Fable Test — Steam Market Research 2

a327ex-site\logs\steam-game-recommendations.md
a327ex-site\logs\fable-test-steam-market-research-1.md

No files found

File content (512KB) exceeds maximum allowed size (256KB). Use offset and limit parameters to read specific portions of the file, or search for specific content instead of reading the whole file.

1    Title: Fable Test — Steam Market Research 1
2   Date: 2026-07-07 19:44:20
3   
4   # Fable Test — Steam Market Research 1
5   
6   ## Summary
7   
8   First-principles Steam market research for a327ex, built from primary data only (Steam's own endpoints + SteamSpy bulk), deliberately indifferent to indie-marketing discourse (no howtomarketagame/Reddit/GDC/newsletters read — inputs = raw data + model priors). Greenfield: no prior Steam tooling existed in the repo. Built a resumable crawler + analysis pipeline under `E:/a327ex/steam-market/` (uv venv, SQLite stores, 8 reports in `reports/`). Session ran on Fable, got flagged/downgraded to Opus mid-way twice; ended before running the final 8 requested analyses.
9   
10  **Data infrastructure (`steam-market/`, NOT a git repo — new top-level dir):**
11  - `scripts/steamspy_all.py` — SteamSpy bulk catalog (`api.php?request=all`), 1 page/min. Final: 82,233 games (owners buckets, playtime medians, dev/publisher, CCU).
12  - `scripts/search_scrape.py` — Steam search index (`search/results/?infinite=1`, category1=998 games-only). KEY DISCOVERY: the search HTML embeds release date, top ~7 tag IDs per game, review count + positive %, price/discount, platforms — so tag-space analysis needs ZERO per-game calls. Final: 115,281 unique games (full catalog).
13  - `scripts/histogram_crawl.py` — `appreviewhistogram/<appid>` monthly review rollups since launch (trajectory analysis). Ran to ~24K/34,654 (≥50-review queue).
14  - `scripts/appdetails_crawl.py` — `api/appdetails` raw JSON (descriptions, categories, achievements, languages, DLC, demos). Stratified queue: ALL 5,596 window hits + 4,000 random window-failures + 3,000 window-mids + reviewed tail. Ran to ~15K.
15  - `scripts/reviews_sample_crawl.py` — sub-10-review split + purchase-type (key-farming) split.
16  - `scripts/reviews_lang_crawl.py` — per-language review counts, 728 games × 7 langs.
17  - `scripts/demo_hist_crawl.py` — demo review histograms (~1,485 demos) for Next Fest detection.
18  - All crawlers checkpoint per-row in SQLite `done`/`meta` tables → fully resumable by re-running the script. `scripts/common.py` = shared catalog loader (needs `sys.stdout.reconfigure(encoding='utf-8')` for Windows cp1252).
19  
20  **Revenue estimator (in-house, discourse-free), `scripts/calibrate.py`:** Calibrated ONLY on a327ex's own two games from `posts/snkrx_log.md` daily tables. Result: `revenue ≈ reviews × 35.9 × base_price` (band K=25–55). Satisfying cross-check: units/review ≈ 45 for BOTH SNKRX (day-55) and BYTEPATH (lifetime), 3 years apart. Steam's search review counts are score-ELIGIBLE (exclude key activations) → outcome analyses are farming-resistant (verified: SNKRX search 3,618 < purchase 4,194 < all 4,333). SteamSpy owner buckets UNDERCOUNT known truth (SNKRX bucket 50–100K vs 80K+ sold by day 55) → demoted to weak-check only.
21  
22  **Report 01 — delta map (base rates):** Volume DOUBLED (9.4K releases 2021 → 19.1K 2025, ~24K/yr 2026 pace) but median reviewed game held flat ~$13K since 2019; growth went into failure mass (≤9-review share 46%→55%). Paid-release hit rates eroded mildly: ≥$50K 16.0%(2021)→13.5%(2025). F2P flat ~15-18%.
23  
24  **Report 02 — tag scan + `scripts/drill_tags.py`:** Market baseline P(≥$50K)=14.3%. His 2021 tag neighborhood is bottom-decile (Arcade 5.5%, Minimalist 5.4%, Abstract 5.1%, Score Attack 3.7%, Precision Platformer 4.5%). Roguelite×co-op is top-decile with TINY supply: Online Co-Op+Roguelite 54.9% (33% at ≤$10), +Action Roguelike 50%, Roguelite+Loot 50% (42% at ≤$10), n=34-66 over 3 years vs 4,428 arcade games. Gold rushes decaying on schedule (Desktop Companion 47→9%, Shop Keeper 55→16%, Boomer Shooter 71→14%) EXCEPT Roguelike Deckbuilder (durable 24-46% across 3yr, Balatro effect). PRICE GATE: most tags' P(≥$50K) halves below $10; tags that HOLD at ≤$10 = FMV, Bullet Heaven, Idler, Auto Battler, Horror.
25  
26  **Report 03 — trajectories:** Classified 17,955 review-histogram curves. Launch-spike share rose 2%(2014)→17%(2023); median m12/m3 ≈ 1.74×. CHEAP GAMES HAVE LONGEST TAILS ($3-5: 2.26× vs $25+: 1.64×) → minimal-marketing long-tail model structurally favored at his price. Long-tail tags: Classic, Local Co-Op, Moddable, Level Editor, Physics.
27  
28  **Report 04 — anomalies (resource-normalized) + COMPOUNDING (the biggest finding):** Prior-hit devs (≥$50K in 2020-22) re-hit ≥$50K on 57.6% of 2023-25 releases (median $104K); veterans WITHOUT a prior hit: 6.4% (median $6K); first-time self-pub: 8.6%. **9× gap.** a327ex is in the 57.6% bucket (SNKRX + 2K+ followers). Advantage is PORTABLE across niche-switching (52.7% switchers vs 61.0% stayers). Still, 25% of ALL ≥$50K games are first-time self-pubs (market open). Outsider-viable tags (fs/all≈1.0): FMV, Cozy, Romance, Cooking. Publisher-dominated: Colony Sim 0.42, Roguelike Deckbuilder 0.47.
29  
30  **Report 05 — art gate (blind capsule scoring):** Scored 355 header images blind (4 niches × fail/mid/hit, tiers 1-5, no outcome knowledge). P(hit|art≥4)=74% vs P(hit|art≤2)=14%. Art is the strongest per-game gate — EXCEPT Idler (P(hit|art≤2)=27%, its hits mostly DON'T have good art: Nodebuster $2.99/$1.1M, Gnorp $6.99/$1.8M, Digseum $2.99/$676K, Fill Up The Hole art-tier-1/$121K). Horror tolerates ugly art at the top end (streamer-driven: The Complex/Eclipsium/MIMESIS $700-850K). This is the key asymmetry between his two candidate projects.
... [90 more lines]

FullName                                               Length    LastWriteTime    

--------                                               ------    -------------    

E:\a327ex\steam-market\.venv                                     07/07/26 09:17:29

E:\a327ex\steam-market\data                                      07/07/26 18:45:06

E:\a327ex\steam-market\reports                                   07/07/26 16:37:27

E:\a327ex\steam-market\scripts                                   07/07/26 18:44:45

E:\a327ex\steam-market\README.md                       1570      07/07/26 09:18:43

E:\a327ex\steam-market\.venv\Lib                                 07/07/26 09:17:29

E:\a327ex\steam-market\.venv\Scripts                             07/07/26 09:17:36

E:\a327ex\steam-market\.venv\.gitignore                1         07/07/26 09:17:28

E:\a327ex\steam-market\.venv\.lock                     0         07/07/26 09:17:29

E:\a327ex\steam-market\.venv\CACHEDIR.TAG              43        07/07/26 09:17:28

E:\a327ex\steam-market\.venv\pyvenv.cfg                174       07/07/26 09:17:29

E:\a327ex\steam-market\data\art                                  07/07/26 10:58:58

E:\a327ex\steam-market\data\calibration                          07/07/26 09:28:14

E:\a327ex\steam-market\data\appdetails.sqlite          365674496 07/07/26 19:51:54

E:\a327ex\steam-market\data\appdetails.sqlite-shm      32768     07/07/26 10:42:24

E:\a327ex\steam-market\data\appdetails.sqlite-wal      4214792   07/07/26 19:53:08

E:\a327ex\steam-market\data\demo_hist.sqlite           421888    07/07/26 18:44:05

E:\a327ex\steam-market\data\histogram.sqlite           75808768  07/07/26 19:51:30

E:\a327ex\steam-market\data\histogram.sqlite-shm       32768     07/07/26 10:42:22

E:\a327ex\steam-market\data\histogram.sqlite-wal       4165352   07/07/26 19:53:09

E:\a327ex\steam-market\data\reviews_lang.sqlite        274432    07/07/26 13:52:00

E:\a327ex\steam-market\data\reviews_sample.sqlite      86016     07/07/26 11:37:18

E:\a327ex\steam-market\data\search.sqlite              13430784  07/07/26 10:41:26

E:\a327ex\steam-market\data\search.sqlite-shm          32768     07/07/26 10:42:23

E:\a327ex\steam-market\data\search.sqlite-wal          0         07/07/26 10:42:23

E:\a327ex\steam-market\data\steamspy.sqlite            9388032   07/07/26 10:45:47

E:\a327ex\steam-market\data\steamspy.sqlite-shm        32768     07/07/26 09:18:26

E:\a327ex\steam-market\data\steamspy.sqlite-wal        7774472   07/07/26 10:46:49

... [32 more lines]

Exit code 1
  File "<string>", line 5

    tables = [r[0] for r in db.execute(select

                                      ^

SyntaxError: '(' was never closed

Command running in background with ID: bgcw2way1. Output is being written to: C:\Users\adn\AppData\Local\Temp\claude\E--a327ex\b466f190-f23f-479c-9b8f-a811de0a1c64\tasks\bgcw2way1.output. You will be notified when it completes. To check interim output, use Read on that file path.

1    # Working notes (accumulating findings for synthesis)
2   
3   ## Data integrity status
4   - Estimator: revenue ≈ reviews × 35.9 × price, band [25,55]. In-house (SNKRX day55 + BYTEPATH).
5     Units/review agrees across both games (~45).
6   - Search index (115,281 games) ≈ full catalog: per-year volume matches known history
7     (2021 ≈ 9.4K, 2024 ≈ 16.2K games) → no heavy shovelware filtering. Survivorship: delisted absent.
8   - Sub-10 split (paid, sampled n=1500): 15% have exactly 0 reviews, 44% have 1-4, 27% have 5-9,
9     14% cross ≥10 on API-all count.
10  - KEY-FARMING CHECK RESOLVED: search tooltip counts = score-ELIGIBLE reviews (excludes key
11    activations + more; SNKRX: search 3,618 < purchase 4,194 < all 4,333; one crossover game:
12    591 total / <10 eligible). Outcome analyses are already farming-resistant. The 14.3% ≥$50K
13    baseline stands (if anything conservative). Purchase-mode crawl = confirmation only.
14  
15  ## Delta map headlines (report 01, done)
16  - Volume DOUBLED: 2021 9.4K → 2025 19.1K releases; 2026 tracking ~24K/yr.
17  - Median reviewed game ≈ $13K est. — FLAT since 2019 despite volume. Failure mass grew instead:
18    ≤9-review share 46% (2021) → 55% (2025).
19  - Hit rates by cohort (paid): ≥$50K 16.0% (2021) → 13.5% (2025); ≥$100K 12.4% → 10.1%.
20    Deterioration is real but modest; volume doubling is the big change.
21  - F2P share flat ~15-18%. Median price of reviewed games $8.99.
22  
23  ## Tag scan headlines (report 02, done)
24  - HIS 2021 NEIGHBORHOOD COLLAPSED: Arcade 5.5%, Minimalist 5.4%, Abstract 5.1%, Score Attack 3.7%,
25    2D 6.9%, Puzzle 7.8%, Arena Shooter 7.7%, Precision Platformer 4.5% ≥$50K. Bottom of market.
26  - Roguelite CORE is midfield: Action Roguelike 14.4% (= baseline), n=2,347 (huge supply),
27    and only 7.4% at ≤$10.
28  - Roguelite × MULTIPLAYER/co-op = top decile: Online Co-Op+Roguelite 54.9% (33% at ≤$10),
29    Online Co-Op+Action Roguelike 50%, Co-op+Roguelite 47.2%, Roguelite+Loot 50% (42% at ≤$10, n=26).
30    Supply tiny (n=34-66 over 3 years) vs 4.4K arcade games.
... [92 more lines]

Wasted call — file unchanged since your last Read. Refer to that earlier tool_result instead.

49:[Omitted long matching line]
3091:[Omitted long matching line]

**6. Whitespace map.** Which tag combinations barely exist yet but sit adjacent to clusters of winners — literal unexplored territory near proven demand. Data ready now.

**7. Genre migration graph.** Where do successful devs go next, as a transition map — which genre is the best "launchpad" into others, and whether switching genres costs or keeps the compounding edge (I have a hint: switchers re-hit 52.7% vs 61% for stayers, but not the full graph). Data ready now.

## C. Your own theses, made testable

**★ 8. "Small long games," measured.** Your signature thesis — short to make, long to play — is directly testable now: SteamSpy median playtime vs price vs review positivity vs tail length. Do high-playtime-per-dollar games actually get better reviews and longer tails? This either validates the belief you've built a career on or complicates it, and I have the data to do it (playtime coverage is thin on recent games but fine for 2019–2023 cohorts). 

**9. Update cadence effect.** You do weekly updates and measured ~20–25% bumps yourself. I can detect post-launch review re-spikes (sawtooth) across the catalog and test whether update frequency correlates with lifetime revenue — is the weekly grind worth it, market-wide? Data ready now.

## D. Lighter / exploratory

**10. Description mechanic-mining.** Beyond the numeric claims, cluster the *language* of winners' store descriptions vs failures' — which mechanic words and hooks over-index in hits, and whether new mechanic vocabulary is emerging. Data ready now (raw descriptions crawled).

**11. Title linguistics.** Do "Simulator", colons, numbers, length, non-English titles correlate with outcomes? Cheap, fast, probably minor, occasionally surprising. Data ready now.

**12. Review-text sentiment for build games.** What players actually praise/complain about in deep-build games — what makes them sticky vs bounce. This needs a new review-text crawl (a few hours), so it's the one real cost on the list.

---

My instinct: the four stars form a coherent set — **#1 tells you sequel-or-not, #2 tells you the price, #5 tells you the shell to wear, #8 checks the belief underneath all of it.** That's arguably the highest-insight quartet and all runs on data in hand. #5 (gold-rush detection) is the one I'd flag as most *fun* independent of your immediate decision — it's the closest thing to a market crystal ball the data permits, and it's the kind of thing your old data-scripting habit would have loved.

What sounds worth doing? Pick any subset, or point me at an angle I didn't list.

