking genre counts**

**Saving genre overlap matrix**

**Checking a few ratios**

**Refining genre labels**

**Checking genre counts**

**Saving genre overlap matrix**

Script completed
Wall time 0.3 seconds
Output:

{"chunk_id":"101144","wall_time_seconds":0.177407747,"exit_code":0,"original_token_count":1367,"output":"     period                    host  build_n  supported_n  observed100  expected100  ratio100  observed1000  expected1000  ratio1000\n  2019_2022           Tower defense       21           21            7        4.867     1.438             2         1.033      1.935\n  2019_2022             Platforming       25           25           12        6.000     2.000             1         1.200      0.833\n  2019_2022   First-person shooting       29           26           14        9.900     1.414            10         5.567      1.796\n  2019_2022 Top-down/arena shooting       39           36            8        4.600     1.739             4         0.800      5.000\n  2019_2022                  Sports        5            5            3        1.350     2.222             1         0.450      2.222\n  2019_2022                  Horror        9            8            3        4.600     0.652             2         2.200      0.909\n  2019_2022                  Puzzle       15           15            7        4.800     1.458             3         1.800      1.667\n  2019_2022     Board/dice/tabletop       17           14            5        4.600     1.087             0         1.000      0.000\n  2019_2022    City/colony building        3            3            2        1.000     2.000             0         1.000      0.000\n  2019_2022              Management       13           13            7        5.200     1.346             2         2.600      0.769\n  2023_2025           Tower defense       86           86           15       13.400     1.119             3         3.200      0.938\n  2023_2025             Platforming       32           32            6        6.000     1.000             2         1.200      1.667\n  2023_2025   First-person shooting       57           56           23       16.467     1.397            12         7.133      1.682\n  2023_2025 Top-down/arena shooting      135          134           27       14.467     1.866             6         4.533      1.324\n  2023_2025                  Sports       17           17            6        3.200     1.875             1         1.000      1.000\n  2023_2025                  Horror       25           25            6       10.000     0.600             3         3.000      1.000\n  2023_2025                  Puzzle       45           45            7        6.600     1.061             2         1.800      1.111\n  2023_2025     Board/dice/tabletop       61           55           10        5.600     1.786             4         0.400     10.000\n  2023_2025    City/colony building       16           16           10        5.400     1.852             3         1.600      1.875\n  2023_2025              Management       49           49           15       15.000     1.000             5         4.000      1.250\n2026_JanAug           Tower defense       71           71            6        4.100     1.463             2         1.500      1.333\n2026_JanAug             Platforming       26           26            6        4.217     1.423             2         0.400      5.000\n2026_JanAug   First-person shooting       35           33            5        7.800     0.641             1         1.733      0.577\n2026_JanAug Top-down/arena shooting       88           88            8        1.883     4.248             3         0.600      5.000\n2026_JanAug                  Sports       18           17            2        0.600     3.333             0         0.200      0.000\n2026_JanAug                  Horror       21           21            7        4.600     1.522             1         1.400      0.714\n2026_JanAug                  Puzzle       37           37            6        3.467     1.731             1         0.733      1.364\n2026_JanAug     Board/dice/tabletop       52           49           12        3.200     3.750             3         0.600      5.000\n2026_JanAug    City/colony building       12           11            1        2.600     0.385             0         0.000        NaN\n2026_JanAug              Management       39           39            9        3.733     2.411             0         0.400      0.000\n                   host  variant  b_n  b_ge100  share  ratio100\n          Tower defense extended  136       27  0.195     2.968\n          Tower defense   direct   57        8  0.082     1.666\n            Platforming extended   71       16  0.026     1.560\n            Platforming   direct   21        6  0.008     2.056\n  First-person shooting extended   93       34  0.093     2.085\n  First-person shooting   direct   15        5  0.015     2.054\nTop-down/arena shooting extended  194       37  0.235     2.829\nTop-down/arena shooting   direct   92       19  0.112     2.681\n                 Sports extended   90       28  0.135     2.290\n                 Sports   direct    8        4  0.012     3.566\n                 Horror extended   56       19  0.016     1.379\n                 Horror   direct   15        3  0.004     1.002\n                 Puzzle extended   72       15  0.014     1.573\n                 Puzzle   direct   29        4  0.006     1.059\n    Board/dice/tabletop extended   74       17  0.140     2.476\n    Board/dice/tabletop   direct   52       12  0.098     2.298\n   City/colony building extended   21       14  0.054     1.963\n   City/colony building   direct   12        8  0.031     2.039\n             Management extended  103       41  0.084     1.219\n             Management   direct   32       10  0.026     1.022\n"}

