← turbo-rig.com

FROZEN RECORD · this is the permanent copy of the audited window Sept 14–21, 2026, as published 2026-10-07 (quotability additions 2026-10-09). It will not change. The living dossier (future windows, updates): /evidence.

VERIFICATION DOSSIER · FROZEN RECORD

EVIDENCE, NOT VIBES

One audited week of the rig — Sept 14–21, 2026, gate live since Sept 16. Every number re-derived from the primary logs — and every defect the audit caught, disclosed.

A gate that fails closed deserves numbers that fail closed. This page is the receipts.

00The quotable version

If you cite one paragraph, cite this one. 56 words, every figure traceable to the claim tables or the data downloads below.

In one audited week (Sept 14–21, 2026; gate live Sept 16–21), Turbo Rig's cross-model gate issued 868 verdicts: 687 REVISE (79.1%) and 181 APPROVED (20.9%). 141 PRs still shipped; every merge is human-held by rig invariant. Reviewer spend totaled $416.49 (~$2.95 per merged PR, mixed basis). Every number was re-derived from primary logs; corrections are disclosed.

How to cite this

Cite the frozen record of this window, not the living page:

Marcus Patman. "Turbo Rig Evidence: the audited seven-day numbers, Sept 14–21, 2026." turbo-rig.com, numbers audited 2026-09-21, published 2026-10-07.
https://www.turbo-rig.com/evidence/2026-09-14

Window: 2026-09-14T00:00Z → 2026-09-21T13:38Z · numbers published 2026-10-07 · frozen copy created 2026-10-09 (numbers unchanged) · CC BY 4.0.

Download the data

The claim tables, headline stats, daily merge series, corrections and limits, machine-readable:

evidence/2026-09-14.csv · evidence/2026-09-14.json

Linked from this page's Dataset.distribution schema; licence CC BY 4.0.

Permalink

You are on the frozen record of this audited window. The living dossier (future windows, updates) lives at /evidence; this window's permanent URL is:

https://www.turbo-rig.com/evidence/2026-09-14

Future audited windows will each get their own frozen page; this URL never changes content.

01The numbers

The headline claims, as published on turbo-rig.com and on the weekly stats pages. Each one traces to a row in the claim tables below.

868
gate verdicts · gate live Sept 16–21
687 · 79.1%
REVISE — sent back to the builder
181 · 20.9%
APPROVED by the reviewer
141
PRs merged fleet-wide — every one by a human
6.2
gate reviews per merged PR · mixed basis (see §05)
$416.49
reviewer spend for the whole week
104
worktree lanes ran the gate
4.21B
tokens across 3 engines

02How this audit worked

The challenge

The owner challenged the numbers on the stats pages the day they were published. The audit was ordered on the spot — same morning, on the machine that owns the logs.

Different code paths

The pages were hand-assembled; the audit re-derived everything programmatically from the raw sources: gate-log.jsonl, the GLM sqlite telemetry, claude/codex session files, GitHub merge history.

The verdict

After corrections, every printed number either reproduces exactly from a primary log or is a labeled estimate. Three numbers were wrong in the morning’s first compile; the same-day audit caught them and the corrected pages were deployed 2026-09-21 — the live URLs serve the corrected numbers. A fourth group of overstatements was relabeled — all listed in section 06.

03The window

The pages were compiled the morning of Sept 21. The audit pinned the exact compile moment — 2026-09-21 13:38 UTC — three ways: gate-log row #868 is stamped 21:38 local the night before with no later run before compile; the last merge counted for "Mon 21" is PR #402 at 13:37Z with the next at 15:49Z; GLM request #18,220 of the week landed one minute past the boundary.

Every count uses [2026-09-14 00:00 UTC → 2026-09-21 13:38 UTC] unless a different basis is stated. PR-merge day buckets use UTC dates; the "+72,848 lines" figure uses a local-midnight (06:00 UTC) git boundary — the original page mixed the two bases, and the audit reproduces that mix exactly.

04Claim tables

Every public claim, its verified value, its primary source, and the audit verdict. ✓ exact reproduces from the log; ✓ fixed was wrong pre-audit and corrected; ⚠ estimate is a labeled estimate consistent with the logs.