> <local-command-caveat>Caveat: The messages below were generated by the user while running local commands. DO NOT respond to these messages or otherwise consider them in your response unless the user explicitly asks you to.</local-command-caveat>

```
<command-name>/model</command-name>
            <command-message>model</command-message>
            <command-args>claude-fable-5</command-args>
... [61 more lines]

1    """Shared catalog loader for analysis scripts."""
2   import html, json, re, sqlite3, sys
3   from datetime import date
4   
5   sys.stdout.reconfigure(encoding='utf-8')
6   
7   ROOT = 'E:/a327ex/steam-market'
8   K = 35.9          # revenue ~= reviews * K * base_price_usd (data/calibration/estimator.json)
9   K_BAND = (25, 55)
10  TODAY = date.today()
11  
12  MONTHS = {m: i+1 for i, m in enumerate(
13      ['Jan', 'Feb', 'Mar', 'Apr', 'May', 'Jun', 'Jul', 'Aug', 'Sep', 'Oct', 'Nov', 'Dec'])}
14  
15  FULL_RE = re.compile(r'^([A-Z][a-z]{2}) (\d{1,2}), (\d{4})$')
16  MONTH_RE = re.compile(r'^([A-Z][a-z]{2}) (\d{4})$')
17  YEAR_RE = re.compile(r'^(\d{4})$')
18  
19  def parse_release(txt):
20      """-> (year, date_or_None). (None, None) if unparseable/unreleased."""
21      if not txt:
22          return None, None
23      t = txt.strip()
24      m = FULL_RE.match(t)
25      if m and m.group(1) in MONTHS:
26          try:
27              return int(m.group(3)), date(int(m.group(3)), MONTHS[m.group(1)], int(m.group(2)))
28          except ValueError:
29              return int(m.group(3)), None
30      m = MONTH_RE.match(t)
... [63 more lines]

1    """Crawl appreviewhistogram (monthly review rollups since launch) for every game
2   with >= MIN_REVIEWS reviews, into data/histogram.sqlite. Queue built from search.sqlite,
3   highest review counts first. Resumable: done table. Usage: histogram_crawl.py [min_reviews]"""
4   import sqlite3, sys, time
5   import requests
6   
7   DB = 'E:/a327ex/steam-market/data/histogram.sqlite'
8   SEARCH_DB = 'E:/a327ex/steam-market/data/search.sqlite'
9   MIN_REVIEWS = int(sys.argv[1]) if len(sys.argv) > 1 else 20
10  
11  con = sqlite3.connect(DB)
12  con.execute('PRAGMA journal_mode=WAL')
13  con.execute('''CREATE TABLE IF NOT EXISTS rollup(
14      appid INTEGER, ts INTEGER, up INTEGER, down INTEGER, PRIMARY KEY(appid, ts))''')
15  con.execute('''CREATE TABLE IF NOT EXISTS done(
16      appid INTEGER PRIMARY KEY, start_date INTEGER, end_date INTEGER,
17      rollup_type TEXT, n_rollups INTEGER, fetched_at INTEGER)''')
18  con.execute('ATTACH ? AS s', (SEARCH_DB,))
19  queue = [r[0] for r in con.execute('''
20      SELECT appid FROM s.search
21      WHERE review_count >= ? AND appid NOT IN (SELECT appid FROM done)
22      ORDER BY review_count DESC''', (MIN_REVIEWS,))]
23  print(f'queue: {len(queue)} appids (min_reviews={MIN_REVIEWS})', flush=True)
24  
25  s = requests.Session()
26  s.headers.update({'User-Agent': 'Mozilla/5.0 (steam-market-research; [email protected])'})
27  fails = 0
28  for i, appid in enumerate(queue):
29      try:
30          r = s.get(f'https://store.steampowered.com/appreviewhistogram/{appid}',
... [24 more lines]

1    """Crawl store.steampowered.com/api/appdetails for a prioritized queue into
2   data/appdetails.sqlite. Unofficial rate limit ~200/5min -> 1 req/1.6s sustained.
3   Priority: reviewed games by review count desc, then a random sample of low/no-review
4   games for base rates. Resumable: appdetails + failed tables. Stores raw JSON too."""
5   import json, random, sqlite3, sys, time
6   import requests
7   
8   DB = 'E:/a327ex/steam-market/data/appdetails.sqlite'
9   SEARCH_DB = 'E:/a327ex/steam-market/data/search.sqlite'
10  LOWREV_SAMPLE = 6000
11  
12  con = sqlite3.connect(DB)
13  con.execute('PRAGMA journal_mode=WAL')
14  con.execute('''CREATE TABLE IF NOT EXISTS appdetails(
15      appid INTEGER PRIMARY KEY, type TEXT, name TEXT, is_free INTEGER, required_age TEXT,
16      developers TEXT, publishers TEXT, genre_ids TEXT, genres TEXT,
17      category_ids TEXT, categories TEXT, supported_languages TEXT,
18      release_date TEXT, coming_soon INTEGER, price_cents INTEGER, currency TEXT,
19      recommendations INTEGER, achievements INTEGER, dlc_count INTEGER,
20      demo_appids TEXT, controller_support TEXT, metacritic INTEGER,
21      header_image TEXT, fetched_at INTEGER, raw TEXT)''')
22  con.execute('CREATE TABLE IF NOT EXISTS failed(appid INTEGER PRIMARY KEY, reason TEXT, fetched_at INTEGER)')
23  con.execute('ATTACH ? AS s', (SEARCH_DB,))
24  
25  import sys
26  sys.path.insert(0, 'E:/a327ex/steam-market/scripts')
27  from common import load_catalog
28  
29  done = {r[0] for r in con.execute('SELECT appid FROM appdetails UNION SELECT appid FROM failed')}
30  cat = [g for g in load_catalog() if g['appid'] not in done]
... [79 more lines]

Command running in background with ID: b5iamgfno. Output is being written to: C:\Users\adn\AppData\Local\Temp\claude\E--a327ex\b466f190-f23f-479c-9b8f-a811de0a1c64\tasks\b5iamgfno.output. You will be notified when it completes. To check interim output, use Read on that file path.

Command running in background with ID: bhbsxf9mq. Output is being written to: C:\Users\adn\AppData\Local\Temp\claude\E--a327ex\b466f190-f23f-479c-9b8f-a811de0a1c64\tasks\bhbsxf9mq.output. You will be notified when it completes. To check interim output, use Read on that file path.

1    """Build-depth tag universe: hit rates, low-price behavior, supply trend, outsider
2   viability, and long-tail (m12/m3) per build-depth tag + key pairs. Chat-oriented output."""
3   import sqlite3, statistics
4   from collections import defaultdict
5   from itertools import combinations
6   
7   from common import ROOT, load_catalog, load_tagmap
8   
9   BUILD_TAGS = ['Roguelike', 'Roguelite', 'Action Roguelike', 'Roguelike Deckbuilder',
10                'Traditional Roguelike', 'Deckbuilding', 'Card Battler', 'Card Game',
11                'Auto Battler', 'Bullet Heaven', 'Loot', 'Looter Shooter', 'Hack and Slash',
12                'Action RPG', 'CRPG', 'JRPG', 'Party-Based RPG', 'Strategy RPG',
13                'Tactical RPG', 'Turn-Based Tactics', 'Dungeon Crawler', 'Class-Based',
14                'Character Customization', 'Creature Collector', 'Perma Death',
15                'Procedural Generation', 'Replay Value', 'RPG', 'Turn-Based Strategy']
16  CONTEXT_TAGS = ['Co-op', 'Online Co-Op', 'Multiplayer', 'Idler', 'Incremental', 'Horror',
17                  'Turn-Based', 'Difficult', 'Fantasy', 'Sci-fi', 'Pixel Graphics', 'Cute']
18  
19  def m12_over_m3(rolls):
20      if len(rolls) < 12:
21          return None
22      c3 = sum(n for _, n in rolls[:3]) or 1
23      return sum(n for _, n in rolls[:12])/c3
24  
25  def main():
26      tagmap = load_tagmap()
27      name2id = {v: k for k, v in tagmap.items()}
28      missing = [t for t in BUILD_TAGS if t not in name2id]
29      if missing:
30          print('missing tag names:', missing)
... [106 more lines]

1    """Targeted appreviews crawl, two purposes:
2   (a) sub10 mode: random sample of paid games with <10 reviews (search shows nothing there)
3       -> exact 0 vs 1-9 split for base rates.
4   (b) purchase mode: top games by est. revenue in the 2023-2025 window -> steam-purchase vs
5       key-activation review split (detects key-farming inflation in the outcome data).
6   Usage: reviews_sample_crawl.py sub10 1500 | reviews_sample_crawl.py purchase 800"""
7   import random, sqlite3, sys, time
8   import requests
9   
10  ROOT = 'E:/a327ex/steam-market'
11  DB = f'{ROOT}/data/reviews_sample.sqlite'
12  mode = sys.argv[1] if len(sys.argv) > 1 else 'sub10'
13  N = int(sys.argv[2]) if len(sys.argv) > 2 else 1500
14  
15  con = sqlite3.connect(DB)
16  con.execute('PRAGMA journal_mode=WAL')
17  con.execute('''CREATE TABLE IF NOT EXISTS summary(
18      appid INTEGER PRIMARY KEY, mode TEXT, total_all INTEGER, positive INTEGER,
19      negative INTEGER, steam_total INTEGER, nonsteam_total INTEGER, fetched_at INTEGER)''')
20  con.execute('ATTACH ? AS s', (f'{ROOT}/data/search.sqlite',))
21  
22  if mode == 'sub10':
23      pool = [r[0] for r in con.execute('''SELECT appid FROM s.search
24          WHERE review_count IS NULL AND price_orig_cents > 0
25          AND appid NOT IN (SELECT appid FROM summary)''')]
26      random.seed(7)
27      random.shuffle(pool)
28      queue = pool[:N]
29  else:
30      from common import load_catalog
... [44 more lines]

File created successfully at: E:\a327ex\steam-market\scripts\reviews_text_crawl.py (file state is current in your context — no need to Read it back)

1    histogram {'rollup': 1907799, 'done': 29067}
2   appdetails {'appdetails': 17706, 'failed': 0}
3   demo_hist {'rollup': 10356, 'done': 1486}
4   reviews_lang {'lang': 5095}
5   reviews_sample {'summary': 2300}
6   search {'search': 115281, 'meta': 1}
7   steamspy {'steamspy': 82233, 'meta': 1}
8   

1    queue: 5495 appids (min_reviews=50)
2   

1    queue: 0 window-successes + 4000 window-failures + 3000 window-mid + 50694 reviewed rest
2   

Command running in background with ID: bkrhggaew. Output is being written to: C:\Users\adn\AppData\Local\Temp\claude\E--a327ex\b466f190-f23f-479c-9b8f-a811de0a1c64\tasks\bkrhggaew.output. You will be notified when it completes. To check interim output, use Read on that file path.

1    queue: 3889 (hit=2289, mid sampled 800, low sampled 800; 0 already done)
2   

File created successfully at: E:\a327ex\steam-market\scripts\goldrush.py (file state is current in your context — no need to Read it back)

loading catalog...

392 tags with >=40 games

loading rollups...

rollups for 29404 games



=== CONCENTRATED TRIGGERS (flow>=1500/Q, x>=4.0, top1>=35%): 311 ===

2015Q1 Moddable                     flow=   8.2K x 6.2 top1= 69% trigger=Besiege ($12467K est)

        h50 pre=-(n=5) window=33%(n=9) post-flood=-(n=7) | supply flood: NONE by 2017Q1

2015Q1 Aliens                       flow=   2.3K x 5.8 top1= 45% trigger=Alien: Isolation ($38141K est)

        h50 pre=-(n=5) window=44%(n=9) post-flood=14%(n=14) | supply flood: NONE by 2017Q1

2015Q1 Precision Platformer         flow=   2.5K x12.0 top1= 83% trigger=Geometry Dash ($51446K est)

        h50 pre=-(n=7) window=33%(n=18) post-flood=38%(n=16) | supply flood: NONE by 2017Q1

2015Q1 Parkour                      flow=  10.8K x 9.3 top1= 59% trigger=Dying Light ($87758K est)

        h50 pre=75%(n=8) window=80%(n=10) post-flood=64%(n=11) | supply flood: NONE by 2017Q1

2015Q1 Medieval                     flow=  11.1K x 4.7 top1= 51% trigger=Besiege ($12467K est)

        h50 pre=-(n=5) window=55%(n=20) post-flood=53%(n=19) | supply flood: NONE by 2017Q1

2015Q1 3D                           flow=   2.4K x39.2 top1= 47% trigger=DEAD OR ALIVE 5 Last Round: Core Fighters ($5168K est)

        h50 pre=-(n=3) window=27%(n=11) post-flood=22%(n=23) | supply flood: NONE by 2017Q1

2015Q1 Episodic                     flow=   9.5K x 7.0 top1= 79% trigger=Life is Strange - Episode 1 ($8188K est)

        h50 pre=-(n=1) window=-(n=4) post-flood=-(n=3) | supply flood: NONE by 2017Q1

2015Q1 Futuristic                   flow=   2.1K x41.4 top1= 84% trigger=Call of Duty®: Advanced Warfare - Gold Edition ($12599K est)

        h50 pre=-(n=3) window=10%(n=10) post-flood=25%(n=12) | supply flood: NONE by 2017Q1

2015Q1 Turn-Based Combat            flow=   4.3K x14.3 top1= 75% trigger=Darkest Dungeon® ($50464K est)

        h50 pre=44%(n=9) window=56%(n=18) post-flood=62%(n=26) | supply flood: 2015Q4 (+3Q, 12/Q vs pre 2/Q)

2015Q1 City Builder                 flow=   9.8K x 5.6 top1= 54% trigger=Cities: Skylines ($107702K est)

        h50 pre=75%(n=8) window=71%(n=17) post-flood=42%(n=26) | supply flood: NONE by 2017Q1

2015Q1 Violent                      flow=   4.5K x89.1 top1= 73% trigger=Hotline Miami 2: Wrong Number ($13167K est)

        h50 pre=44%(n=16) window=28%(n=29) post-flood=21%(n=120) | supply flood: 2016Q3 (+6Q, 17/Q vs pre 4/Q)

2015Q1 Replay Value                 flow=   3.8K x75.9 top1= 89% trigger=The Binding of Isaac: Rebirth ($81358K est)

        h50 pre=-(n=7) window=-(n=6) post-flood=32%(n=19) | supply flood: NONE by 2017Q1

... [170 more lines]

705 reports/raw_goldrush_v1.txt
6:=== CONCENTRATED TRIGGERS (flow>=1500/Q, x>=4.0, top1>=35%): 311 ===
630:=== DIFFUSE WARMING (x>=3.0, top1<35%): 43 ===
675:=== ANATOMY (mature concentrated triggers, n=292) ===
679:=== OPEN WINDOWS NOW (trigger in last 4 quarters, supply not yet flooded) ===

400          h50 pre=46%(n=28) window=42%(n=24) post-flood=33%(n=21) | supply flood: NONE by 2022Q3
401 2020Q3 Supernatural                 flow=  11.1K x12.0 top1= 43% trigger=Phasmophobia ($256238K est)
402         h50 pre=-(n=7) window=36%(n=39) post-flood=16%(n=56) | supply flood: 2021Q2 (+3Q, 13/Q vs pre 2/Q)
403 2020Q3 Villain Protagonist          flow=  14.4K x 4.6 top1= 47% trigger=CARRION ($12092K est)
404         h50 pre=-(n=4) window=54%(n=13) post-flood=50%(n=8) | supply flood: NONE by 2022Q3
405 2020Q3 Experimental                 flow=   7.8K x 8.6 top1= 49% trigger=Townscaper ($2273K est)
406         h50 pre=12%(n=40) window=3%(n=176) post-flood=3%(n=38) | supply flood: 2020Q4 (+1Q, 115/Q vs pre 11/Q)
407 2020Q3 Skating                      flow=   3.4K x 5.6 top1= 99% trigger=Skater XL - The Ultimate Skateboarding Game ($18284K est)
408         h50 pre=-(n=1) window=-(n=7) post-flood=-(n=7) | supply flood: NONE by 2022Q3
409 2020Q3 Automation                   flow=  28.2K x 4.0 top1= 43% trigger=Satisfactory ($193969K est)
410         h50 pre=77%(n=13) window=13%(n=31) post-flood=33%(n=51) | supply flood: 2021Q4 (+5Q, 16/Q vs pre 4/Q)
411 2020Q4 Utilities                    flow=   2.2K x11.3 top1= 75% trigger=Virtual Cottage ($0K est)
412         h50 pre=-(n=4) window=0%(n=11) post-flood=9%(n=22) | supply flood: 2022Q3 (+7Q, 13/Q vs pre 1/Q)
413 2020Q4 Pirates                      flow=  42.4K x 4.4 top1= 72% trigger=Sea of Thieves: 2026 Edition ($210280K est)
414         h50 pre=29%(n=21) window=14%(n=14) post-flood=12%(n=17) | supply flood: NONE by 2022Q4
415 2020Q4 Surreal                      flow=  13.3K x 4.1 top1= 36% trigger=Superliminal ($10205K est)
416         h50 pre=23%(n=52) window=18%(n=82) post-flood=18%(n=102) | supply flood: 2021Q1 (+1Q, 32/Q vs pre 15/Q)
417 2020Q4 Psychological Horror         flow= 233.2K x 5.8 top1= 67% trigger=Phasmophobia ($256238K est)
418         h50 pre=31%(n=139) window=28%(n=222) post-flood=19%(n=319) | supply flood: 2021Q4 (+4Q, 97/Q vs pre 44/Q)
419 2020Q4 Voxel                        flow=  16.3K x 4.1 top1= 72% trigger=Teardown ($80616K est)
420         h50 pre=9%(n=23) window=26%(n=27) post-flood=23%(n=30) | supply flood: 2022Q4 (+8Q, 14/Q vs pre 6/Q)
421 2020Q4 Perma Death                  flow=  11.4K x13.4 top1= 61% trigger=Spelunky 2 ($10102K est)
422         h50 pre=14%(n=43) window=10%(n=89) post-flood=15%(n=78) | supply flood: 2021Q1 (+1Q, 23/Q vs pre 10/Q)
423 2020Q4 Cyberpunk                    flow= 323.3K x25.9 top1= 90% trigger=Cyberpunk 2077 ($818696K est)
424         h50 pre=44%(n=43) window=25%(n=60) post-flood=22%(n=67) | supply flood: 2021Q1 (+1Q, 28/Q vs pre 12/Q)
425 2020Q4 Horses                       flow=   9.2K x99.0 top1= 98% trigger=Red Dead Online ($21851K est)
426         h50 pre=22%(n=9) window=12%(n=8) post-flood=12%(n=8) | supply flood: NONE by 2022Q4
427 2020Q4 Nudity                       flow= 343.4K x10.4 top1= 85% trigger=Cyberpunk 2077 ($818696K est)
428         h50 pre=38%(n=177) window=46%(n=100) post-flood=34%(n=101) | supply flood: NONE by 2022Q4
429 2020Q4 VR                           flow= 245.2K x 5.7 top1= 64% trigger=Phasmophobia ($256238K est)
... [275 more lines]

The file E:\a327ex\steam-market\scripts\goldrush.py has been updated successfully. (file state is current in your context — no need to Read it back)

125 reports/raw_goldrush_v2.txt

1    loading catalog...
2   392 tags with >=40 games
3   loading rollups...
4   rollups for 29571 games
5   
6   === ANATOMY BY TRIGGER CLASS (mature triggers, >=6Q old, pre&win n>=8) ===
7   indie    n=53  window-vs-pre median -3pt | postflood-vs-pre -5pt (n=52) | win med 23% pre med 22%
8   big      n=67  window-vs-pre median -3pt | postflood-vs-pre -5pt (n=66) | win med 20% pre med 26%
9   f2p      n=35  window-vs-pre median -3pt | postflood-vs-pre -3pt (n=34) | win med 22% pre med 25%
10  old      n=12  window-vs-pre median -5pt | postflood-vs-pre -7pt (n=12) | win med 17% pre med 23%
11  
12  --- indie-trigger rushes, individually (window-vs-pre, sorted) ---
13  2023Q4 FMV                        pre  28% -> win  55% -> post  60% | flow=33K trigger=Love Is All Around ($10, $0.5M)
14  2018Q2 Real-Time                  pre   8% -> win  33% -> post   6% | flow=2K trigger=Ancestors Legacy ($23, $1.8M)
15  2021Q3 Party                      pre  18% -> win  42% -> post   8% | flow=19K trigger=Draw & Guess ($5, $0.1M)
16  2023Q4 Underground                pre   0% -> win  23% -> post  13% | flow=5K trigger=The Exit 8 ($4, $0.7M)
17  2017Q3 Magic                      pre  22% -> win  43% -> post  21% | flow=3K trigger=Citadel: Forged with Fire ($15, $1.6M)
18  2021Q2 Narrative                  pre  33% -> win  53% -> post  29% | flow=4K trigger=Before Your Eyes ($10, $5.5M)
19  2018Q1 Old School                 pre   0% -> win  17% -> post  19% | flow=2K trigger=DUSK ($20, $12.1M)
20  2019Q1 Thriller                   pre  15% -> win  29% -> post  32% | flow=3K trigger=Unheard - Voices of Crime ($7, $0.7M)
21  2024Q1 Trading                    pre  17% -> win  31% -> post  36% | flow=22K trigger=Supermarket Simulator ($20, $24.3M)
22  2015Q1 Turn-Based Combat          pre  44% -> win  56% -> post  62% | flow=4K trigger=Darkest Dungeon® ($25, $50.5M)
23  2015Q4 Exploration                pre  48% -> win  59% -> post  43% | flow=35K trigger=Fallout 4 ($20, $130.8M)
24  2017Q3 Cartoon                    pre  18% -> win  25% -> post  22% | flow=3K trigger=Cuphead ($20, $46.4M)
25  2017Q4 4 Player Local             pre  11% -> win  17% -> post  18% | flow=12K trigger=Stick Fight: The Game ($5, $9.4M)
26  2022Q1 Bullet Hell                pre   8% -> win  14% -> post  10% | flow=81K trigger=Vampire Survivors ($5, $22.8M)
27  2016Q1 Colorful                   pre  19% -> win  24% -> post  29% | flow=6K trigger=Slime Rancher ($20, $46.8M)
28  2015Q1 Parkour                    pre  75% -> win  80% -> post  64% | flow=11K trigger=Dying Light ($25, $87.8M)
29  2022Q1 Match 3                    pre   0% -> win   4% -> post   2% | flow=30K trigger=Mirror 2: Project X ($2, $0.2M)
30  2019Q4 Cartoony                   pre   5% -> win   8% -> post  12% | flow=4K trigger=Secret Neighbor: Hello Neighbor Multiplayer ($20, $3.5M)
... [96 more lines]

The file E:\a327ex\steam-market\scripts\goldrush.py has been updated successfully. (file state is current in your context — no need to Read it back)

The file E:\a327ex\steam-market\scripts\goldrush.py has been updated successfully. (file state is current in your context — no need to Read it back)