**Checking recent comparisons**

**Checking recent comparisons**

**Checking threshold comparison counts**

**Checking recent comparisons**

**Checking threshold comparison counts**

**Checking recent comparisons**

**Checking threshold comparison counts**

**Preparing genre evidence table**

**Checking recent comparisons**

**Checking threshold comparison counts**

**Preparing genre evidence table**

**Checking recent comparisons**

**Checking threshold comparison counts**

**Preparing genre evidence table**

**Checking classification sensitivity**

Script completed
Wall time 1.5 seconds
Output:

{}
{"chunk_id":"9b055b","wall_time_seconds":1.000928847,"session_id":34745,"original_token_count":0,"output":""}
{"chunk_id":"8a921a","wall_time_seconds":0.176876147,"exit_code":0,"original_token_count":754,"output":"                   host              variant  build_n  build_ge100  supported_n  observed100  expected100  ratio100\n  First-person shooting                  all       57           24           56           23       16.467     1.397\n  First-person shooting              no_core       43           17           41           15       11.000     1.364\n  First-person shooting             no_rogue       37           12           35           10        7.800     1.282\n  First-person shooting no_prior_current1000       48           17           47           16       11.183     1.431\n  First-person shooting      no_prior_credit       38           14           34           10        7.300     1.370\nTop-down/arena shooting                  all      135           27          134           27       14.467     1.866\nTop-down/arena shooting              no_core       31            4           31            4        1.800     2.222\nTop-down/arena shooting             no_rogue       22            5           22            5        1.200     4.167\nTop-down/arena shooting no_prior_current1000      123           18          121           17       10.800     1.574\nTop-down/arena shooting      no_prior_credit       82            8           79            7        5.600     1.250\n                 Sports                  all       17            6           17            6        3.200     1.875\n                 Sports              no_core        9            3            9            3        2.400     1.250\n                 Sports             no_rogue       11            2           11            2        1.800     1.111\n                 Sports no_prior_current1000       15            5           15            5        1.600     3.125\n                 Sports      no_prior_credit        8            3            8            3        1.600     1.875\n    Board/dice/tabletop                  all       61           15           55           10        5.600     1.786\n    Board/dice/tabletop              no_core       21            1           20            1        1.400     0.714\n    Board/dice/tabletop             no_rogue       29            6           25            3        3.200     0.938\n    Board/dice/tabletop no_prior_current1000       54           10           52            9        4.600     1.957\n    Board/dice/tabletop      no_prior_credit       40            9           38            8        1.900     4.211\n   City/colony building                  all       16           10           16           10        5.400     1.852\n   City/colony building              no_core        4            2            4            2        1.200     1.667\n   City/colony building             no_rogue        9            6            9            6        3.200     1.875\n   City/colony building no_prior_current1000       16           10           16           10        5.200     1.923\n   City/colony building      no_prior_credit       11            8           11            8        3.350     2.388\n"}

**Comparing trait evidence**

**Comparing trait evidence**

Script completed
Wall time 5.4 seconds
Output:

{"chunk_id":"4dd4f3","wall_time_seconds":0.194864567,"exit_code":0,"original_token_count":285,"output":" year                    host    n  strict  extended  share\n 2025        Action roguelike  393     133       166  0.338\n 2025             Platforming  648      10        18  0.015\n 2025   First-person shooting  252      17        24  0.067\n 2025 Top-down/arena shooting  192      30        44  0.156\n 2025                  Sports  173       6        30  0.035\n 2025                  Horror  973      11        15  0.011\n 2025                  Puzzle 1274      14        26  0.011\n 2025     Board/dice/tabletop  124      13        17  0.105\n 2025    City/colony building  102       8        10  0.078\n 2026        Action roguelike  675     326       392  0.483\n 2026             Platforming  838      26        38  0.031\n 2026   First-person shooting  322      35        42  0.109\n 2026 Top-down/arena shooting  297      88       106  0.296\n 2026                  Sports  254      18        55  0.071\n 2026                  Horror 1224      21        26  0.017\n 2026                  Puzzle 1503      37        47  0.025\n 2026     Board/dice/tabletop  255      52        68  0.204\n 2026    City/colony building  133      12        15  0.090\n"}
{"chunk_id":"51549a","wall_time_seconds":5.000684073,"session_id":34745,"original_token_count":0,"output":""}