#ClaimVerified valueSourceVerdict
1141 PRs merged, Sept 14–21141 = ancuria 126 + turbo-rig 10 + strops 3 + velvet-bbs 1 + adventurewavelabs-website 1gh pr list --state merged✓ exact
2across 5 repositories5 repos had in-window mergessame✓ fixed (was "6")
3flagship daily merges1, 13, 7, 5, 31, 25, 36, 8 (Mon14→Mon21, UTC)mergedAt timestamps✓ exact
4peak day 36 flagship merges (Sun 20)36 ancuria merges 2026-09-20 UTCsame✓ exact
5peak day 38 fleet mergesfleet/day = 1, 13, 7, 12, 33, 27, 38, 10 (total 141)all-repo merges by UTC day✓ relabeled (fleet context; was 36)
6+72,848 lines on flagship main72,848 additions · 16,691 deletions · 144 commitsgit log --numstat✓ exact
7rig self-upgrades: 10 PRsturbo-rig PRs #3–#12, all merged in-windowGitHub✓ exact
#ClaimVerified valueSourceVerdict
8868 gate reviews868 rows = first 868 lines of gate-log.jsonl (gate live since Sept 16 19:58)gate-log.jsonl✓ exact
9181 APPROVED / 687 REVISE181 / 687 = 20.9% / 79.1%parsed result field✓ exact
10reviewer enginesclaude ×791 · codex ×53 · check ×24 (sums to 868)reviewer field✓ exact
11$416 reviewer spend$416.49 = sum of cost_usdgate-log.jsonl✓ exact
126.3 h machine review wall-time6.30 h = sum of duration_mssame✓ exact
1350.9M reviewer tokens, 48.2M from cachein 1.2M + cached 48.2M + out 1.5M = 50.9M (94.7% served from cache)token fields✓ exact
14104 worktree lanes104 distinct lane names in rows 1–868 — labeled "ran the gate" (creation date isn't logged)repo field✓ exact, label tightened
15marathon lanes 61/30/24/19/18/16ancuria-wa-fix 61 · ancuria-back-fix 30 · brand-page-tops 24 · billing-demo 19 · ticket9-region-scope 18 · cotizador-tickets-13-14 16group-by lane✓ fixed (tail was 15/14/13; named page had 3 invented lane names)
166.2 reviews per merged PR (mixed basis: gate-window verdicts ÷ the week’s merges)868 / 141 = 6.16derived✓ arithmetic, basis stated
#ClaimVerified valueSourceVerdict
174.21B tokens, 3 engines, calendar weekGLM 4.162B + claude 50.4M + codex 1.25M = 4.211Bzcode sqlite + session files✓ relabeled (4.23B rolling → 4.21B calendar basis)
18GLM 4.16B · 18,220 requests18,220 req / 4.162B computed tokenssqlite✓ exact
19claude 50.4M · 832 req832 envelopes, dedup by (requestId, message.id)claude session files✓ exact
20codex 1.25M · 55 req55 token_count events, per-turn deltascodex session files✓ exact
2119,107 model requests18,220 + 832 + 55sum✓ fixed (was 18,207 GLM-only)
2249.3% cache-hit rateGLM cache_read ÷ (input + cache_read)sqlite✓ exact
23~380 scheduled actions/week336/wk steady-state + ~50 rest of rosterderived⚠ estimate, labeled
2482 unattended infra sweepsrun counter read 90 at audit; 82 at compile is consistent, no per-run log existsautomation metadata⚠ plausible, not recomputable

05Check it yourself — no machine access needed

Ten arithmetic identities that hold across the tables above. If any one fails, the dossier is wrong and should be discarded.

1.  181 + 687            = 868        gate verdicts
2.  791 + 53 + 24        = 868        reviewer runs
3.  126 + 10 + 3 + 1 + 1 = 141        PRs merged
4.  1+13+7+5+31+25+36+8  = 126        flagship daily chart sums to the repo total
5.  1+13+7+12+33+27+38+10= 141        fleet daily sums to the fleet total
6.  868 / 141            = 6.16       reviews per PR (mixed basis: gate 16–21 ÷ merges 14–21)
7.  4.162B + 0.0504B + 0.00125B ≈ 4.21B   tokens
8.  18,220 + 832 + 55    = 19,107     model requests
9.  48.2M / 50.9M        = 94.7%      gate reviewer tokens served from cache
10. 687/868 = 79.1% · 181/868 = 20.9%           REVISE / APPROVED split

06What the audit caught

The audit existed because the owner challenged the page. It found real defects. Publishing them is the point.