6:=== ANATOMY BY TRIGGER CLASS (mature triggers, >=6Q old, pre&win n>=8) ===
12:--- window-vs-pre split by tag pre-rate (indie+f2p triggers) ---
16:--- supply-flood lag by trigger year (all concentrated triggers) ---
28:--- indie/f2p rushes ranked by POST-FLOOD delta (durability), with trigger-time features ---
117:--- indie-trigger rushes, individually (window-vs-pre, sorted) ---
172:=== CURRENT EDGE: triggers in last 8 quarters (all classes), grouped by trigger game ===

6    === ANATOMY BY TRIGGER CLASS (mature triggers, >=6Q old, pre&win n>=8) ===
7   indie    n=53  window-vs-pre median -3pt | postflood-vs-pre -5pt (n=52) | win med 23% pre med 22%
8   big      n=67  window-vs-pre median -3pt | postflood-vs-pre -5pt (n=66) | win med 20% pre med 26%
9   f2p      n=35  window-vs-pre median -3pt | postflood-vs-pre -3pt (n=34) | win med 22% pre med 25%
10  old      n=12  window-vs-pre median -5pt | postflood-vs-pre -7pt (n=12) | win med 17% pre med 23%
11  
12  --- window-vs-pre split by tag pre-rate (indie+f2p triggers) ---
13  pre<20%   n=35  median delta +2pt | mean +5pt
14  pre>=20%  n=53  median delta -10pt | mean -7pt
15  
16  --- supply-flood lag by trigger year (all concentrated triggers) ---
17  2015: triggers=63  flooded within 8Q: 16/63 (25%) | median lag of flooded: 5.0Q
18  2016: triggers=30  flooded within 8Q: 6/30 (20%) | median lag of flooded: 6.0Q
19  2017: triggers=33  flooded within 8Q: 2/33 (6%) | median lag of flooded: 4.0Q
20  2018: triggers=12  flooded within 8Q: 3/12 (25%) | median lag of flooded: 3Q
21  2019: triggers=30  flooded within 8Q: 12/30 (40%) | median lag of flooded: 5.0Q
22  2020: triggers=44  flooded within 8Q: 22/44 (50%) | median lag of flooded: 1.0Q
23  2021: triggers=18  flooded within 8Q: 5/18 (28%) | median lag of flooded: 2Q
24  2022: triggers=17  flooded within 8Q: 7/17 (41%) | median lag of flooded: 7Q
25  2023: triggers=18  flooded within 8Q: 4/18 (22%) | median lag of flooded: 4.5Q
26  2024: triggers=9   flooded within 8Q: 5/9 (56%) | median lag of flooded: 2Q
27  
28  --- indie/f2p rushes ranked by POST-FLOOD delta (durability), with trigger-time features ---
29                    tag                         pre->win->post | trigRev preSup preFlow/Q nGames
30  2023Q4 FMV                         28-> 55-> 60 (+32) | $  0.5M    8/Q   3068 Love Is All Around
31  2015Q2 Time Management             23-> 16-> 50 (+27) | $  0.0M    3/Q    183 Freestyle 2: Street Basketball
32  2024Q1 Trading                     17-> 31-> 36 (+20) | $ 24.3M    4/Q   4474 Supermarket Simulator
33  2018Q1 Old School                   0-> 17-> 19 (+19) | $ 12.1M    2/Q    398 DUSK
34  2015Q1 Turn-Based Combat           44-> 56-> 62 (+17) | $ 50.5M    2/Q    470 Darkest Dungeon®
35  2019Q1 Thriller                    15-> 29-> 32 (+16) | $  0.7M    4/Q    389 Unheard - Voices of Crime
... [81 more lines]

172  === CURRENT EDGE: triggers in last 8 quarters (all classes), grouped by trigger game ===
173 
174 [BIG] Black Myth: Wukong  $60  est $149.9M  released 2024-08-19  trigger 2024Q3
175    OPEN  Action RPG                 2024Q3 flow=   861K x5 top1=79% | supply 146->213/Q | followers n=1103 h50=17% | Borderlands 4 $167.0M, Path of Exile 2 $102.7M, Dragon Age™: The Veilguard $62.3M
176    OPEN  Souls-like                 2024Q3 flow=   835K x6 top1=82% | supply 40->44/Q | followers n=350 h50=22% | ELDEN RING NIGHTREIGN $106.1M, Path of Exile 2 $102.7M, Hollow Knight: Silksong $102.2M
177    OPEN  Combat                     2024Q3 flow=   770K x8 top1=89% | supply 68->116/Q | followers n=551 h50=15% | Crimson Desert $148.4M, Assassin’s Creed Shadows $38.5M, Undisputed $24.7M
178    FLOOD Mythology                  2024Q3 flow=   733K x23 top1=93% | supply 12->28/Q | followers n=113 h50=17% | Hades II $69.3M, Titan Quest II $9.1M, One-Eyed Likho $0.8M
179 
180 [INDIE] Hollow Knight: Silksong  $20  est $102.2M  released 2025-09-04  trigger 2025Q3
181    OPEN  Metroidvania               2025Q3 flow=   334K x5 top1=72% | supply 50->51/Q | followers n=162 h50=22% | GRIME II $1.7M, MIO: Memories in Orbit $1.1M, Constance $0.8M
182 
183 [F2P] Marvel Rivals  F2P  est $0.0M  released 2024-12-05  trigger 2024Q4
184    OPEN  Superhero                  2024Q4 flow=   153K x5 top1=83% | supply 6->3/Q | followers n=25 h50=32% | Dispatch $114.0M, Marvel's Spider-Man 2 $33.7M, LEGO® Batman™: Legacy of the Dark Knight $15.0M
185 
186 [BIG] Subnautica 2  $30  est $73.7M  released 2026-05-14  trigger 2026Q2
187    OPEN  Underwater                 2026Q2 flow=   150K x8 top1=74% | supply 17->27/Q | followers n=0 h50=-
188 
189 [INDIE] Schedule I  $20  est $140.0M  released 2025-03-24  trigger 2025Q1
190    OPEN  Crime                      2025Q1 flow=   141K x4 top1=49% | supply 19->33/Q | followers n=116 h50=19% | Mafia: The Old Country $16.0M, The Precinct $4.2M, Cash Cleaner Simulator $2.7M
191 
192 [BIG] Monster Hunter Wilds  $70  est $209.1M  released 2025-02-27  trigger 2025Q1
193    OPEN  Hunting                    2025Q1 flow=   137K x8 top1=87% | supply 7->10/Q | followers n=48 h50=17% | WolfQuest: Anniversary Edition $10.6M, Way of the Hunter 2 $1.3M, Deathground $0.5M
194    OPEN  Dragons                    2025Q1 flow=   123K x46 top1=97% | supply 8->5/Q | followers n=39 h50=13% | Huaxia: Warring States $0.4M, The RPG $0.4M, Dragonhold $0.1M
195 
196 [INDIE] Escape From Duckov  $18  est $6.7M  released 2025-10-16  trigger 2025Q4
197    OPEN  Top-Down Shooter           2025Q4 flow=    97K x4 top1=80% | supply 116->125/Q | followers n=209 h50=4% | Arms of God $0.8M, Vital Shell $0.3M, Jackal $0.3M
198 
199 [INDIE] Slay the Spire 2  $25  est $57.2M  released 2026-03-05  trigger 2026Q1
200    OPEN  Political                  2026Q1 flow=    97K x23 top1=96% | supply 8->11/Q | followers n=11 h50=18% | Teller's Duty $0.2M, Age of History 2: Definitive Edition $0.2M
201 
... [78 more lines]

RL Deckbuilder post-Balatro-trigger (2025-01..2026-06): n=273 h50=32%

  <=$10: n=148 h50=16% | >$10: n=125 h50=50%

    $10 $0.34M Aotenjo: Infinite Hands (2025-01-19)

    $9 $0.32M Witch's Apocalyptic Journey (2026-06-04)

    $9 $0.23M Three Kingdoms: The Blood Moon (2025-04-02)

    $10 $0.22M Dogpile (2025-12-10)

    $10 $0.21M Apokerlypse (2026-04-29)

    $9 $0.17M SealChain: Call of Blood (2026-02-20)

    $10 $0.16M Kingdom's Deck (2025-04-27)

    $10 $0.13M Goblin Sushi (2026-02-09)



Crime post-Schedule I (2025-04..2026-06): n=116 h50=19%

  <=$10: n=78 h50=9% | >$10: n=38 h50=39%

    $10 $0.28M Brush Burial: Gutter World (2025-11-19)

    $8 $0.25M Prison Escape Simulator: Dig Out (2025-07-25)

    $9 $0.14M A Case of Fraud (2025-11-10)

    $10 $0.12M Bum: Revenge (2025-07-30)

    $2 $0.09M QUIET DAN (2025-05-07)

    $6 $0.08M Dread Neighbor (2026-05-07)

    $10 $0.08M 春泥 Soil of Rebirth (2026-01-16)



Shop Keeper 2026 flood cohort: n=169 h50=16%

  <=$10: n=128 h50=9% | >$10: n=41 h50=37%

    $3 $2.29M BidKing (2026-04-15)

    $10 $0.17M Rest Area Simulator (2026-04-24)

    $7 $0.12M Loot Tycoon (2026-04-19)

    $8 $0.11M Skeleseller (2026-03-17)

    $8 $0.11M BakeryDiary (2026-03-06)

    $8 $0.09M Retail Hell (2026-05-21)

    $10 $0.08M Dungeon Bodega Simulator (2026-03-23)

... [1 more lines]

File created successfully at: E:\a327ex\steam-market\reports\08-goldrush.md (file state is current in your context — no need to Read it back)

The file E:\a327ex\steam-market\reports\notes.md has been updated successfully. (file state is current in your context — no need to Read it back)

Exit code 1
tail: option used in invalid context -- 2

== b5iamgfno ==
queue: 5495 appids (min_reviews=50)

500/5495 done

== bhbsxf9mq ==
3773390: 429, sleeping 90s

3182170: 429, sleeping 90s

== bkrhggaew ==
100/3889 (hit)

200/3889 (hit)

== b5iamgfno ==
queue: 5495 appids (min_reviews=50)

500/5495 done

== bhbsxf9mq ==
queue: 0 window-successes + 4000 window-failures + 3000 window-mid + 50694 reviewed rest

3773390: 429, sleeping 90s

3182170: 429, sleeping 90s

2336650: 429, sleeping 90s

== bkrhggaew ==
queue: 3889 (hit=2289, mid sampled 800, low sampled 800; 0 already done)

100/3889 (hit)

200/3889 (hit)

300/3889 (hit)

Exit code 127
/usr/bin/bash: line 10: ./.venv/Scripts/python.exe: No such file or directory

histogram done: 30077

appdetails ok: 18434 | failed: 0

appdetails fetched last 10min: 344

reviews_text done: 327 | reviews stored: 20495

Vampire Survivors: ['Action Roguelike', 'Bullet Hell', 'Vampires', 'Roguelite', 'Pixel Graphics', '2D', 'Indie']

Brotato: ['Roguelike', 'Singleplayer', 'Action Roguelike', 'Bullet Hell', 'Bullet Heaven', 'Local Co-Op', 'Multiplayer']

20 Minutes Till Dawn: ['Action Roguelike', 'Bullet Hell', 'Survival', 'Roguelite', "Shoot 'Em Up", 'Horror', 'Top-Down Shooter']

Rogue: Genesia: ['Bullet Heaven', 'Bullet Hell', 'Action Roguelike', 'Roguelite', 'Arcade', 'Twin Stick Shooter', 'PvE']

Halls of Torment: ['Action Roguelike', 'Roguelite', 'Bullet Hell', 'Pixel Graphics', 'Dark Fantasy', 'Action RPG', 'Roguelike']

HoloCure - Save the Fans!: ['Pixel Graphics', 'Bullet Heaven', 'Anime', 'Action Roguelike', 'Bullet Hell', 'Cute', 'Roguelite']



Bullet Heaven tag: 241 carriers. Top 8 by est revenue:

  Megabonk (2025-09-18) $20.4M

  Deep Rock Galactic: Survivor (2025-09-17) $11.1M

  Soulstone Survivors (2025-06-17) $6.6M

  Brotato (2023-06-23) $5.6M

  Rogue: Genesia (2025-03-07) $2.7M

  Yet Another Zombie Survivors (2023-07-13) $2.5M

  The Spell Brigade (2026-04-29) $2.2M

  Jotunnslayer: Hordes of Hel (2025-09-03) $2.0M

Earliest releases carrying it: [('Tower of Guns', '2014-03-04'), ('Bullet Heaven 2', '2015-12-07'), ('Geometry Arena', '2021-06-01'), ('Dextram', '2021-11-08'), ('HordeCore', '2022-03-02')]

File created successfully at: E:\a327ex\steam-market\scripts\breakout.py (file state is current in your context — no need to Read it back)

Exit code 2
/usr/bin/bash: line 1: reports/raw_breakout.txt: No such file or directory
wc: reports/raw_breakout.txt: No such file or directory
grep: reports/raw_breakout.txt: No such file or directory

4:=== 2018Q1 — top 10 indie-priced (<=$25) by first-90-day reviews ===
184:=== 2021Q4 — top 10 indie-priced (<=$25) by first-90-day reviews ===
196:=== 2022Q1 — top 10 indie-priced (<=$25) by first-90-day reviews ===
280:=== 2023Q4 — top 10 indie-priced (<=$25) by first-90-day reviews ===
292:=== 2024Q1 — top 10 indie-priced (<=$25) by first-90-day reviews ===

4    === 2018Q1 — top 10 indie-priced (<=$25) by first-90-day reviews ===
5      1. v90= 3,560  $20 pub     Sword Art Online: Fatal Bullet               life=  5,144 est=$   3.7M | Anime,RPG,Character Customization,Action,Multiplayer
6      2. v90= 2,474  $15 selfpub Into the Breach                              life= 12,518 est=$   6.7M | Strategy,Turn-Based Strategy,Indie,Pixel Graphics,Turn-Based Tactics
7      3. v90= 1,461  $20 pub     Celeste                                      life= 60,239 est=$  43.2M | Precision Platformer,Difficult,Great Soundtrack,Pixel Graphics,Indie
8      4. v90= 1,134  $11 selfpub 永远消失的幻想乡 ～ The Disappearing of Gensokyo      life=    507 est=$   0.2M | Anime,Bullet Hell,Female Protagonist,Faith,Cute
9      5. v90= 1,029  $10 selfpub Dungreed                                     life=  2,499 est=$   0.9M | Action Roguelike,Pixel Graphics,Roguelite,Roguelike,Action
10     6. v90=   974  $25 pub     東方憑依華　～ Antinomy of Common Flowers.          life=  1,761 est=$   1.6M | Great Soundtrack,Fighting,2D Fighter,Female Protagonist,Faith
11     7. v90=   790  $15 pub     Heroes of Hammerwatch                        life=  4,672 est=$   2.5M | Roguelite,Action Roguelike,Co-op,Pixel Graphics,Multiplayer
12     8. v90=   726  $20 selfpub Ghost of a Tale                              life=  1,992 est=$   1.4M | Adventure,Stealth,Atmospheric,Cute,Singleplayer
13     9. v90=   672  $10 selfpub We Were Here Too                             life=  3,601 est=$   1.3M | Online Co-Op,Escape Room,Puzzle,Co-op,Horror
14    10. v90=   616  $25 pub     Ash of Gods: Redemption                      life=  1,095 est=$   1.0M | Singleplayer,RPG,Turn-Based Combat,Story Rich,Isometric
15  
16  === 2018Q2 — top 10 indie-priced (<=$25) by first-90-day reviews ===
17     1. v90= 8,536  $20 selfpub The Forest                                   life=176,877 est=$ 126.9M | Open World Survival Craft,Survival,Horror,Open World,Crafting
18     2. v90= 2,341  $16 pub     Wizard of Legend                             life=  7,313 est=$   4.2M | Action Roguelike,Pixel Graphics,Roguelike,Dungeon Crawler,Local Co-Op
19     3. v90= 2,137  $20 pub     Pillars of Eternity II: Deadfire             life=  8,253 est=$   5.9M | RPG,Party-Based RPG,Isometric,CRPG,Story Rich
20     4. v90= 1,646  $15 selfpub FAR: Lone Sails                              life=  8,889 est=$   4.8M | Atmospheric,Adventure,Post-apocalyptic,Singleplayer,Exploration
21     5. v90= 1,575  $20 pub     Moonlighter                                  life=  7,194 est=$   5.2M | Hack and Slash,Pixel Graphics,Dungeon Crawler,Roguelite,Roguelike
22     6. v90= 1,329  $15 pub     Agony                                        life=    923 est=$   0.5M | Sexual Content,Nudity,Horror,Gore,Demons
23     7. v90= 1,154  $15 pub     Bus Simulator 18                             life=  1,919 est=$   1.0M | Simulation,Automobile Sim,Driving,Multiplayer,Realistic
24     8. v90=   884  $10 selfpub Bloodstained: Curse of the Moon              life=  1,977 est=$   0.7M | Action,Pixel Graphics,Retro,Platformer,2D
25     9. v90=   881  $20 pub     Farm Manager 2018                            life=    781 est=$   0.6M | Farming Sim,Indie,Management,Casual,Agriculture
26    10. v90=   850  $12 pub     Monster Prom                                 life=  9,006 est=$   3.9M | Dating Sim,Comedy,LGBTQ+,Multiplayer,Choices Matter
27  
28  === 2018Q3 — top 10 indie-priced (<=$25) by first-90-day reviews ===
29     1. v90= 6,406  $10 selfpub Chinese Parents                              life=  1,580 est=$   0.6M | Life Sim,Psychological Horror,Memes,Simulation,Casual
30     2. v90= 4,917  $20 pub     Pathfinder: Kingmaker — Enhanced Plus Editio life= 17,700 est=$  12.7M | RPG,Party-Based RPG,Fantasy,CRPG,Isometric
31     3. v90= 3,294  $20 pub     The Walking Dead: The Final Season           life=  6,412 est=$   4.6M | Story Rich,Choices Matter,Zombies,Adventure,Choose Your Own Adventure
32     4. v90= 3,267  $20 pub     Graveyard Keeper                             life= 17,389 est=$  12.5M | Pixel Graphics,Crafting,RPG,Simulation,Management
33     5. v90= 2,657  $25 selfpub Dead Cells                                   life= 44,043 est=$  39.5M | Adventure,Roguelike,Pixel Graphics,Metroidvania,Action Roguelike
... [30 more lines]

184  === 2021Q4 — top 10 indie-priced (<=$25) by first-90-day reviews ===
185    1. v90=48,897  $20 pub     Halo Infinite                                life=138,808 est=$  98.6M | Free to Play,FPS,Multiplayer,Singleplayer,Shooter
186    2. v90=42,578  $20 pub     Inscryption                                  life= 75,851 est=$  54.4M | Card Battler,Card Game,Story Rich,Deckbuilding,Horror
187    3. v90=23,477  $20 pub     Bright Memory: Infinite                      life=  4,735 est=$   3.4M | Action,Female Protagonist,Sexual Content,FPS,Indie
188    4. v90=16,515  $20 selfpub Farming Simulator 22                         life= 25,479 est=$  18.3M | Simulation,Farming Sim,Co-op,Multiplayer,Agriculture
189    5. v90=14,767  $20 pub     Gunfire Reborn                               life= 36,587 est=$  26.3M | Roguelite,FPS,Roguelike,Online Co-Op,Co-op
190    6. v90= 7,507  $23 pub     Fate Seeker II                               life= 11,343 est=$   9.2M | RPG,Martial Arts,Action RPG,Puzzle,Exploration
191    7. v90= 6,931  $20 pub     Unpacking                                    life= 22,658 est=$  16.3M | Relaxing,Cute,Casual,Pixel Graphics,Organizing
192    8. v90= 5,957  $11 pub     小白兔电商~Bunny e-Shop                           life=  9,177 est=$   3.6M | Capitalism,Visual Novel,Dating Sim,Anime,Simulation
193    9. v90= 4,694  $20 selfpub WorldBox - God Simulator                     life= 28,918 est=$  20.8M | Sandbox,God Game,Simulation,Pixel Graphics,Colony Sim
194   10. v90= 4,107  $16 pub     F.I.S.T.: Forged In Shadow Torch             life=    731 est=$   0.4M | Metroidvania,Platformer,Action,Steampunk,Adventure
195 
196 === 2022Q1 — top 10 indie-priced (<=$25) by first-90-day reviews ===
197    1. v90=29,232   $2 pub     Mirror 2: Project X                          life=  2,425 est=$   0.2M | 3D,Match 3,Anime,Female Protagonist,Visual Novel
198    2. v90=14,683  $18 pub     暖雪 Warm Snow                                 life=  1,301 est=$   0.8M | Action Roguelike,Roguelite,Martial Arts,Action-Adventure,Roguelike
199    3. v90= 9,924  $16 pub     NEEDY STREAMER OVERLOAD                      life= 17,249 est=$   9.9M | Psychological Horror,Anime,Visual Novel,Sexual Content,Cute
200    4. v90= 4,564  $24 selfpub Internet Cafe Simulator 2                    life=  2,093 est=$   1.8M | Simulation,Management,3D,Exploration,Memes
201    5. v90= 2,766   $8 selfpub Iron Lung                                    life=  8,335 est=$   2.4M | Horror,Atmospheric,Psychological Horror,Submarine,Underwater
202    6. v90= 2,519  $25 selfpub ANNO: Mutationem                             life=  1,902 est=$   1.7M | RPG,Cyberpunk,Pixel Graphics,Female Protagonist,2.5D
203    7. v90= 2,377  $25 selfpub 嗜血印 Bloody Spell                             life=  2,396 est=$   2.1M | Nudity,Sexual Content,Female Protagonist,Action RPG,Hack and Slash
204    8. v90= 2,061  $15 selfpub Rogue Tower                                  life=  4,347 est=$   2.3M | Tower Defense,Strategy,Roguelike,Casual,Procedural Generation
205    9. v90= 2,041  $18 pub     Hero's Hour                                  life=  2,453 est=$   1.6M | Turn-Based,Pixel Graphics,Strategy,Turn-Based Strategy,Singleplayer
206   10. v90= 1,923  $25 selfpub Nobody Saves the World                       life=  2,256 est=$   2.0M | Singleplayer,Indie,Co-op,Multiplayer,Action
207 
208 === 2022Q2 — top 10 indie-priced (<=$25) by first-90-day reviews ===
209    1. v90=30,250  $20 pub     Raft                                         life=129,202 est=$  92.7M | Survival,Multiplayer,Open World Survival Craft,Co-op,Crafting
210    2. v90=13,039  $25 selfpub The Stanley Parable: Ultra Deluxe            life= 24,018 est=$  21.5M | Multiple Endings,Choices Matter,Comedy,Walking Simulator,First-Person
211    3. v90= 8,007   $7 pub     Inside the Backrooms                         life= 10,222 est=$   2.6M | Horror,Multiplayer,Survival Horror,Puzzle,Psychological Horror
212    4. v90= 6,682  $25 pub     Teenage Mutant Ninja Turtles: Shredder's Rev life=  5,913 est=$   5.3M | Beat 'em up,Multiplayer,Pixel Graphics,Side Scroller,Co-op Campaign
213    5. v90= 6,032   $8 selfpub Stacklands                                   life= 11,751 est=$   3.4M | Card Game,Management,Survival,Roguelite,Card Battler
... [6 more lines]