**Planning text-length matching**

**Planning text-length matching**

**Checking text-length sensitivity**

**Planning text-length matching**

**Checking text-length sensitivity**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"d7ebc7","wall_time_seconds":1.00018696,"session_id":85886,"original_token_count":0,"output":""}
{"chunk_id":"4aacf8","wall_time_seconds":0.0000012,"exit_code":0,"original_token_count":152,"output":"                           min    max\nhost                                 \nAction RPG               1.706  1.921\nAction roguelike         1.138  1.237\nAuto battler             1.618  1.829\nBoard/dice/tabletop      1.579  2.264\nCity/colony building     1.843  2.000\nDeckbuilding             1.887  2.207\nFirst-person shooting    1.322  1.397\nHorror                   0.600  0.682\nManagement               0.978  1.128\nPlatforming              0.900  1.071\nPuzzle                   0.955  1.186\nSports                   1.875  2.000\nTop-down/arena shooting  1.866  2.132\nTower defense            1.023  1.264\n"}

**Assessing genre expectations**

**Assessing genre expectations**

**Prioritizing inspection targets**

**Assessing genre expectations**

**Prioritizing inspection targets**

**Assessing genre expectations**

**Prioritizing inspection targets**

**Assessing genre expectations**

**Prioritizing inspection targets**

**Checking sparse-case ratios**

**Assessing genre expectations**

**Prioritizing inspection targets**

**Checking sparse-case ratios**

**Assessing genre expectations**

**Prioritizing inspection targets**

**Checking sparse-case ratios**

**Testing publisher-removal sensitivity**

**Assessing genre expectations**

**Prioritizing inspection targets**

**Checking sparse-case ratios**

**Testing publisher-removal sensitivity**

**Assessing genre expectations**

**Prioritizing inspection targets**

**Checking sparse-case ratios**

**Testing publisher-removal sensitivity**

Script completed
Wall time 0.4 seconds
Output:

{"chunk_id":"2c2f9b","wall_time_seconds":0.000001589,"exit_code":0,"original_token_count":823,"output":"   period                      host  build_n  supported_n  observed100  expected100  ratio100\n2023_2025          Action roguelike      461          429           97       87.333     1.111\n2023_2025              Deckbuilding      589          299           65       28.617     2.271\n2023_2025              Auto battler       88           39           14        5.767     2.428\n2023_2025                Action RPG       71           60           25       13.767     1.816\n2023_2025              Tactical RPG      132          118           37       27.867     1.328\n2023_2025             Tower defense       86           78           15       12.000     1.250\n2023_2025               Platforming       32           32            6        4.800     1.250\n2023_2025              Metroidvania        6            5            3        2.200     1.364\n2023_2025     First-person shooting       57           50           18       15.800     1.139\n2023_2025   Top-down/arena shooting      135          125           24       13.367     1.796\n2023_2025               Shoot-em-up      122          103           16       14.100     1.135\n2023_2025                    Racing        5            3            0        0.533     0.000\n2023_2025        Driving simulation        6            5            1        0.400     2.500\n2023_2025   Flight/space simulation       11            8            1        1.067     0.938\n2023_2025                    Sports       17           17            6        3.933     1.525\n2023_2025                    Rhythm       11            5            1        1.133     0.882\n2023_2025             Stealth/heist        7            2            1        1.083     0.923\n2023_2025                    Horror       25           24            6        8.800     0.682\n2023_2025         Survival/crafting       74           67           14       13.067     1.071\n2023_2025                    Puzzle       45           42            7        4.533     1.544\n2023_2025                   Match-3        7            5            1        0.000       NaN\n2023_2025                   Pinball        1            0            0        0.000       NaN\n2023_2025       Board/dice/tabletop       61           48           10        4.467     2.239\n2023_2025      City/colony building       16           11            7        4.700     1.489\n2023_2025                Automation        8            5            2        1.400     1.429\n2023_2025                Management       49           48           14       13.800     1.014\n2023_2025                   Farming        9            4            2        0.950     2.105\n2023_2025                   Fishing        5            1            0        0.200     0.000\n2023_2025                  Fighting        6            5            4        0.400    10.000\n2023_2025                Beat-em-up        5            0            0        0.000       NaN\n2023_2025        Real-time strategy       24           16            5        3.483     1.435\n2023_2025       Turn-based strategy      168          156           48       32.483     1.478\n2023_2025 Adventure/point-and-click        1            1            0        0.000       NaN\n2023_2025      Job/hobby simulation        5            4            3        2.067     1.452\n"}
{"chunk_id":"898c7c","wall_time_seconds":0.211656259,"exit_code":0,"original_token_count":372,"output":"                   host definition  build_n  supported_n  observed100  expected100  ratio100\nTop-down/arena shooting     strict      135          134           27       14.467     1.866\nTop-down/arena shooting     direct       92           91           19       10.000     1.900\nTop-down/arena shooting   extended      194          191           37       18.867     1.961\n  First-person shooting     strict       57           56           23       16.467     1.397\n  First-person shooting     direct       15           14            4        3.600     1.111\n  First-person shooting   extended       93           92           33       25.667     1.286\n                 Sports     strict       17           17            6        3.200     1.875\n                 Sports     direct        8            8            4        1.200     3.333\n                 Sports   extended       90           89           27       19.300     1.399\n    Board/dice/tabletop     strict       61           55           10        5.600     1.786\n    Board/dice/tabletop     direct       52           48            9        4.600     1.957\n    Board/dice/tabletop   extended       74           68           12        6.350     1.890\n   City/colony building     strict       16           16           10        5.400     1.852\n   City/colony building     direct       12           12            8        3.600     2.222\n   City/colony building   extended       21           20           13        7.000     1.857\n"}

**Reviewing narrower evidence**

**Reviewing narrower evidence**

**Assessing category evidence**

**Reviewing narrower evidence**

**Assessing category evidence**

**Reviewing narrower evidence**

**Assessing category evidence**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"9de0c2","wall_time_seconds":1.001354102,"session_id":27885,"original_token_count":0,"output":""}

**Verifying final-pass fields**

**Verifying final-pass fields**

**Checking comparison denominators**

**Verifying final-pass fields**

**Checking comparison denominators**

**Verifying final-pass fields**

**Checking comparison denominators**

**Validating source audit**

**Verifying final-pass fields**

**Checking comparison denominators**

**Validating source audit**

**Verifying final-pass fields**

**Checking comparison denominators**

**Validating source audit**

**Verifying matched-pair constraints**

Script completed
Wall time 0.3 seconds
Output:

{"chunk_id":"8d6132","wall_time_seconds":0.000001217,"exit_code":0,"original_token_count":676,"output":"FINAL SPARSE PASSES\n                 host  b_n  b_ge100  responding_developers  matched_ratio100\n          Action RPG   71       31                     30             1.830\n              Sports   17        6                      6             1.875\n Board/dice/tabletop   61       15                     14             1.786\nCity/colony building   16       10                     10             1.852\n  Real-time strategy   24        8                      8             1.765\nMATCHED REMOVALS\n                    host          removal                   removed  build_n  build_ge100  supported_n  observed100  expected100  ratio100\n                 Sports     developer_id  name:excellent rectangle       16            5           16            5        2.800     1.786\n                 Sports     publisher_id  name:excellent rectangle       16            5           16            5        2.800     1.786\n                 Sports top3_build_games   1566200|1742360|2995450       14            3           14            3        2.400     1.250\n   City/colony building     developer_id          name:sparkypants       15            9           15            9        4.400     2.045\n   City/colony building     publisher_id      name:klei publishing       15            9           15            9        4.400     2.045\n   City/colony building top3_build_games   1270580|2290000|2346410       13            7           13            7        3.400     2.059\n    Board/dice/tabletop     developer_id        name:ironoak games       60           14           55           10        5.600     1.786\n    Board/dice/tabletop     publisher_id          name:curve games       60           14           55           10        5.600     1.786\n    Board/dice/tabletop top3_build_games   1676840|1689500|3444020       58           12           54            9        5.200     1.731\n  First-person shooting     developer_id           name:sgtokidoki       56           23           55           22       15.867     1.387\n  First-person shooting     publisher_id           name:sgtokidoki       56           23           55           22       15.867     1.387\n  First-person shooting top3_build_games    671860|1272080|2406770       54           21           53           20       14.800     1.351\nTop-down/arena shooting     developer_id name:awesome games studio      134           26          133           26       13.867     1.875\nTop-down/arena shooting     publisher_id name:awesome games studio      134           26          133           26       13.867     1.875\nTop-down/arena shooting top3_build_games   1981820|2163330|2441270      132           24          131           24       13.467     1.782\n"}
{"chunk_id":"82cd9a","wall_time_seconds":0.175044082,"exit_code":0,"original_token_count":1234,"output":"                     host    n  b_n  b_ge100  c_ge100  b_rate100  c_rate100  supported_n  observed100  matched_expected100  matched_ratio100\n         Action roguelike 1696  461      111      229      0.241      0.185          460          110               93.750             1.173\n             Deckbuilding  724  589      160       16      0.272      0.119          496          117               60.700             1.928\n             Auto battler  201   88       31       22      0.352      0.195           78           25               14.250             1.754\n               Action RPG  739   71       31      120      0.437      0.180           67           29               15.850             1.830\n             Tactical RPG  895  132       43      147      0.326      0.193          130           42               33.350             1.259\n            Tower defense  699   86       15       62      0.174      0.101           86           15               13.400             1.119\n              Platforming 2735   32        6      249      0.188      0.092           32            6                6.000             1.000\n             Metroidvania  253    6        3       67      0.500      0.271            6            3                2.600             1.154\n    First-person shooting 1005   57       24      171      0.421      0.180           56           23               16.467             1.397\n  Top-down/arena shooting  825  135       27       63      0.200      0.091          134           27               14.467             1.866\n              Shoot-em-up  651  122       22       67      0.180      0.127          120           21               16.400             1.280\n                   Racing  608    5        2       79      0.400      0.131            4            1                1.600             0.625\n       Driving simulation  743    6        1      154      0.167      0.209            6            1                1.600             0.625\n  Flight/space simulation  410   11        2       49      0.182      0.123           11            2                0.800             2.500\n                   Sports  666   17        6       98      0.353      0.151           17            6                3.200             1.875\n                   Rhythm  238   11        3       42      0.273      0.185           10            3                1.800             1.667\n            Stealth/heist  291    7        6       75      0.857      0.264            3            2                1.200             1.667\n                   Horror 3607   25        6      653      0.240      0.182           25            6               10.000             0.600\n        Survival/crafting 1176   74       17      266      0.230      0.241           73           17               14.800             1.149\n                   Puzzle 5099   45        7      527      0.156      0.104           45            7                6.600             1.061\n                  Match-3  185    7        1        8      0.143      0.045            7            1                0.200             5.000\n                  Pinball   45    1        0        4      0.000      0.091            0            0                0.000               NaN\n      Board/dice/tabletop  530   61       15       48      0.246      0.102           55           10                5.600             1.786\n     City/colony building  386   16       10      130      0.625      0.351           16           10                5.400             1.852\n               Automation  168    8        2       68      0.250      0.425            7            2                1.400             1.429\n               Management 1228   49       15      385      0.306      0.327           49           15               15.000             1.000\n                  Farming  319    9        5      105      0.556      0.339            9            5                2.600             1.923\n                  Fishing  122    5        2       29      0.400      0.248            3            1                1.667             0.600\n                 Fighting  241    6        4       46      0.667      0.196            6            4                1.000             4.000\n               Beat-em-up   93    5        1       17      0.200      0.193            2            0                0.000               NaN\n       Real-time strategy  313   24        8       69      0.333      0.239           23            8                4.533             1.765\n      Turn-based strategy  998  168       53      147      0.315      0.177          166           52               37.850             1.374\nAdventure/point-and-click  857    1        0      149      0.000      0.174            1            0                0.000               NaN\n     Job/hobby simulation  402    5        4      120      0.800      0.302            4            3                1.467             2.045\n"}