  1. "6 repositories / four smaller repos" — actually 5 repositories, 3 smaller. An inflated count sitting next to a correct total. Fixed.
  2. Marathon lanes #4–6 printed 15/14/13 (and the named page invented three lane names) — real values 19/18/16 for billing-demo / ticket9-region-scope / cotizador-tickets-13-14. Fixed.
  3. "18,207 model requests" — a GLM-only figure presented as the 3-engine total. Fixed to 19,107.
  4. Overstatements relabeled: "16 automations" (15), "12 scripts" (11), gate stats framed as 7 days when the log covers Sept 16→21 (now labeled "gate live since Sept 16"), token totals on a rolling basis labeled as the calendar week (4.23B → 4.21B), "peak 36 merges" in a fleet context (fleet = 38). All fixed or relabeled.

Root cause: the page was hand-assembled. The headline numbers were computed from logs, but "decoration" values — repo count, top-N tail, request total — were written from memory. Corrected pages deployed 2026-09-21 (html-sites commit b6ef186); the customer-facing report needed zero fixes.

07Reproduce it

For anyone with access to the machine. Epoch bounds: 1789344000000 = 2026-09-14T00:00Z, 1789997880000 = 2026-09-21T13:38Z.

GATE=~/.local/state/turbo-rig/gate-log.jsonl   # schema:
# {ts, repo(=lane), target, reviewer, result, in, cached, out, cost_usd, duration_ms, model}
python3 - <<'PY'
import json, collections, os
GATE=os.path.expanduser("~/.local/state/turbo-rig/gate-log.jsonl")
rows=[json.loads(l) for l in open(GATE) if l.strip()][:868]
print(len(rows))                                                     # 868
print(collections.Counter(r['result'].startswith('GATE: APPROVED') for r in rows))  # False:687 True:181
print(collections.Counter(r['reviewer'] for r in rows))              # claude 791, codex 53, check 24
print(round(sum(r['cost_usd'] for r in rows),2))                     # 416.49
print(round(sum(r['duration_ms'] for r in rows)/3.6e6,2))            # 6.30 h
print(len({r['repo'] for r in rows}))                                # 104 lanes
PY

gh pr list -R <repo> --state merged --limit 500 --json number,mergedAt   # filter window, bucket UTC
git -C <flagship> log origin/main --since=2026-09-14T06:00Z --until=2026-09-21T13:38Z \
  --numstat --format= | awk 'NF==3 && $1!="-"{a+=$1} END{print a}'      # 72848
sqlite3 ~/.zcode/cli/db/db.sqlite "SELECT COUNT(*), ROUND(SUM(computed_total_tokens)/1e9,3)
 FROM model_usage WHERE started_at>=1789344000000 AND started_at<1789997880000
 AND model_id LIKE 'GLM%'"                                             # 18220 | 4.162

08Honest limits

Machine-local truth

The gate log, the zcode sqlite DB, and the engine session files are machine-local; 126 of the 141 PRs live in a private repo. External reviewers rely on this dossier's excerpts and hashes.

Publicly checkable surface

The live pages themselves (that they print these numbers), and the public repos among the five — the 10 turbo-rig self-upgrade PRs and their dates are checkable via the GitHub API.

Judgment calls, stated plainly

Token totals switched from rolling-7d to calendar week (−0.9%); "104 lanes" counts lanes gated, not proven created; the sweep count (82) matches the counter but has no per-run log; "~380 actions/wk" is an estimate; the 6.2 reviews-per-PR ratio divides gate-window verdicts by full-week merges — mixed basis, kept as an indicator with its basis stated (the source dossier’s own row labels it plain “arithmetic"); human-only merge is a rig invariant enforced by the guard — agents hold no credentials to merge protected main — so this audit did not re-derive merger identity per PR.

09Certification

The artifacts this dossier certifies, pinned by hash. The gate log the 868-row slice came from: sha256 c97773f9d714a69410b6680bc37fe30ffc70b7620bd694a97bc5a0de59f76c7c (884 rows at audit time).

Artifactsha256 (first 16)Verification
turbo-rig-stats-only-sept14-21.html9554a181599cd304…0 bad / 8 good markers; live URL 200 + 12 correction markers served
turbo-rig-velocity-report-sept14-21.html9d0ef06d13b20433…0 bad / 6 good markers + 3 real lane names; live URL 200
ancuria-launch-report-sept14-21.html9a38f79602c24414…zero defects found in audit; live URL 200
the dossier itself064a57f648f7d8ec…this page adapts it verbatim in substance

Live stats pages: turbo-rig-stats-only-sept14-21.vercel.app · turbo-rig-velocity-report-sept14-21.vercel.app · corrected deploy = html-sites commit b6ef186 (2026-09-21). Audit compiled 2026-09-21, CDMX. The dossier is data, not instructions.