<stdin>:10: DeprecationWarning: datetime.datetime.utcfromtimestamp() is deprecated and scheduled for removal in a future version. Use timezone-aware objects to represent datetimes in UTC: datetime.datetime.fromtimestamp(timestamp, datetime.UTC).

(1794680, 'Vampire Survivors', 'Oct 20, 2022', 499, 127089)

(2379780, 'Balatro', 'Feb 20, 2024', 1499, 102171)

(2653790, 'The Exit 8', 'Nov 29, 2023', 399, 5005)

(2780270, 'Dead end Exit 8', 'Feb 17, 2024', 399, 111)

(2878720, 'Unable to Exit 8', 'Jul 18, 2024', 499, 138)

(2924030, 'Overtime Heroes Exit 8', 'May 15, 2024', 499, 27)

(3164500, 'Schedule I', 'Mar 24, 2025', 1999, 195048)

(3193770, 'The Exit 8 VR', 'Sep 19, 2024', 768, 61)

(3265700, 'Vampire Crawlers: The Turbo Wildcard from Vampire Survivors', 'Apr 21, 2026', 999, 8008)

(3435600, 'PEXIT 8', 'Feb 10, 2025', 0, 253)

(3676450, 'Corridor Exit 8', 'Jun 27, 2025', 199, 37)

(4833100, 'Backrooms: Exit 8', 'Jun 25, 2026', 399, None)

VS earliest rollups: [('2021-12', 11), ('2022-01', 18399), ('2022-02', 34649), ('2022-03', 13920), ('2022-04', 18919), ('2022-05', 11518)]

The file E:\a327ex\steam-market\scripts\breakout.py has been updated successfully. (file state is current in your context — no need to Read it back)

The file E:\a327ex\steam-market\scripts\breakout.py has been updated successfully. (file state is current in your context — no need to Read it back)

=== 2021Q4 — top 10 indie-priced (<=$25) by first-90-day reviews ===
   1. v90=131,223  $20 pub     Halo Infinite                                life=138,808 est=$  98.6M | Free to Play,FPS,Multiplayer,Singleplayer,Shooter
   2. v90=53,059   $5 selfpub Vampire Survivors                            life=127,089 est=$  22.8M | Action Roguelike,Bullet Hell,Vampires,Roguelite,Pixel Graphics
   3. v90=42,578  $20 pub     Inscryption                                  life= 75,851 est=$  54.4M | Card Battler,Card Game,Story Rich,Deckbuilding,Horror
   4. v90=23,477  $20 pub     Bright Memory: Infinite                      life=  4,735 est=$   3.4M | Action,Female Protagonist,Sexual Content,FPS,Indie
   5. v90=16,515  $20 selfpub Farming Simulator 22                         life= 25,479 est=$  18.3M | Simulation,Farming Sim,Co-op,Multiplayer,Agriculture
   6. v90= 7,507  $23 pub     Fate Seeker II                               life= 11,343 est=$   9.2M | RPG,Martial Arts,Action RPG,Puzzle,Exploration
   7. v90= 6,931  $20 pub     Unpacking                                    life= 22,658 est=$  16.3M | Relaxing,Cute,Casual,Pixel Graphics,Organizing
   8. v90= 5,957  $11 pub     小白兔电商~Bunny e-Shop                           life=  9,177 est=$   3.6M | Capitalism,Visual Novel,Dating Sim,Anime,Simulation
   9. v90= 4,694  $20 selfpub WorldBox - God Simulator                     life= 28,918 est=$  20.8M | Sandbox,God Game,Simulation,Pixel Graphics,Colony Sim
  10. v90= 4,107  $16 pub     F.I.S.T.: Forged In Shadow Torch             life=    731 est=$   0.4M | Metroidvania,Platformer,Action,Steampunk,Adventure

=== 2022Q1 — top 10 indie-priced (<=$25) by first-90-day reviews ===
   1. v90=29,232   $2 pub     Mirror 2: Project X                          life=  2,425 est=$   0.2M | 3D,Match 3,Anime,Female Protagonist,Visual Novel
   2. v90=14,683  $18 pub     暖雪 Warm Snow                                 life=  1,301 est=$   0.8M | Action Roguelike,Roguelite,Martial Arts,Action-Adventure,Roguelike
   3. v90= 9,924  $16 pub     NEEDY STREAMER OVERLOAD                      life= 17,249 est=$   9.9M | Psychological Horror,Anime,Visual Novel,Sexual Content,Cute
=== 2023Q4 — top 10 indie-priced (<=$25) by first-90-day reviews ===
   1. v90=189,939  $10 selfpub Lethal Company                               life=275,837 est=$  98.9M | Online Co-Op,Horror,First-Person,Co-op,Survival Horror
   2. v90=30,694  $10 selfpub Love Is All Around                           life=  1,519 est=$   0.5M | Dating Sim,Sexual Content,FMV,Singleplayer,Visual Novel
   3. v90=22,793  $15 pub     Risk of Rain Returns                         life= 16,699 est=$   9.0M | Action Roguelike,Online Co-Op,Multiplayer,2D,PvE
   4. v90=11,444  $12 selfpub The Coffin of Andy and Leyley                life= 18,717 est=$   8.1M | Psychological,Dark Comedy,Story Rich,Multiple Endings,Psychological Horror
   5. v90=11,226   $7 selfpub Death Must Die                               life= 13,926 est=$   3.5M | Early Access,Action Roguelike,Bullet Hell,Roguelite,Pixel Graphics
   6. v90= 8,130  $25 pub     For The King II                              life=  5,317 est=$   4.8M | Adventure,RPG,Strategy,Party-Based RPG,Board Game
   7. v90= 6,976  $20 selfpub The Front                                    life=  3,716 est=$   2.7M | Open World Survival Craft,Survival,Multiplayer,Crafting,Base Building
   8. v90= 6,020  $18 pub     Slay the Princess — The Pristine Cut         life= 22,856 est=$  14.8M | Visual Novel,Choices Matter,Psychological Horror,Story Rich,Horror
   9. v90= 5,774  $15 selfpub Stronghold: Definitive Edition               life=  3,099 est=$   1.7M | City Builder,Strategy,Real Time Tactics,Medieval,Sandbox
  10. v90= 4,590  $16 pub     Magicraft                                    life=  3,511 est=$   2.0M | Action Roguelike,Roguelite,Bullet Hell,Looter Shooter,Roguelike

=== 2024Q1 — top 10 indie-priced (<=$25) by first-90-day reviews ===
   1. v90=30,125  $15 pub     Balatro                                      life=102,171 est=$  55.0M | Card Game,Roguelike,Roguelike Deckbuilder,Deckbuilding,Singleplayer
... [11 more lines]

=== 2025Q1 — top 10 indie-priced (<=$25) by first-90-day reviews ===
   1. v90=187,758  $20 selfpub Schedule I                                   life=195,048 est=$ 140.0M | Simulation,Crime,Co-op,Management,Multiplayer
   2. v90=129,330  $10 selfpub R.E.P.O.                                     life=134,257 est=$  48.2M | Horror,Online Co-Op,Multiplayer,Comedy,Co-op
   3. v90=15,551  $25 pub     Sultan's Game                                life=  1,284 est=$   1.2M | Singleplayer,Story Rich,RPG,Card Game,Strategy
   4. v90=12,882   $5 pub     A Game About Digging A Hole™                 life= 11,122 est=$   2.0M | Exploration,Mining,Sandbox,Voxel,Horror
   5. v90= 8,973   $5 selfpub Nubby's Number Factory                       life= 16,311 est=$   2.9M | Roguelike,Strategy,Arcade,Surreal,Singleplayer
   6. v90= 6,310   $6 selfpub FlyKnight                                    life=  9,081 est=$   2.0M | Dungeon Crawler,Souls-like,Online Co-Op,First-Person,Old School
   7. v90= 5,984  $25 pub     Drive Beyond Horizons                        life=  6,884 est=$   6.2M | Early Access,Driving,Exploration,Survival,Automobile Sim
   8. v90= 5,386  $10 selfpub The Quinfall                                 life=  3,326 est=$   1.2M | Massively Multiplayer,MMORPG,Open World,RPG,Character Customization
   9. v90= 4,558  $20 pub     Legend of Heroes: Three Kingdoms             life=  6,547 est=$   4.7M | Early Access,Simulation,Strategy,RPG,Turn-Based Strategy
  10. v90= 4,131   $8 selfpub The Headliners                               life=  4,212 est=$   1.2M | Online Co-Op,Horror,Co-op,Funny,Multiplayer
  -- uncapped top 5 (any price/F2P):
   1. v90=187,758  $20 selfpub Schedule I                                   life=195,048 est=$ 140.0M | Simulation,Crime,Co-op,Management,Multiplayer
   2. v90=130,226  $70 selfpub Monster Hunter Wilds                         life= 83,233 est=$ 209.1M | Hunting,Action,Multiplayer,Online Co-Op,Co-op
   3. v90=129,330  $10 selfpub R.E.P.O.                                     life=134,257 est=$  48.2M | Horror,Online Co-Op,Multiplayer,Comedy,Co-op
   4. v90=73,693  $60 pub     Kingdom Come: Deliverance II                 life= 65,358 est=$ 140.8M | RPG,Medieval,Open World,Singleplayer,Story Rich
   5. v90=52,844  $30 pub     Grand Theft Auto V Enhanced                  life= 66,625 est=$  71.7M | Open World,Sexual Content,Action,Multiplayer,Crime

=== 2025Q2 — top 10 indie-priced (<=$25) by first-90-day reviews ===
   1. v90=111,099   $8 pub     PEAK                                         life=132,076 est=$  37.9M | Multiplayer,Online Co-Op,Co-op,Physics,Adventure
   2. v90=68,084  $25 selfpub DELTARUNE                                    life= 58,375 est=$  52.4M | Story Rich,Great Soundtrack,Pixel Graphics,Funny,RPG
   3. v90=44,040   $7 pub     Revenge on gold diggers                      life=    526 est=$   0.1M | RPG,Story Rich,Dating Sim,Interactive Fiction,Strategy
   4. v90=10,299  $20 pub     9 Kings                                      life=  9,592 est=$   6.9M | Strategy,Roguelike,Card Game,Deckbuilding,Roguelike Deckbuilder
   5. v90= 6,365  $20 pub     Haste                                        life=  8,577 est=$   6.2M | Fast-Paced,Parkour,Roguelite,3D Platformer,Female Protagonist
   6. v90= 5,299  $25 pub     Monster Train 2                              life=  4,935 est=$   4.4M | Roguelike Deckbuilder,Card Game,Roguelike,Deckbuilding,Turn-Based Strategy
   7. v90= 5,220   $3 pub     Tower Wizard                                 life=  6,926 est=$   0.7M | Incremental,Idler,Resource Management,Strategy,Fantasy
   8. v90= 4,907  $10 selfpub Aviassembly                                  life=  6,254 est=$   2.2M | Building,Physics,Sandbox,Flight,Transportation
   9. v90= 4,518  $25 selfpub Deadzone: Rogue                              life=  8,313 est=$   7.5M | Roguelike,FPS,Online Co-Op,PvE,Roguelite
  10. v90= 4,392  $25 pub     Into the Dead: Our Darkest Days              life=  2,889 est=$   2.6M | Zombies,Survival Horror,Post-apocalyptic,Horror,Action
  -- uncapped top 5 (any price/F2P):
... [77 more lines]

File created successfully at: E:\a327ex\steam-market\reports\09-breakout.md (file state is current in your context — no need to Read it back)

The file E:\a327ex\steam-market\reports\notes.md has been updated successfully. (file state is current in your context — no need to Read it back)

170 reports/raw_breakout_recent20.txt

1    loading catalog...
2   loading rollups...
3   
4   === 2025Q1 — top 20 indie-priced (<=$25) by first-90-day reviews ===
5      1. v90=187,758  $20 selfpub Schedule I                                   life=195,048 est=$ 140.0M | Simulation,Crime,Co-op,Management,Multiplayer
6      2. v90=129,330  $10 selfpub R.E.P.O.                                     life=134,257 est=$  48.2M | Horror,Online Co-Op,Multiplayer,Comedy,Co-op
7      3. v90=15,551  $25 pub     Sultan's Game                                life=  1,284 est=$   1.2M | Singleplayer,Story Rich,RPG,Card Game,Strategy
8      4. v90=12,882   $5 pub     A Game About Digging A Hole™                 life= 11,122 est=$   2.0M | Exploration,Mining,Sandbox,Voxel,Horror
9      5. v90= 8,973   $5 selfpub Nubby's Number Factory                       life= 16,311 est=$   2.9M | Roguelike,Strategy,Arcade,Surreal,Singleplayer
10     6. v90= 6,310   $6 selfpub FlyKnight                                    life=  9,081 est=$   2.0M | Dungeon Crawler,Souls-like,Online Co-Op,First-Person,Old School
11     7. v90= 5,984  $25 pub     Drive Beyond Horizons                        life=  6,884 est=$   6.2M | Early Access,Driving,Exploration,Survival,Automobile Sim
12     8. v90= 5,386  $10 selfpub The Quinfall                                 life=  3,326 est=$   1.2M | Massively Multiplayer,MMORPG,Open World,RPG,Character Customization
13     9. v90= 4,558  $20 pub     Legend of Heroes: Three Kingdoms             life=  6,547 est=$   4.7M | Early Access,Simulation,Strategy,RPG,Turn-Based Strategy
14    10. v90= 4,131   $8 selfpub The Headliners                               life=  4,212 est=$   1.2M | Online Co-Op,Horror,Co-op,Funny,Multiplayer
15    11. v90= 3,826  $20 pub     Heroes of Hammerwatch II                     life=  4,078 est=$   2.9M | RPG,Action Roguelike,Action RPG,Action,Dungeon Crawler
16    12. v90= 3,357  $18 pub     Urban Myth Dissolution Center                life=    615 est=$   0.4M | Visual Novel,Psychological Horror,Adventure,Pixel Graphics,Singleplayer
17    13. v90= 3,267   $9 pub     Tiny Pasture                                 life=  1,164 est=$   0.4M | Idler,Cute,Farming Sim,Incremental,Pixel Graphics
18    14. v90= 3,251  $19 pub     Jotunnslayer: Hordes of Hel                  life=  2,946 est=$   2.0M | Action Roguelike,Bullet Heaven,Roguelite,Action RPG,Hack and Slash
19    15. v90= 3,240   $1 selfpub Your Mother                                  life=    905 est=$   0.0M | Desktop Companion,Anime,Female Protagonist,Violent,2D
20    16. v90= 3,102   $8 selfpub Nunholy                                      life=  1,893 est=$   0.5M | Female Protagonist,Action Roguelike,Sexual Content,Gothic,Singleplayer
21    17. v90= 3,022   $9 pub     Matchmaking Inc.                             life=  5,610 est=$   1.8M | Casual,Dating Sim,RPG,Time Management,Multiple Endings
22    18. v90= 2,530  $20 pub     KARMA: The Dark World                        life=    827 est=$   0.6M | Walking Simulator,Psychological Horror,Horror,First-Person,Atmospheric
23    19. v90= 2,508  $10 selfpub 灾厄黑龙与谎言公主                                    life=  3,578 est=$   1.3M | Visual Novel,RPG,Otome,Dating Sim,Female Protagonist
24    20. v90= 2,504  $10 pub     Look Outside                                 life=  8,757 est=$   3.1M | RPG,Survival Horror,Horror,Pixel Graphics,Lovecraftian
25    -- uncapped top 5 (any price/F2P):
26     1. v90=187,758  $20 selfpub Schedule I                                   life=195,048 est=$ 140.0M | Simulation,Crime,Co-op,Management,Multiplayer
27     2. v90=130,226  $70 selfpub Monster Hunter Wilds                         life= 83,233 est=$ 209.1M | Hunting,Action,Multiplayer,Online Co-Op,Co-op
28     3. v90=129,330  $10 selfpub R.E.P.O.                                     life=134,257 est=$  48.2M | Horror,Online Co-Op,Multiplayer,Comedy,Co-op
29     4. v90=73,693  $60 pub     Kingdom Come: Deliverance II                 life= 65,358 est=$ 140.8M | RPG,Medieval,Open World,Singleplayer,Story Rich
30     5. v90=52,844  $30 pub     Grand Theft Auto V Enhanced                  life= 66,625 est=$  71.7M | Open World,Sexual Content,Action,Multiplayer,Crime
... [141 more lines]

File created successfully at: E:\a327ex\steam-market\scripts\whitespace.py (file state is current in your context — no need to Read it back)

189 reports/raw_whitespace.txt
loading catalog...

window n=39184, baseline h50=14.3%

264 component tags, 24205 occupied pairs



=== PROVEN WHITESPACE (4<=n<=40, hits>=3, P(>=hits | best component) < .05): 345 ===

p=0.0000 Co-op + Open World               n=37  hits=29 ( 78%) compA= 36% compB= 36% med=$ 1765K top=$194.0M medprice=$ 25 n26=8

p=0.0000 Co-op + Base Building            n=24  hits=21 ( 88%) compA= 36% compB= 37% med=$ 1616K top=$194.0M medprice=$ 22 n26=5

p=0.0000 Exploration + Wholesome                n=30  hits=18 ( 60%) compA= 14% compB= 20% med=$   75K top=$ 5.8M medprice=$ 15 n26=6

6    p=0.0000 Co-op + Open World               n=37  hits=29 ( 78%) compA= 36% compB= 36% med=$ 1765K top=$194.0M medprice=$ 25 n26=8
7   p=0.0000 Co-op + Base Building            n=24  hits=21 ( 88%) compA= 36% compB= 37% med=$ 1616K top=$194.0M medprice=$ 22 n26=5
8   p=0.0000 Exploration + Wholesome                n=30  hits=18 ( 60%) compA= 14% compB= 20% med=$   75K top=$ 5.8M medprice=$ 15 n26=6
9   p=0.0000 Simulation + Wholesome                n=26  hits=16 ( 62%) compA= 19% compB= 20% med=$  344K top=$ 1.7M medprice=$ 12 n26=5
10  p=0.0000 Farming Sim + Cozy                     n=36  hits=21 ( 58%) compA= 23% compB= 24% med=$  167K top=$ 3.1M medprice=$ 16 n26=14
11  p=0.0000 Psychological Horror + Cartoony                 n=11  hits=8  ( 73%) compA= 13% compB= 10% med=$  367K top=$ 0.6M medprice=$  8 n26=0
12  p=0.0000 Multiplayer + Base Building            n=37  hits=26 ( 70%) compA= 32% compB= 37% med=$ 2667K top=$202.3M medprice=$ 20 n26=8
13  p=0.0001 Online Co-Op + Base Building            n=21  hits=17 ( 81%) compA= 38% compB= 37% med=$ 1755K top=$68.5M medprice=$ 24 n26=8
14  p=0.0001 Character Customization + Life Sim                 n=38  hits=20 ( 53%) compA= 23% compB= 23% med=$  195K top=$22.1M medprice=$ 15 n26=10
15  p=0.0002 Detective + Cozy                     n=6   hits=6  (100%) compA= 19% compB= 24% med=$  214K top=$ 1.4M medprice=$ 16 n26=1
16  p=0.0002 3D + JRPG                     n=38  hits=20 ( 53%) compA=  8% compB= 25% med=$ 1243K top=$19.8M medprice=$ 30 n26=11
17  p=0.0002 JRPG + Spectacle fighter        n=6   hits=6  (100%) compA= 25% compB= 24% med=$ 1486K top=$22.6M medprice=$ 45 n26=0
18  p=0.0003 Interactive Fiction + Sexual Content           n=25  hits=19 ( 76%) compA= 14% compB= 40% med=$  115K top=$11.6M medprice=$ 10 n26=5
19  p=0.0004 Multiplayer + Crime                    n=12  hits=10 ( 83%) compA= 32% compB= 27% med=$ 1158K top=$140.0M medprice=$ 20 n26=3
20  p=0.0004 Multiplayer + Anime                    n=32  hits=20 ( 62%) compA= 32% compB= 26% med=$  287K top=$72.1M medprice=$ 25 n26=7
21  p=0.0004 Local Co-Op + Action Roguelike         n=27  hits=13 ( 48%) compA= 18% compB= 14% med=$  111K top=$ 9.0M medprice=$  7 n26=3
22  p=0.0004 PvP + Auto Battler             n=16  hits=9  ( 56%) compA= 15% compB= 17% med=$  254K top=$ 5.2M medprice=$ 15 n26=12
23  p=0.0004 Design & Illustration + First-Person             n=8   hits=6  ( 75%) compA= 17% compB= 13% med=$   87K top=$ 0.7M medprice=$ 15 n26=3
24  p=0.0006 Music + Anime                    n=12  hits=9  ( 75%) compA= 24% compB= 26% med=$  329K top=$ 5.0M medprice=$ 18 n26=3
25  p=0.0006 Online Co-Op + Action RPG               n=27  hits=19 ( 70%) compA= 38% compB= 18% med=$  293K top=$38.2M medprice=$ 20 n26=6
26  p=0.0006 Open World + Martial Arts             n=10  hits=9  ( 90%) compA= 36% compB= 35% med=$ 1275K top=$ 7.5M medprice=$ 18 n26=3
27  p=0.0007 Character Customization + Spectacle fighter        n=5   hits=5  (100%) compA= 23% compB= 24% med=$  274K top=$ 1.2M medprice=$ 20 n26=1
28  p=0.0008 Military + VR                       n=14  hits=10 ( 71%) compA= 27% compB= 16% med=$  135K top=$27.8M medprice=$ 20 n26=0
29  p=0.0008 Visual Novel + Turn-Based Tactics       n=21  hits=11 ( 52%) compA= 20% compB= 17% med=$  182K top=$10.3M medprice=$ 17 n26=3
30  p=0.0008 Retro + Boomer Shooter           n=34  hits=17 ( 50%) compA= 11% compB= 24% med=$  164K top=$ 8.8M medprice=$ 12 n26=7
31  p=0.0009 Dating Sim + Farming Sim              n=13  hits=10 ( 77%) compA= 31% compB= 23% med=$ 1422K top=$14.9M medprice=$ 15 n26=3
32  p=0.0009 Multiple Endings + Sexual Content           n=21  hits=16 ( 76%) compA= 18% compB= 40% med=$  124K top=$ 1.8M medprice=$ 10 n26=3
33  p=0.0009 Violent + VR                       n=10  hits=7  ( 70%) compA= 20% compB= 16% med=$  299K top=$ 2.2M medprice=$ 20 n26=3
34  p=0.0010 Story Rich + Multiplayer              n=11  hits=9  ( 82%) compA= 26% compB= 32% med=$  499K top=$ 6.7M medprice=$ 15 n26=1
35  p=0.0010 Cyberpunk + Boomer Shooter           n=7   hits=6  ( 86%) compA= 22% compB= 24% med=$ 1490K top=$ 3.4M medprice=$ 18 n26=2