**Checking omitted source text**

**Checking omitted source text**

**Checking source wording**

**Checking omitted source text**

**Checking source wording**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"23d777","wall_time_seconds":1.001266155,"session_id":88154,"original_token_count":0,"output":""}

**Auditing group counts**

**Auditing group counts**

**Checking sample coverage**

**Auditing group counts**

**Checking sample coverage**

**Auditing group counts**

**Checking sample coverage**

[
  {
    "type": "text",
    "text": "Script completed\nWall time 1.2 seconds\nOutput:\n"
  },
  {
    "type": "text",
    "text": "{}"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"16d4b5\",\"wall_time_seconds\":1.0017193,\"session_id\":92010,\"original_token_count\":0,\"output\":\"\"}"
  },
  {
    "type": "text",
    "text": "{\"chunk_id\":\"e38cd6\",\"wall_time_seconds\":0.00000144,\"exit_code\":0,\"original_token_count\":7,\"output\":\"Saved two PNG/SVG figures.\\n\"}"
  },
  {
    "type": "image",
    "note": "[image: shown below]"
  },
  {
    "type": "image",
    "note": "[image: shown below]"
  }
]

**Checking validation guard**

**Checking validation guard**

**Checking source validation**

**Checking validation guard**

**Checking source validation**

**Checking validation guard**

**Checking source validation**

**Checking validation guard**

**Checking source validation**

**Checking validation guard**

**Checking source validation**

**Checking validation guard**

**Checking source validation**

**Checking validation guard**

**Checking source validation**

Script completed
Wall time 0.8 seconds
Output:

{"chunk_id":"bfb889","wall_time_seconds":0.000001648,"exit_code":0,"original_token_count":15,"output":"Passed 675 checks; matched records 12844 peer links 62343\n"}
{"chunk_id":"15d605","wall_time_seconds":0.594828254,"exit_code":0,"original_token_count":47,"output":"             games  strict  extended\nperiod                              \n2019_2022    27645     691      1294\n2023_2025    35806    1911      3051\n2026_JanAug  12409    1375      1810\n"}

**Documenting revision caveat**

**Documenting revision caveat**

**Checking cohort comparability**

**Documenting revision caveat**

**Checking cohort comparability**

**Documenting revision caveat**

**Checking cohort comparability**

**Documenting revision caveat**

**Checking cohort comparability**

**Separating missing-price peers**

Script completed
Wall time 0.7 seconds
Output:

{"chunk_id":"9855e1","wall_time_seconds":0.560139103,"exit_code":0,"original_token_count":1035,"output":"                   host   appid                                          game  reviews\n       Action roguelike 1307660                                    Ravenbound      770\n       Action roguelike 1504250                                  Paper Planet      380\n       Action roguelike 1721110                                       Abyssus     4968\n       Action roguelike 2008510                                 Go Mecha Ball      372\n       Action roguelike 2173300                                Skull Survivor        1\n       Action roguelike 2500680                                    Tiny Chaos       43\n       Action roguelike 2870580                           Cubed and Dangerous       23\n       Action roguelike 3636470                        Dream Abyss: Survivors       47\n           Tactical RPG 3317660                      Kaamos: Puzzle Roguelike       73\n  First-person shooting 1721110                                       Abyssus     4968\nTop-down/arena shooting 2008510                                 Go Mecha Ball      372\nTop-down/arena shooting 2500680                                    Tiny Chaos       43\nTop-down/arena shooting 2870580                           Cubed and Dangerous       23\n            Shoot-em-up 2500680                                    Tiny Chaos       43\n            Shoot-em-up 3787880 Pattern Survivors: Bullet Hell — Create Chaos        5\n      Survival/crafting 2173300                                Skull Survivor        1\n      Survival/crafting 2332410                                  Hero of Fate      495\n      Survival/crafting 3636470                        Dream Abyss: Survivors       47\n                 Puzzle 3317660                      Kaamos: Puzzle Roguelike       73\n    Turn-based strategy 3317660                      Kaamos: Puzzle Roguelike       73\nExplore a gorgeous roguelite hockey world where every choice matters on your journey to restore hockey to its former glory. Hire unique players, manage your benchwarmers, discover synergies between them, choose new abilities, bribe the referees, visit the Blademaster for upgrades and a lot more. CRAFT YOUR TEAM Take charge of a ragtag group of hockey players in a roguelite journey across a mysterious land. As the newly appointed captain, every choice matters on your path to discover the secret of hockey's last bastion. Hire superstars, utilize the synergy found between players, and harness ruthless abilities, all in favor of crafting an unbeatable team. LOSING IS ONLY THE BEGINNING Easier said than done, as the ultimate dream team doesn’t appear out of thin air. Use losses as an opportunity to visit the mighty Blademaster in between runs and permanently upgrade your hero for the battles ahead. When the Center is still flubbing every single face-off, there’s no shame in lining the ref’s pockets with some sweet “maple syrup.” DEEP AND REWARDING GAMEPLAY Combining modern precision with fast-paced 90s-inspired action, Tape to Tape offers a unique experience in the form of a simcade hockey experience. Bone-crushing body checks and stellar saves will keep players shaking in their skates, but mastering the dynamic puck physics system will secure the victory. Players and goalies will react to the puck bouncing off sticks and ricocheting off faces creating organic and emergent battles on the ice. Thankfully, there is no penalty when a tooth is lost. KEY FEATURES: A deep Roguelite campaign featuring countless powerups, events, boss battles and unique players Create and customize your teams and players with power-ups, artifacts, and personalized color options Over 40 different teams, each with unique synergies and colorful players Local, couch and Steam Remote Play multiplayer for up to 10 players Innovative goaltenders that react organically to any shot, deflected puck or rebound -goals never feel unfair A unique co-op campaign delivering a roguelite experience like no other Challenge your friends in Tournament, Play now or Shootout modes or play the Campaign cooperatively A unique gorgeous blend of hand-drawn 2D sprites and animations in a 3D environment LOCAL/STEAM RE\n"}

**Recomputing known-price aggregates**

**Recomputing known-price aggregates**

**Recomputing known-price aggregates**

**Checking missing-price handling**

**Recomputing known-price aggregates**

**Checking missing-price handling**

**Recomputing known-price aggregates**

**Checking missing-price handling**

**Recomputing known-price sensitivity**

Script completed
Wall time 0.8 seconds
Output:

{"chunk_id":"d920d0","wall_time_seconds":0.662788538,"exit_code":0,"original_token_count":182,"output":"   period                    host variant  supported_n  observed100  expected100  ratio100\n2023_2025        Action roguelike     all          452          106       92.117     1.151\n2023_2025     Board/dice/tabletop     all           55           10        5.600     1.786\n2023_2025    City/colony building     all           16           10        5.400     1.852\n2023_2025   First-person shooting     all           55           22       15.867     1.387\n2023_2025                  Puzzle     all           44            7        6.400     1.094\n2023_2025                  Sports     all           17            6        3.200     1.875\n2023_2025 Top-down/arena shooting     all          131           26       13.600     1.912\n"}
{}

**Preparing finalize.py code**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"f0d4f2","wall_time_seconds":1.000842765,"session_id":73218,"original_token_count":0,"output":""}

**Checking sample counts**

**Checking sample counts**

**Checking sample counts**

**Checking sample counts**

Script completed
Wall time 1.2 seconds
Output:

{}
{"chunk_id":"2c4ee5","wall_time_seconds":1.001644917,"session_id":67308,"original_token_count":0,"output":""}
{"chunk_id":"49be2a","wall_time_seconds":0.000001396,"exit_code":0,"original_token_count":7,"output":"Saved two PNG/SVG figures.\n"}

**Checking sensitivity claim**

**Checking sensitivity claim**

**Checking sensitivity claim**

**Checking sensitivity claim**

**Checking sensitivity claim**

**Checking audit reproducibility**

Script completed
Wall time 0.0 seconds
Output:

null

Script completed
Wall time 0.0 seconds
Output:

{"chunk_id":"d48589","wall_time_seconds":0.000001281,"exit_code":0,"original_token_count":15,"output":"Passed 675 checks; matched records 12844 peer links 62343\n"}
{}