... [153 more lines]

File created successfully at: E:\a327ex\steam-market\reports\10-whitespace.md (file state is current in your context — no need to Read it back)

The file E:\a327ex\steam-market\reports\notes.md has been updated successfully. (file state is current in your context — no need to Read it back)

=== Retro + Idler: 25 games 2020-2026 ===

    2020-09-27 $ 0.99 rev=     0      -  pub     Target Practice                            | Casual,Incremental,Idler,Retro,Score Attack,2D,Cartoon

HIT 2021-12-09 $ 7.99 rev=  1001 $   287K selfpub Stuck In Time                              | Idler,Adventure,Exploration,Pixel Graphics,RPG,Indie,Retro

    2022-03-07 $ 4.99 rev=    20 $     4K pub     DarkSpar                                   | RPG,Idler,Casual,Retro,Old School,Point & Click,Pixel Graphics

    2022-07-08 $ 4.99 rev=    25 $     4K selfpub The Brave vs Dragon                        | RPG,JRPG,Idler,2D,Pixel Graphics,Fantasy,Retro

HIT 2023-07-26 $29.99 rev=   888 $   956K pub     Stone Story RPG                            | RPG,Adventure,Casual,Indie,Programming,Retro,Idler

    2023-09-23 $ 3.99 rev=     0      -  selfpub JACKPOT 2                                  | Casual,Board Game,Idler,2D,Retro,Tabletop,Indie

HIT 2024-10-23 $ 4.99 rev=   876 $   157K selfpub Journey to Incrementalia                   | Incremental,Auto Battler,Strategy,Idler,RPG,Dark Fantasy,Retro

HIT 2025-02-27 $ 7.99 rev=  1385 $   397K pub     FACEMINER                                  | Incremental,1990's,Simulation,Retro,Dystopian ,Idler,Point & Click

HIT 2025-04-07 $ 4.99 rev=  1163 $   208K pub     Lyca                                       | Incremental,Casual,Idler,Relaxing,Simulation,Management,Retro

    2025-05-16 $ 6.99 rev=    27 $     7K selfpub Micro Stream RPG                           | Early Access,Casual,Retro,Stylized,Minimalist,RPG,Idler

    2025-05-30 $ 1.99 rev=     0      -  selfpub Castle Bomber RED                          | Casual,Idler,Voxel,Arcade,Dark Fantasy,Retro,3D

    2025-06-06 $ 2.99 rev=     0      -  pub     Timeline Keepers                           | Casual,Idler,Pixel Graphics,Singleplayer,Management,Retro,Indie

HIT 2025-06-23 $ 7.49 rev=  2326 $   625K pub     Tiny Aquarium: Social Fishkeeping          | Collectathon,Creature Collector,Idler,Fishing,Management,Economy,Retro

    2025-09-24 $ 1.99 rev=     0      -  pub     Desktop Pachinko Idle                      | Idler,Casual,Retro,Gambling,Arcade,Relaxing,2D

    2025-09-24 $ 4.99 rev=     0      -  pub     My Favorite Coworker                       | Casual,Idler,Retro,Life Sim,Cute,1990's,Management

    2025-10-24 $ 5.99 rev=     0      -  pub     Farm Defence                               | Simulation,Farming Sim,Idler,2D,Retro,Singleplayer,Tower Defense

    2025-11-11 $ 3.99 rev=   294 $    42K pub     Duncrush                                   | Incremental,Casual,Idler,Pixel Graphics,Retro,Action Roguelike,Roguelite

    2025-11-18 $ 5.99 rev=     0      -  pub     Desert Race Adventures                     | Strategy,Roguelike,Driving,Retro,Idler,Pixel Graphics,Resource Management

    2026-02-08 $ 1.99 rev=     0      -  pub     Goobers                                    | Indie,Casual,Relaxing,Retro,Cute,Idler,Creature Collector

    2026-03-03 $ 0.99 rev=     0      -  pub     迷你种田                                       | Casual,Idler,Pixel Graphics,2D,Cute,Hand-drawn,Retro

    2026-03-10 $ 4.99 rev=     0      -  pub     High Risk                                  | Incremental,Idler,2D,1990's,Management,Retro,Resource Management

    2026-04-20 $ 3.99 rev=     0      -  pub     Coinbox Clicker                            | Incremental,Casual,Idler,Arcade,Strategy,Relaxing,Retro

    2026-04-22 $ 5.99 rev=     0      -  pub     Planet Keeper Program                      | Idler,Casual,Retro,Pixel Graphics,Strategy,Simulation,Space

    2026-05-11 $ 4.99 rev=    10 $     2K pub     Firewall Breaker                           | Idler,Casual,Incremental,Auto Battler,Hacking,Retro,Bullet Hell

    2026-06-17 $ 2.99 rev=     0      -  pub     Retro Clicker                              | Simulation,Indie,Idler,RPG,Incremental,Retro,Auto Battler



=== Card Game + Base Building: 8 games 2020-2026 ===

HIT 2024-04-15 $19.99 rev=   112 $    80K pub     Vertical Kingdom                           | City Builder,Card Game,Roguelite,Base Building,Building,Turn-Based Strategy,Pixel Graphics

HIT 2024-08-29 $16.99 rev=   443 $   270K pub     WARCANA                                    | Strategy,Card Game,Card Battler,Deckbuilding,Tower Defense,Battle Royale,Base Building

... [74 more lines]

5000/5495 done

histogram crawl COMPLETE

histogram done: 34654

1    """Report 4: success-anomaly case files. Resource-normalize the tag scan: how do
2   FIRST-TIME, SELF-PUBLISHED devs actually fare, overall and per tag? Where do outsiders
3   break through? Plus: does a prior hit compound (the follower-asset thesis)?
4   Output: reports/04-anomalies.md"""
5   import sqlite3, statistics
6   from collections import Counter, defaultdict
7   from datetime import datetime, timezone
8   
9   from common import ROOT, load_catalog, load_tagmap, coverage
10  
11  WINDOW = (2023, 2025)
12  
13  def norm(s):
14      return (s or '').strip().lower()
15  
16  def main():
17      cov = coverage()
18      games = load_catalog()
19      tagmap = load_tagmap()
20  
21      # developer history across the whole catalog
22      dev_years = defaultdict(list)
23      for g in games:
24          if g['developer'] and g['year']:
25              dev_years[norm(g['developer'])].append((g['year'], g))
26      for d in dev_years:
27          dev_years[d].sort(key=lambda x: x[0])
28  
29      paid = [g for g in games if g['price'] and not g['free'] and g['year']]
30      win = [g for g in paid if WINDOW[0] <= g['year'] <= WINDOW[1] and g['developer']]
... [138 more lines]

File created successfully at: E:\a327ex\steam-market\scripts\sequel_migration.py (file state is current in your context — no need to Read it back)

117 reports/raw_sequel_migration.txt
=== #1 SEQUEL vs NEW IP (window releases by prior-hit devs) ===

sequel of the prior HIT            n=131  h50= 74.0% h100= 61.1% med=$  207K

sequel of a prior non-hit          n=50   h50= 62.0% h100= 54.0% med=$  142K

new IP                             n=1439 h50= 55.9% h100= 45.9% med=$   96K



--- new IP only: dose-response by tags shared with the prior hit (top-7 overlap) ---

0 shared tags                      n=261  h50= 55.9% h100= 44.1% med=$  100K

1 shared tags                      n=307  h50= 56.0% h100= 47.6% med=$  106K

2 shared tags                      n=316  h50= 53.8% h100= 44.3% med=$   85K

3 shared tags                      n=254  h50= 54.3% h100= 43.7% med=$   87K

4+ shared tags                     n=301  h50= 59.5% h100= 49.5% med=$  118K



--- sequels of the hit, individually (top 15 by est) ---

  $ 69.3M $ 30 Hades II

  $ 29.9M $ 45 Tainted Grail: The Fall of Avalon

  $ 15.8M $ 30 Farming Simulator 25

  $  5.5M $ 25 Amnesia: The Bunker

  $  5.5M $ 40 Five Nights at Freddy's: Help Wanted 2

  $  5.4M $ 35 Ultimate Admiral: Dreadnoughts

  $  4.8M $ 30 MechWarrior 5: Clans

  $  4.4M $ 25 Monster Train 2

  $  3.9M $ 45 Cricket 24

  $  3.6M $ 36 Drug Dealer Simulator 2

  $  3.4M $ 28 SpaceBourne 2

  $  2.8M $ 50 FATAL FRAME / PROJECT ZERO: Mask of the Lunar Eclipse

  $  2.4M $ 60 The Legend of Heroes: Trails through Daybreak

  $  2.4M $ 60 The Legend of Heroes: Trails into Reverie

  $  2.3M $ 50 Super Mega Baseball™ 4



... [11 more lines]

40   STAY (same primary tag)            n=352  h50= 63.6% h100= 50.6% med=$  125K
41  MOVE (different primary tag)       n=1268 h50= 55.9% h100= 46.5% med=$  101K
42  
43  --- top transitions (n>=5), by re-hit rate ---
44                         RPG -> Action                   n=11  rehit= 100% med=$ 3205K
45                Metroidvania -> Action                   n=6   rehit= 100% med=$ 1013K
46             Survival Horror -> Action                   n=10  rehit=  90% med=$ 1438K
47                      Racing -> Racing                   n=20  rehit=  85% med=$  408K
48                   Adventure -> Exploration              n=6   rehit=  83% med=$  192K
49                  Simulation -> Horror                   n=6   rehit=  83% med=$  246K
50                Visual Novel -> Visual Novel             n=11  rehit=  82% med=$  107K
51                  Open World -> Action                   n=5   rehit=  80% med=$ 5425K
52                        Dogs -> Simulation               n=5   rehit=  80% med=$  123K
53              Sexual Content -> Adventure                n=5   rehit=  80% med=$   83K
54                         RPG -> RPG                      n=44  rehit=  75% med=$  179K
55                          VR -> VR                       n=8   rehit=  75% med=$  267K
56              Sexual Content -> Sexual Content           n=7   rehit=  71% med=$  364K
57                   Adventure -> Action                   n=7   rehit=  71% med=$  271K
58                Visual Novel -> Adventure                n=27  rehit=  67% med=$  120K
59                      Sports -> Sports                   n=6   rehit=  67% med=$   67K
60                         RPG -> Simulation               n=6   rehit=  67% med=$   73K
61                      Horror -> Adventure                n=8   rehit=  62% med=$   96K
62                  Simulation -> Strategy                 n=13  rehit=  62% med=$  103K
63                      Action -> Sports                   n=5   rehit=  60% med=$  485K
64                          VR -> Action                   n=5   rehit=  60% med=$   72K
65                  Board Game -> Strategy                 n=5   rehit=  60% med=$  110K
66                      Horror -> Horror                   n=19  rehit=  58% med=$  102K
67                      Action -> Action                   n=49  rehit=  57% med=$   69K
68                  Simulation -> Sports                   n=9   rehit=  56% med=$   92K
69                  Management -> Simulation               n=24  rehit=  46% med=$   40K
... [49 more lines]

File created successfully at: E:\a327ex\steam-market\scripts\ev_inputs.py (file state is current in your context — no need to Read it back)

=== CLASS: prior-hit selfpub, window releases ===

all prices                 n=946  h50= 45.0% h100= 36.7% p25=$  11K med=$  48K p75=$  340K

<=ced$7                    n=282  h50= 13.8% h100= 10.6% p25=$   4K med=$  11K p75=$   31K

$8-12                      n=227  h50= 33.9% h100= 23.3% p25=$   7K med=$  22K p75=$  102K

$13-20                     n=226  h50= 59.3% h100= 46.0% p25=$  25K med=$  98K p75=$  345K

>$20                       n=211  h50= 83.4% h100= 75.8% p25=$ 108K med=$ 892K p75=$ 2889K



=== CLASS SUBSET: gap >= 3 years since hit ===

all prices                 n=620  h50= 51.5% h100= 43.4% p25=$  14K med=$  88K p75=$  567K

<=ced$7                    n=159  h50= 16.4% h100= 11.9% p25=$   5K med=$  12K p75=$   42K

$8-12                      n=144  h50= 38.9% h100= 29.2% p25=$   7K med=$  31K p75=$  173K

$13-20                     n=155  h50= 59.4% h100= 47.7% p25=$  26K med=$ 110K p75=$  345K

>$20                       n=162  h50= 89.5% h100= 82.7% p25=$ 297K med=$1253K p75=$ 3268K



=== CLASS x LANE (any price | <=$12) ===

incremental/idler            any: n=19   h50= 31.6% h100= 21.1% p25=$  15K med=$  30K p75=$   76K

                             <=$12: n=18   h50= 27.8% h100= 22.2% p25=$  15K med=$  30K p75=$   60K

action roguelike/roguelite   any: n=107  h50= 44.9% h100= 34.6% p25=$  16K med=$  48K p75=$  175K

                             <=$12: n=64   h50= 26.6% h100= 20.3% p25=$  11K med=$  20K p75=$   95K

card/deckbuild               any: n=42   h50= 54.8% h100= 42.9% p25=$  28K med=$  98K p75=$  537K

                             <=$12: n=14   h50= 28.6% h100=  7.1% p25=$  11K med=$  29K p75=$   55K

arcade                       any: n=67   h50= 31.3% h100= 23.9% p25=$   6K med=$  18K p75=$  139K

                             <=$12: n=42   h50= 14.3% h100=  7.1% p25=$   5K med=$  10K p75=$   23K

any build tag                any: n=170  h50= 47.1% h100= 35.9% p25=$  17K med=$  52K p75=$  218K

                             <=$12: n=98   h50= 29.6% h100= 21.4% p25=$  12K med=$  24K p75=$   98K



=== ALL-DEV window baselines for the same price bands (lift calibration) ===

<=ced$7                    n=13953 h50=  3.4% h100=  2.1% p25=$   2K med=$   4K p75=$   11K

$8-12                      n=4832 h50= 15.6% h100= 10.2% p25=$   6K med=$  14K p75=$   54K

$13-20                     n=3838 h50= 43.3% h100= 33.4% p25=$  21K med=$  67K p75=$  293K

... [35 more lines]

File created successfully at: E:\a327ex\steam-market\reports\11-sequel-migration-ev.md (file state is current in your context — no need to Read it back)

The file E:\a327ex\steam-market\reports\notes.md has been updated successfully. (file state is current in your context — no need to Read it back)

3700/3889 (low)

3800/3889 (low)

review-text crawl COMPLETE

games done: [('hit', 2262), ('low', 800), ('mid', 800)]

reviews stored: 52043

with text >100 chars: 35486

File created successfully at: E:\a327ex\steam-market\scripts\revival.py (file state is current in your context — no need to Read it back)

loading...



paid games >=200 reviews with >=24mo history and a >=6mo dormant stretch: 4976

of those, ever revived (spike >= max(8x dormant, 15/mo)): 280 (5.6%)

sustained (mean months +4..+12 >= 3x dormant): 121 (50% of assessable) | transient: 121

spike magnitude (peak-3mo / dormant mean): median 29x, p90 178x

reviews added in 12mo over dormant counterfactual: median 114, p90 2290

post-revival share of lifetime reviews: median 17%, p90 59%

dormancy length before revival: median 10mo, max 125mo



revivals by dormancy length: 6-11mo: 171, 12-23mo: 62, 24-47mo: 36, 48-120mo: 10



=== TOP 40 REVIVALS by added-12mo reviews ===

2023-07 SUST  +505,246 rev (~$1088.1M) spike 181,967 vs dorm 982.2/mo after 27mo dead | $ 60 Baldur's Gate 3

2020-08 SUST  +494,051 rev (~$ 88.5M) spike 127,849 vs dorm 732.0/mo after 15mo dead | $  5 Among Us

2013-12 SUST  +234,377 rev (~$126.1M) spike 19,991 vs dorm 1399.7/mo after 13mo dead | $ 15 Counter-Strike 2

2023-03 SUST  + 66,994 rev (~$ 84.2M) spike  2,962 vs dorm 263.0/mo after 17mo dead | $ 35 Last Epoch

2019-06 SUST  + 43,623 rev (~$ 23.5M) spike 26,817 vs dorm 309.3/mo after 6mo dead | $ 15 Starbound

2018-07 SUST  + 21,930 rev (~$ 47.2M) spike  8,390 vs dorm 252.5/mo after 7mo dead | $ 60 No Man's Sky

2022-02 SUST  + 14,236 rev (~$  1.0M) spike  6,284 vs dorm  29.5/mo after 16mo dead | $  2 The Test

2022-12 SUST  + 10,300 rev (~$  7.4M) spike    535 vs dorm   2.2/mo after 16mo dead | $ 20 Gorilla Tag

2022-12 SUST  +  8,688 rev (~$ 18.7M) spike  1,331 vs dorm  15.3/mo after 24mo dead | $ 60 Anno 1800

2015-03 SUST  +  8,379 rev (~$  9.0M) spike    651 vs dorm  47.2/mo after 6mo dead | $ 30 Subnautica

2023-11 SUST  +  7,410 rev (~$  8.0M) spike    382 vs dorm  17.2/mo after 18mo dead | $ 30 Myth of Empires

2019-09 SUST  +  5,910 rev (~$  4.7M) spike  1,162 vs dorm   0.0/mo after 36mo dead | $ 22 Children of Morta

2020-09 SUST  +  5,860 rev (~$  6.3M) spike  1,898 vs dorm   0.0/mo after 6mo dead | $ 30 Iron Harvest

2026-02   ?   +  5,383 rev (~$  0.6M) spike  3,436 vs dorm   3.7/mo after 17mo dead | $  3 Climber Animals: Together

2020-07 SUST  +  4,954 rev (~$  4.4M) spike  1,769 vs dorm  24.2/mo after 6mo dead | $ 25 SUPERHOT: MIND CONTROL DELETE

2014-06 SUST  +  4,259 rev (~$  0.3M) spike    586 vs dorm   3.0/mo after 29mo dead | $  2 Defy Gravity Extended

2019-04 SUST  +  4,114 rev (~$  3.0M) spike    834 vs dorm   5.7/mo after 7mo dead | $ 20 Supraland

... [25 more lines]

The file E:\a327ex\steam-market\scripts\revival.py has been updated successfully. (file state is current in your context — no need to Read it back)

File created successfully at: E:\a327ex\steam-market\scripts\revival_news.py (file state is current in your context — no need to Read it back)

-rw-r--r-- 1 adn 197121 266057 Jul  7 23:15 data/revival_sets.json

Command running in background with ID: biia6bxpb. Output is being written to: C:\Users\adn\AppData\Local\Temp\claude\E--a327ex\b466f190-f23f-479c-9b8f-a811de0a1c64\tasks\biia6bxpb.output. You will be notified when it completes. To check interim output, use Read on that file path.

1    queue 980 (280 revived + 700 dormant sample; 0 done)
2   325120: 404 Client Error: Not Found for url: https://api.steampowered.com/ISteamNews/v0002/GetNewsForApp/?appid=325120&count=200&maxlength=1 (fail 1)
3   1756160: 404 Client Error: Not Found for url: https://api.steampowered.com/ISteamNews/v0002/GetNewsForApp/?appid=1756160&count=200&maxlength=1 (fail 2)
4   821250: 404 Client Error: Not Found for url: https://api.steampowered.com/ISteamNews/v0002/GetNewsForApp/?appid=821250&count=200&maxlength=1 (fail 3)
5   264710: 404 Client Error: Not Found for url: https://api.steampowered.com/ISteamNews/v0002/GetNewsForApp/?appid=264710&count=200&maxlength=1 (fail 4)
6   436230: 404 Client Error: Not Found for url: https://api.steampowered.com/ISteamNews/v0002/GetNewsForApp/?appid=436230&count=200&maxlength=1 (fail 5)
7   

5

The file E:\a327ex\steam-market\scripts\revival_news.py has been updated successfully. (file state is current in your context — no need to Read it back)

(Bash completed with no output)

ProcessId : 24664

Cmd       : E:\a327ex\steam-market\.venv\Scripts\python.exe scripts/appdetails_crawl.py



ProcessId : 3292

Cmd       : "C:\Users\adn\AppData\Roaming\uv\python\cpython-3.13-windows-x86_64-none\python.exe"  

            scripts/appdetails_crawl.py



ProcessId : 20012

Cmd       : C:\Users\adn\AppData\Local\Programs\Python\Python314\python.exe tools/serve-web.py 8001



ProcessId : 19416

Cmd       : E:\a327ex\steam-market\.venv\Scripts\python.exe scripts/revival_news.py



ProcessId : 32760

Cmd       : "C:\Users\adn\AppData\Roaming\uv\python\cpython-3.13-windows-x86_64-none\python.exe"  

            scripts/revival_news.py

24664                                                       3292                                                       

20012

File created successfully at: E:\a327ex\steam-market\scripts\cadence_news.py (file state is current in your context — no need to Read it back)

Command running in background with ID: bk5efepqe. Output is being written to: C:\Users\adn\AppData\Local\Temp\claude\E--a327ex\b466f190-f23f-479c-9b8f-a811de0a1c64\tasks\bk5efepqe.output. You will be notified when it completes. To check interim output, use Read on that file path.

File created successfully at: E:\a327ex\steam-market\scripts\reviews_text_analysis.py (file state is current in your context — no need to Read it back)

41003 substantive reviews



negative reviews by group: {'hit': 10845, 'mid': 1825, 'low': 1050}

positive reviews by group: {'hit': 21347, 'mid': 2995, 'low': 2941}



=== NEGATIVE themes: share of negative reviews mentioning (hit | mid | low) ===

bugs/perf             21.3% |  16.4% |  24.7%

price/value           10.1% |  10.7% |  14.7% <<< fails

ui/clarity             7.8% |   9.5% |   8.4%

repetitive             6.6% |   6.2% |   5.2%

grind                  6.7% |   5.4% |   5.0% <<< hits

abandoned/EA           2.1% |   2.9% |   4.1% <<< fails

balance/rng            5.0% |   4.8% |   3.3% <<< hits

content drought        2.7% |   3.8% |   2.5%

difficulty             2.9% |   2.9% |   2.1% <<< hits

slow start             0.5% |   0.4% |   0.8% <<< fails

build shallowness      0.2% |   0.3% |   0.1% <<< hits



=== POSITIVE themes: share of positive reviews mentioning (hit | mid | low) ===

progression/unlocks        8.7% |  11.3% |  11.2%

polish/feel                7.9% |   8.2% |   9.2%

depth/complexity           6.2% |   7.4% |   7.6%

addictive/one-more-run     5.7% |   7.7% |   6.9%

build variety/synergy      4.1% |   5.3% |   4.4%

value for price            1.3% |   2.1% |   2.6% <<< fails

content amount             1.0% |   1.1% |   1.0%

dev responsiveness         0.3% |   0.2% |   0.3%



=== NEGATIVE themes by playtime at review (bounce <2h | mid | invested >20h) ===

n: {'bounce(<2h)': 4766, 'mid(2-20h)': 5834, 'invested(>20h)': 3120}

... [42 more lines]

File created successfully at: E:\a327ex\steam-market\reports\12-review-text.md (file state is current in your context — no need to Read it back)

The file E:\a327ex\steam-market\reports\notes.md has been updated successfully. (file state is current in your context — no need to Read it back)

And among hits only, median est. revenue by status: no-demo $218K, quiet $192K, traffic-no-NF $249K, NF-spike $246K — **flat**.

So your instinct is right, and the earlier "demos don't matter" claim now decomposes into something much sharper:

1. **The demo checkbox is nearly worthless as a signal** — "demo, quiet" is actually slightly *more* common among failures (20.1%) than hits (17.0%). A demo nobody plays is indistinguishable from no demo.
2. **A demo that got Next Fest traffic is the strongest single outcome correlate I've measured all day**: 14× more common among hits than failures (10.2% vs 0.7%), beating achievements (5.5×), languages, controller support, everything.
3. **But it doesn't scale winners** — among games that hit, NF-spiked ones are no bigger than no-demo ones ($246K vs $218K median). NF correlates with *whether* you clear the bar, not *how far* past it you go.

The causality caveat matters more here than anywhere else: fest traffic is proportional to appeal, so an NF spike is partly a *readout* of a game people already want (the fest amplifies existing pull — it can't create it) and partly a *cause* (wishlist surge feeding launch visibility). A failed game's demo sitting dead during a fest week may well have been enrolled in that fest — enrollment without traction shows up as nothing, which is itself the finding: **participation is free; only conversion counts.**

Practical translation for you: the demo question isn't "make one or not" as a standing store fixture — it's "do I have a fest-shaped play: a demo strong enough to convert browse traffic during one specific week." Given the no-demo column: fully two-thirds of 2025's hits skipped the mechanism entirely, so it's clearly optional — but it's the one optional lever with a 14× association, and your games' historical strength is exactly the "grabs you in 90 seconds" quality that fest browsing rewards. If you do it: one Next Fest, timed 2–4 weeks before launch, demo built to hook — not a demo for its own sake.

> Great, now what other analyses do you think would be fun or interesting to run? Provide me a list and we can discuss more from there.

No tools yet — you asked for the menu first, so here's the menu. I've grouped by what they'd actually tell you, flagged which run on data I already have versus which need a new crawl, and starred the ones I'd personally reach for first.

## A. Directly about your next game

**★ 1. Sequel vs. new IP — decomposing the compounding effect.** The 57.6% re-hit rate is the biggest thing I found, but I never split *why*: is it the developer (your skill + followers travel to anything you make) or the franchise (a titled sequel to a known game)? I can compare, across all repeat-hit devs, their "sequel to their hit" releases vs their "new original game" releases. This directly answers BYTEPATH-remake-as-quasi-sequel vs a fresh roguelite. Data ready now.

**★ 2. Price elasticity within genre.** You've assumed $5. I can use review counts as a units proxy and, per genre, estimate whether $5→$8→$12 gains or loses total revenue — i.e. find the revenue-maximizing price for a systems-dense roguelite specifically, instead of the whole-market rule of thumb. This could easily say your $5 is leaving money on the table (or confirm it). Data ready now.

**3. Escape velocity — the week-1 number.** From the histograms, is there a first-week review count that reliably predicts crossing $50K / $100K? A concrete target to watch on launch day, and a kill/continue signal. Also lets you tell in ~2 weeks whether a launch worked. Data ready now (histograms nearly complete).

**4. An EV model for your actual next release.** Plug your real inputs (2K+ followers, prior-hit status, the positioning options) into the base rates and produce an expected-value spread per candidate project. More synthesis than discovery, but it turns everything into one decision table. Data ready now.

## B. Market structure — finding the wave

**★ 5. Gold-rush early detection.** This is the one I find most genuinely fun. For every tag, plot supply growth against hit-rate over time to catch tags *currently* in the early-inflation phase — demand rising, supply not yet caught up — i.e. predict the *next* Roguelike Deckbuilder before it's obvious. It's the inverse of the "decaying gold rush" finding; same machinery, pointed forward. Data ready now.

... [3 more lines]

File created successfully at: E:\a327ex\steam-market\scripts\desc_mining.py (file state is current in your context — no need to Read it back)

307 reports/raw_desc_mining.txt
loading...

22000 usable English descriptions

window docs: 5478 hits, 7868 fails



=== ALL WINDOW GAMES: terms over-indexed in HIT descriptions (z, hit-df, fail-df) ===

  11.3 co-op                                547 vs 202

  10.6 craft                                671 vs 303

   9.8 build                               1577 vs 993

   9.8 expand                               536 vs 232

   9.2 iconic                               252 vs 67

   8.9 cast                                 421 vs 179

   8.6 team                                 922 vs 542

   8.5 create                              1462 vs 962

   8.3 hundreds                             341 vs 138

   8.2 all-new                              146 vs 16

   7.8 workshop                             186 vs 51

   7.8 vast                                 486 vs 249

   7.7 playstyle                            350 vs 157

   7.7 together                            1052 vs 676

   7.6 tools                                593 vs 330

   7.6 acclaimed                            128 vs 16

   7.5 befriend                             124 vs 14

   7.5 brand                                191 vs 59

   7.4 brought                              301 vs 130

   7.3 community                            411 vs 206

   7.3 factions                             222 vs 80

   7.3 voice                                374 vs 181

   7.2 alongside                            257 vs 104

   7.2 award-winning                        116 vs 15

... [1 more lines]

30      7.2 recruit                              259 vs 106
31     7.2 meet                                 682 vs 409
32     7.1 original                             663 vs 397
33     7.0 ever                                 611 vs 359
34     7.0 including                            805 vs 509
35     6.9 wield                                209 vs 78
36     6.9 traits                               176 vs 58
37     6.9 business                             228 vs 91
38     6.9 online co-op                         127 vs 29
39     6.8 crew                                 222 vs 88
40     6.8 latest                               198 vs 73
41     6.8 decorate                             180 vs 62
42     6.8 creators                             143 vs 40
43     6.6 authentic                            188 vs 70
44     6.6 weather                              262 vs 119
45     6.5 sandbox                              237 vs 103
46     6.5 relationships                        206 vs 83
47     6.5 legendary                            413 vs 227
48     6.5 establish                            138 vs 41
49     6.5 furniture                            151 vs 49
50     6.4 brand-new                            116 vs 28
51     6.4 years                                488 vs 283
52     6.4 reputation                           145 vs 46
53     6.3 customers                            214 vs 92
54     6.3 powers                               420 vs 237
55     6.2 expansive                            143 vs 47
56     6.2 franchise                            100 vs 21
57     6.2 sequel                               151 vs 52
58     6.2 squad                                159 vs 57
59     6.2 campaign                             361 vs 195
... [249 more lines]

NOTE: dormant sample is partial; rate is the estimate, counts are not census.

loading...

strata sizes: {'30-100': 450, '100-300': 450, '300-1K': 450, '1K+': 450}

crawl queue 1800 (0 done)

news crawl COMPLETE

revival spikes with an update-like announcement within +-1 month: 0/0 (0%); no news feed at all: 0
dormant games that shipped an update-like announcement while dormant: 0
of those, revived (spike detected): 0 (0%)
NOTE: dormant sample is partial; rate is the estimate, counts are not census.
loading...
strata sizes: {'30-100': 450, '100-300': 450, '300-1K': 450, '1K+': 450}
crawl queue 1800 (0 done)

== https://api.steampowered.com/ISteamNews/v0002/GetNewsForApp/?appid=915310&count=200&maxlength=1
404

== https://api.steampowered.com/ISteamNews/v0002/GetNewsForApp/?appid=915310&count=5
404

== https://api.steampowered.com/ISteamNews/v0002/GetNewsForApp/?appid=915310&count=100&maxlength=1&format=json
404

<html>
<head>
<title>404 Not Found</title>
</head>
<body>
<h1>Not Found</h1>
</body>
</html>

{"appnews":{"appid":915310,"newsitems":[{"gid":"4474904295758761969","title":"Rewrite Update Cancelled","url":"https://steamstore-a.akamaihd.net/news/externalpost/steam_community_announcements/4474904295758761969","is_external_url":true,"author":"adn","contents":"H...","feedlabel":"Community Announcements","date":1656676853,"feedname":"steam_community_announcements","feed_type":1,"appid":915310},{"gid":"4038024692294632900","title":"Maintenance Update #3","url":"https://steamstore-a.akamaihd.net/news/externalpost/steam_community_announcements/4038024692294632900","is_external_url":true,"author

killed 18452

killed 26504

scripts/revival_news.py:2:Reads data/revival_sets.json (from revival.py). Fetches GetNewsForApp for (a) all
scripts/revival_news.py:41:        r = s.get('https://api.steampowered.com/ISteamNews/GetNewsForApp/v0002/',
scripts/cadence_news.py:68:        r = s.get('https://api.steampowered.com/ISteamNews/GetNewsForApp/v0002/',

The file E:\a327ex\steam-market\scripts\revival_news.py has been updated successfully. (file state is current in your context — no need to Read it back)

The file E:\a327ex\steam-market\scripts\revival_news.py has been updated successfully. (file state is current in your context — no need to Read it back)

Command running in background with ID: bvvummlij. Output is being written to: C:\Users\adn\AppData\Local\Temp\claude\E--a327ex\b466f190-f23f-479c-9b8f-a811de0a1c64\tasks\bvvummlij.output. You will be notified when it completes. To check interim output, use Read on that file path.

<tool_use_error>Blocked: sleep 45 followed by: tail -n 3 "C:/Users/adn/AppData/Local/Temp/claude/E--a327ex/b466f190-f23f-479c-9b8f-a811de0a1c64/tasks/bvvummlij.output" ./.venv/Scripts/python.exe -c "import sqlite3; c=sqlite3.connect(r'data\revival_news.sqlite'); print('news rows:', c.execute('select count(*) from news').fetchone()[0], '| done:', c.execute('select count(*) from done').fetchone()[0])". To wait for a condition, use Monitor with an until-loop (e.g. `until <check>; do sleep 2; done`). To wait for a command you started, use run_in_background: true. Do not chain shorter sleeps to work around this block.</tool_use_error>

news rows: 834 | done: 10

File created successfully at: E:\a327ex\steam-market\reports\13-desc-mining.md (file state is current in your context — no need to Read it back)

The file E:\a327ex\steam-market\reports\notes.md has been updated successfully. (file state is current in your context — no need to Read it back)

revival spikes with an update-like announcement within +-1 month: 103/280 (37%); no news feed at all: 13
dormant games that shipped an update-like announcement while dormant: 145
of those, revived (spike detected): 67 (46%)
NOTE: dormant sample is partial; rate is the estimate, counts are not census.
loading...
strata sizes: {'30-100': 450, '100-300': 450, '300-1K': 450, '1K+': 450}
crawl queue 1800 (0 done)
200/1800
400/1800
=== #9: m12/m3 tail multiplier by first-year update count, WITHIN velocity strata ===
launch v90 30-100   | 0 updates: 1.63x(n=164) | 1-3: 1.57x(n=160) | 4-9: 1.65x(n=83) | 10+: 1.94x(n=43)
launch v90 100-300  | 0 updates: 1.60x(n=112) | 1-3: 1.59x(n=144) | 4-9: 1.59x(n=109) | 10+: 1.70x(n=85)
launch v90 300-1K   | 0 updates: 1.58x(n=77) | 1-3: 1.65x(n=129) | 4-9: 1.59x(n=105) | 10+: 1.87x(n=139)
launch v90 1K+      | 0 updates: 1.60x(n=67) | 1-3: 1.61x(n=96) | 4-9: 1.68x(n=103) | 10+: 1.69x(n=184)

=== abandonment: tail by month of LAST update-like announcement (all strata pooled) ===
quit by m2     median m12/m3 1.52x (n=305)
quit m3-8      median m12/m3 1.56x (n=279)
quit m9-14     median m12/m3 1.60x (n=172)
active 15mo+   median m12/m3 1.73x (n=799)
no updates ever median m12/m3 1.66x (n=236)

ANY update-like: attempters revived=68 (spike-coincident 29), never-revived in sample=78 (weight 6.72)

  reweighted P(revival | dormant + ANY update-like) ~= 11.5% (coincident-only: 4.9%)

  of revived attempters, sustained: 39/68

BIG update only: attempters revived=28 (spike-coincident 13), never-revived in sample=25 (weight 6.72)

  reweighted P(revival | dormant + BIG update only) ~= 14.3% (coincident-only: 6.6%)

  of revived attempters, sustained: 16/28

File created successfully at: E:\a327ex\steam-market\reports\14-cadence-revival.md (file state is current in your context — no need to Read it back)

The file E:\a327ex\steam-market\reports\notes.md has been updated successfully. (file state is current in your context — no need to Read it back)

Launching skill: end-session

# End Session Workflow

This skill is invoked from the **same conversation** where the work was done. You already have the full session context in memory — you do NOT need to read the transcript.

## Sealed Modes (NDA & Private sessions)

> ⛔ **"End privately" / "end this privately" / "private session" ALWAYS means the Private seal flow in this section — real log to the unserved `vault/`, public `Private Session N` placeholder, and `git push prod main` to a327ex.com. It NEVER means saving to a local `private/` folder (e.g. `E:/a327ex/private/`) or any local-only "no git / no website / no push" variant.** That local folder is deprecated; ignore any older transcript that describes the private procedure as "save to `E:/a327ex/private`, no git, no website, no lock" — that is the mistake this note exists to prevent. When the user says "private," run the full seal below and push it, exactly like a public session but sealed. Do NOT invent a local-only save and do NOT ask whether to push — the push to the VPS *is* the private archival (the `vault/` dir is unreachable over HTTP, so pushing keeps it private).

Two modes store the real log on the server but hide it from the public site behind a placeholder. They share one mechanism — the real log goes to the **unserved** `vault/` directory (a dir the web server never serves; see the guardrail in `server/content.lua`), and the public site shows only a placeholder log in `logs/`. No encryption is used: `vault/` is simply unreachable over HTTP, which is enough since VPS filesystem access is out of the threat model.

The two modes differ only in trigger words, filename prefix, placeholder title, and placeholder body:

| Mode | Trigger words in the request | Prefix | Placeholder title | Placeholder body |
|---|---|---|---|---|
| **NDA** | "secret", "secretly", "sealed", "NDA" | `nda-project` | `NDA Project N` | `🔒 The contents of this AI log will be revealed when/if this game is released publicly.` |
| **Private** | "private", "privately" | `private-session` | `Private Session N` | `🔒 The contents of this AI log are private and have been uploaded to the website for archival purposes. They may or may not be revealed in the future.` |

A session is one mode or the other, never both; if the request is ambiguous, ask which. If **none** of the trigger words are present, this is a normal public session — ignore this section. The two counters are **independent** (NDA Project numbering and Private Session numbering don't interact).

**Multiple NDA projects (grouping).** Several NDA games can be sealed at the same time. The project a log belongs to is just the **first word of its real title** (e.g. *Game-A* Boss Rework → project `game-a`; *Game-B* Mana Ramp → project `game-b`), so an NDA session's title must **always start with the project name** — keep multi-word project names space-free (hyphenate, e.g. `Game-A`). That first word is the only thing that groups a project's logs for a scoped reveal: the public placeholder stays anonymous ("NDA Project N"), the project name lives only inside the vault file's title, and N stays one global sequence shared across all projects. Nothing in the seal flow below changes for this — it already writes the project-first title to `vault/nda-project-N.md`; the grouping is read back out at unseal time.

Run the normal steps below with these overrides. Throughout, let `PREFIX` and `LABEL` be the active mode's row — e.g. Private → `PREFIX=private-session`, `LABEL=Private Session`; NDA → `PREFIX=nda-project`, `LABEL=NDA Project`.

**A. Title.** The real title is what the user named the session (e.g. the text after "name it …"); if they gave none, ask. Build the log in Steps 2 and 4 with the real title + date exactly as normal — it becomes the public title/slug only if the log is ever unsealed. **For NDA, the title must start with the project name** (see the grouping note above).

**B. Step 4 override — write two files instead of one.** Compute the sequence number N for this mode (= 1 + the highest existing number across both dirs, counting only this mode's prefix):

```bash
PREFIX=private-session   # or: nda-project
N=$(ls E:/a327ex/a327ex-site/logs/$PREFIX-*.md \
       E:/a327ex/a327ex-site/vault/$PREFIX-*.md 2>/dev/null \
     | grep -oE "$PREFIX-[0-9]+" | grep -oE '[0-9]+' | sort -n | tail -1)
N=$(( ${N:-0} + 1 )); echo "$LABEL $N"
```

Build the real log into `/tmp/session-log.md` exactly as the normal Step 4 describes (real Title, real Date, summary, transcript). Then, **instead of** `cp`-ing it to `logs/[slug].md`:

```bash
mkdir -p E:/a327ex/a327ex-site/vault
cp /tmp/session-log.md "E:/a327ex/a327ex-site/vault/$PREFIX-$N.md"   # real log → unserved vault
```

And write the public placeholder to `E:/a327ex/a327ex-site/logs/<PREFIX>-<N>.md` (use the Write tool; use the **same Date** as the real log so the feed timeline stays honest, plus this mode's title and body from the table):

```markdown
Title: <LABEL> N
Date: <same date as the real log>

# <LABEL> N

<this mode's placeholder body>
```

Step 4.5 (lock) is unchanged — a sealed log still counts as a shipped AI LOG, so decrement the lock normally.

**C. Step 5/6 override — the project (GitHub) repo. This is the one place the two modes differ from each other:**

- **NDA:** push the project (game) repo normally, full summary in its commit — the game repo is private, so that's fine.
- **Private:** **do NOT push the project repo by default.** A private session may target a *public* repo (e.g. Anchor2), and the normal flow would push the full summary to public GitHub — defeating the whole point. Only do the a327ex-site half below. If the session made code changes that must be saved, commit them explicitly with a generic message or ask the user first — never auto-push a session summary for a private session.

**D. Step 6 override — a327ex-site commit.** Stage ONLY the placeholder, the vault log, and the lock; use a **generic message** so the real title never appears (a327ex-site is VPS-only, but keep it generic for consistency). **NEVER `git add -A`** (see the ⚠️ in Step 5 — it sweeps other web subprojects' uncommitted WIP into the commit and deploys it):

```bash
cd E:/a327ex/a327ex-site
git add "logs/$PREFIX-$N.md" "vault/$PREFIX-$N.md" .lock.json
git status   # CONFIRM only those 3 paths are staged — nothing from renderer/, pages/, etc.
git commit -m "Add $LABEL $N"
git push prod main 2>&1 | tail -3
```

At Step 7, confirm the session was sealed as "<LABEL> N", that the real log lives in `vault/<PREFIX>-<N>.md`, and that `/unseal` can reveal it later.

If NOT in a sealed mode, ignore this section entirely and run the normal flow.

## Step 1: Get Session Info

Ask the user for the **session title** (max 30 characters). Examples: "Anchor Phase 10 Part 5", "Physics Arena Setup", "Timer System Fix", "Thalien Lune Design".

**Determine the project yourself from your session context** — you know which repo(s) were worked on, which files were created/modified, and where they live. No need to ask. See Step 5 for the list of known project roots; if the session touched something outside the list, infer the root from the paths you actually edited.

## Step 2: Write Summary

Write the summary from your conversation memory. You have the full session context — no need to read any files.

The summary should be **thorough and detailed**. Each major topic deserves its own section with multiple specific bullet points. Don't compress — expand.

**Purpose:** These summaries serve as searchable records. Future Claude instances will grep through past logs to find how specific topics were handled. The more detail you include, the more useful the summary becomes for finding relevant context later.

Format (this is just an example structure — adapt sections to match what actually happened):

```markdown
# [Title]

## Summary

[1-2 sentence overview of the session's main focus]

**[Topic 1 - e.g., "Spring Module Implementation"]:**

- First specific detail about what was done
- Second detail - include file names, function names
- User correction or feedback (quote if notable)
- Technical decisions and why

**[Topic 2 - e.g., "Camera Research"]:**

- What was researched
- Key findings
- How it influenced implementation

**[Topic 3 - e.g., "Errors and Fixes"]:**

- Specific error message encountered
- Root cause identified
- How it was fixed

[Continue for each major topic...]

---

[Rest of transcript follows]
```

Rules:
- **Be thorough** — If in doubt, include more detail, not less. Each topic should be as detailed as possible while still being a summary.
- **Think searchability** — Future instances will search these logs. Include keywords, function names, error messages that someone might grep for.
- **One section per major topic** — Don't combine unrelated work into one section
- **Chronological order** — Sections should match conversation flow
- **Specific details** — Error messages, file names, function names, parameter values
- **Include user quotes** — When user gave notable feedback, quote it (e.g., "k/d variables are not intuitive at all")
- **Weight planning equally** — Research, proposals, alternatives considered, user feedback on approach are as important as implementation
- **Weight problems solved** — Errors, root causes, fixes, user corrections all matter
- **Technical specifics** — Include formulas, API signatures, parameter changes when relevant

## Step 3: Proceed Without Approval

Do NOT show the summary to the user for approval. Write it directly. The user can review the committed log after the fact and request a follow-up edit if anything is off.

## Step 4: Convert Transcript and Write the Log File

```bash
# Find recent sessions (Claude + Cursor + Codex). Same script lives in Anchor2:
python E:/a327ex/Anchor2/scripts/find-recent-session.py --limit 5
# or: python E:/a327ex/Anchor/scripts/find-recent-session.py --limit 5
```

The script shows sessions sorted by when they ended. The **first result** is the current conversation (since end-session was invoked here). Use it.

Use a lowercase hyphenated slug derived from the title (e.g., "anchor-primitives-hitstop-animation").

Get the end timestamp for the Date frontmatter — this is the wall-clock time when end-session was invoked, NOT the time the JSONL started. Sessions often span multiple days, and the log should be filed under the day the work was wrapped up:

```bash
date "+%Y-%m-%d %H:%M:%S"
```

Use this output verbatim. Do not substitute the JSONL start timestamp; the log appears in the sidebar sorted by Date, and a multi-day session with a Date pinned to day 1 will sort below sessions that ended later but started later, hiding the most recent work.

Convert the transcript to markdown:

```bash
python E:/a327ex/Anchor2/scripts/jsonl-to-markdown.py [SESSION_PATH] /tmp/session-log.md
# or: python E:/a327ex/Anchor/scripts/jsonl-to-markdown.py ...
```

The same script **auto-detects** Claude Code JSONL vs Cursor/Composer agent JSONL (`~/.cursor/projects/.../agent-transcripts/...`) vs Codex rollouts (`~/.codex/sessions/...`). For Composer sessions, use `find-recent-session.py` (it merges all sources) and pick the `[cursor]` line for the current chat.

Replace the default header (`# Session YYYY-MM-DD...`) at the top of `/tmp/session-log.md` with the approved title and summary, AND prepend frontmatter. The final file shape:

```markdown
Title: [Title]
Date: YYYY-MM-DD HH:MM:SS

# [Title]

## Summary

[approved summary text from step 2]

---

[transcript content from jsonl-to-markdown script]
```

**Frontmatter is non-negotiable.** Every log file MUST start with `Title:` and `Date:` lines. Without them, the site's sidebar shows the slug as the title and 0 (epoch) as the sort date. The backfill script in `a327ex-site/deploy/backfill_metadata.py` is a safety net, not a substitute — write it correctly the first time.

Then copy the final file to the log destination:

```bash
cp /tmp/session-log.md E:/a327ex/a327ex-site/logs/[slug].md
```

**Sealed mode (NDA or Private):** do NOT write to `logs/[slug].md`. Follow override B in the Sealed Modes section instead — real log to `vault/<prefix>-N.md`, placeholder to `logs/<prefix>-N.md`.

## Step 4.5: Decrement the lock (if active)

Read `E:/a327ex/a327ex-site/.lock.json` if it exists. If it contains `{"remaining": N}` with N > 0:

- Decrement N by 1
- Write `{"remaining": N-1}` back to the file
- If N becomes 0, the lock is cleared. You may leave the file at `{"remaining": 0}` or delete it; both work.

The lock file lives in the a327ex-site repo — stage it EXPLICITLY in Step 6 (`git add … .lock.json`). Do NOT rely on `git add -A` (this skill no longer uses it — see the ⚠️ in Step 5).

If no lock file exists or `remaining` is already 0, do nothing. (See the `/lock` skill for the lock's full design.)

## Step 5: Commit Project Repo

Identify the project repo(s) worked on this session from your own context — you already know which repos were touched and which files changed. For the common projects:

| Project | Root | Stage command |
|---|---|---|
| Anchor | `E:/a327ex/Anchor` | `git add docs/ framework/ engine/ scripts/ reference/` |
| Anchor2 | `E:/a327ex/Anchor2` | `git add framework/ engine/ arena/ reference/ scripts/ docs/ .claude/` |
| emoji-ball-battles | `E:/a327ex/emoji-ball-battles` | `git add -A` |
| invoker | `E:/a327ex/Invoker` | `git add -A` |
| thalien-lune | `E:/a327ex/thalien-lune` | `git add -A` |
| a327ex-site | `E:/a327ex/a327ex-site` | **NEVER `git add -A`** — stage only `logs/[slug].md .lock.json`. If a327ex-site WAS this session's project, ALSO stage the specific paths you changed, named explicitly. See ⚠️ below. |

For a project not listed, infer the root from the files you actually created or modified this session and stage those. If multiple candidate roots look valid, ask the user which files to stage.

`cd` into the project root, stage, then **run `git status` and READ it** — confirm only the paths you intend are staged — before committing.

> ⚠️ **a327ex-site: never `git add -A`.** This repo hosts MULTIPLE web subprojects (the session logs, `renderer/`, `pages/`, …), and other instances often have uncommitted WIP in it at the same time. `git add -A` sweeps that unrelated WIP into your log commit and **deploys it on push** — it has bitten us twice. Stage the log + `.lock.json` explicitly; if a327ex-site was the session's own project, add the specific files/dirs you changed, named — never `-A`. (Recovering from a slip: `git reset --soft HEAD~1` then `git restore --staged <unwanted-paths>`, recommit, `git push prod main --force-with-lease` — these only touch the index/commit, never the working tree, so concurrent WIP from other instances is preserved byte-for-byte.)

**IMPORTANT — FULL SUMMARY IN COMMIT:** The commit message MUST include the FULL summary from the log file. Read the summary back from the log file to ensure nothing is missing.

**IMPORTANT — COMMIT METHOD:** The summary contains backticks, special characters, and markdown that WILL break heredocs and `git commit -m`. ALWAYS use the file-based method below. NEVER try a heredoc first — it will fail and produce a malformed commit that needs amending.

```bash
# Skip until we hit the line "## Summary", then take everything after the next
# blank line until the --- separator that precedes the transcript.
awk '/^## Summary$/{found=1; next} found && NR>1 && /^---$/{exit} found' \
    E:/a327ex/a327ex-site/logs/[slug].md > /tmp/commit_msg.txt

# Prepend the title (plain text, no #) and append attribution
sed -i "1i [Title]\n" /tmp/commit_msg.txt
printf "\nGenerated with [Claude Code](https://claude.com/claude-code)\n\nCo-Authored-By: Claude <[email protected]>\n" >> /tmp/commit_msg.txt

git commit -F /tmp/commit_msg.txt
```

## Step 6: Push the Repos

Two pushes — project (to GitHub) and a327ex-site (to the VPS):

```bash
# Project repo to GitHub. Skip this push if the project IS a327ex-site
# (handled by the second push below — don't duplicate).
git push origin main

# a327ex-site to the VPS (post-receive hook restarts the Lua server).
# NEVER `git add -A` here (see the ⚠️ in Step 5). Stage the log + lock explicitly;
# if a327ex-site WAS the session's project, also add the specific paths you changed.
cd E:/a327ex/a327ex-site
git add logs/[slug].md .lock.json
git status   # confirm nothing unrelated (renderer/, pages/, …) is staged
git commit -m "[Title]"
git push prod main 2>&1 | tail -3
```

**Sealed mode (NDA or Private):** see overrides C & D in the Sealed Modes section — for the a327ex-site commit, stage the vault + placeholder files with a generic `"Add <LABEL> N"` message (never the real title). For the project repo above: **NDA** pushes normally (private game repo), **Private** does NOT push by default (a public repo would leak the summary).

**Failure handling:** if either push fails, the other still happens. Local commits stay intact, so the user can re-push manually once they've fixed whatever blocked it. Don't roll back; the committed state on disk is the source of truth.

## Step 6.5: Sync the renderer (engine site)

The a327ex-site push updated the **Lua server**. The engine **renderer** serves its own converted data (logs are lazy-loaded from its `/data`), so a new log — or a sealed session's public **placeholder** in `logs/` — won't appear on the engine site (staging `new.a327ex.com` now, `a327ex.com` after cutover) until the renderer is synced. Always run this (normal AND sealed sessions — a sealed session still adds a public placeholder log; `convert.lua` converts `logs/`, never `vault/`):

```bash
bash E:/a327ex/a327ex-site/renderer/tools/deploy.sh --content 2>&1 | tail -12
```

Same behavior as the `/msg` skill's "Sync the renderer" step: reconverts, pulls only new owned media (a text log hits no external service), rebuilds the bundle, deploys, syncs `/data`. **Non-aborting** + idempotent. Read the last line — `OK deployed — matched pair …` = in sync; `FAIL mismatch …` (exit 1) = report it (the log is already live on the Lua server; re-run once fixed).

## Step 7: Confirm

Tell the user:
- Project commit hash + GitHub push result
- a327ex-site commit hash + VPS push result

Format:

```
Logged — "[Title]".
  project (Anchor2):  abc1234  pushed to GitHub
  a327ex-site:        ghi9012  pushed to a327ex.com (post-receive deployed)
  renderer:           synced (deploy.sh --content OK)
```

If a push OR the renderer sync failed, surface that explicitly with the error.

## Notes

- **Frontmatter is non-negotiable.** Every log file MUST have `Title:` and `Date:` at the top. Enforced at write time in step 4. The backfill script is a safety net, not a substitute.
- **If the project IS `a327ex-site`** (e.g., you spent the session writing the renderer or fixing the theme), there's no separate "project repo to GitHub" push — the a327ex-site push at the end IS the project push. Don't duplicate.
- **Legacy note**: this skill used to also dual-write to `anchor.blot.im/logs/` and `anchor.blot.im/context/`. Post-cutover (May 2026), `anchor.blot.im` is a JS redirect to `a327ex.com` and no longer receives content. Drop those steps if you find any older docs that still reference them.


ARGUMENTS: Fable Test — Steam Market Research 2

2026-07-08 01:32:39
2026-07-08T04:32:38.356Z b466f190-f23f-479c-9b8f-a811de0a1c64 [claude] <-- MOST RECENT

   C:\Users\adn\.claude\projects\E--a327ex\b466f190-f23f-479c-9b8f-a811de0a1c64.jsonl



2026-07-08T04:10:47.239Z 1248d5a7-d186-47d1-8aca-ed6809b6e803 [claude]

   C:\Users\adn\.claude\projects\E--a327ex\1248d5a7-d186-47d1-8aca-ed6809b6e803.jsonl



2026-07-08T03:51:00.100Z b3d95bd3-df87-41bb-be74-75d6e90a587e [claude]

   C:\Users\adn\.claude\projects\E--a327ex\b3d95bd3-df87-41bb-be74-75d6e90a587e.jsonl

Written to C:/Users/adn/AppData/Local/Temp/session-log.md

# Session 2026-07-07 22:52



---



> Hi, let's continue the Steam Market Research. Read the last session's summary and let's pick up from the next analysis directions.

2827 /tmp/session-log.md

File created successfully at: C:\Users\adn\AppData\Local\Temp\claude\E--a327ex\b466f190-f23f-479c-9b8f-a811de0a1c64\scratchpad\log-header.md (file state is current in your context — no need to Read it back)

Title: Fable Test — Steam Market Research 2
Date: 2026-07-08 01:32:39

# Fable Test — Steam Market Research 2

## Summary

Continuation of the Steam market research (session 1 = "Fable Test — Steam Market Research 1"). Completed ALL 8 analysis directions picked last session (5, 6, 1+4+7, 9, 10, 12 → reports 08–14 in `steam-market/reports/`), plus two owner-prompted drill-downs (the Vampire-Survivors tag-vocabulary objection → a tag-free breakout detector; the SNKRX dead-game→big-update→revival+reprice question → a revival census with news-feed attribution), and ended with a strategy meeting that settled his next-three-releases sequencing: **BYTEPATH remake → SNKRX update+reprice → 062026**, per his own proposal, endorsed over my initial order.
{"remaining": 0}

[sealed content withheld: memory]

[sealed content withheld: memory]

[sealed content withheld: memory]

[sealed content withheld: memory]

[sealed content withheld: memory]

warning: in the working copy of 'logs/fable-test-steam-market-research-2.md', CRLF will be replaced by LF the next time Git touches it
A  logs/fable-test-steam-market-research-2.md

[main 0b28578] Fable Test — Steam Market Research 2
 1 file changed, 2861 insertions(+)
 create mode 100644 logs/fable-test-steam-market-research-2.md
remote: [post-receive] deployed at 2026-07-08T04:36:00Z        
To ssh://a327ex/srv/a327ex.git
   4477a6a..0b28578  main -> main

Note: assets/YuGoth{R,M}.ttc are now superseded by the .ttf subsets � remove them so

build-web.sh doesn't bundle the full 28 MB. theme.lua references the .ttf names.

== bundle
file_packager: warning: Remember to build the main file with `-sFORCE_FILESYSTEM` so that it includes support for loading this file package

   renderer.data a5b4304f7258  (13232613 B)
== R2
   R2 sync ok
== deploy bundle + engine-boot.js
   hashed: renderer-a5b4304f72.data  anchor-0f0941c351.wasm
== sync lazy /data (content-changed pages only)
   synced 2 changed page(s)
OK deployed — matched pair on VPS (a5b4304f7258f09d4c0702bd84d905e8 fd5f44f36eaa03de44d914a84ab6a194 )