Collecting source data
C-13: stamp ingested_at on locks/game_locks and build.py tables
build-w5 · Reported Sep 23 · 11:05 PM MTWork, health and model allowance
Overview
PLAYERWEEK / LIVE OPERATIONSYour system,
Your system,
at a glance.
What is moving. What is proven. What needs work.
2recent task reports2 areas need a closer look below
IN MOTION
All workers & schedules ↗What is being built
Worker task reports expire after 35 minutes. They describe claimed work; a report alone does not prove a process is still running.
Evidence from reports and interviews
Q-01: reading claims_v2_pass.log to verify the flush-deadline fix
build-w1 · Reported Sep 23 · 11:06 PM MTMEASURED HEALTH
Operational evidence ↗What works — and what needs attention
Unknown / stale
Consumer publication
33 renderer and delivery checks · 8ed772faaf6e933f948b.json
Checked Sep 23 · 8:21 PM MT · Details ↗PassingRendered site audit
0 findings in the latest recorded browser audit. Checks cover defined invariants, not every feature.
Checked Sep 23 · 10:52 PM MT · Details ↗Needs attentionPayload coverage
32,629 cells checked · 14 coverage warnings · 0 blockers. Historical forecasts intentionally withheld can trigger coverage warnings.
Checked Sep 23 · 11:05 PM MT · Details ↗PassingAutomation checks
11/11 required operational checks pass. This measures the pipeline; forecast quality is tracked separately.
Checked Sep 23 · 11:06 PM MT · Details ↗Measured resultForecast edge evidence
n=33 · mean -0.09 pts · 95% [-0.73, +0.55] · gate closed (interval includes 0) · ~50 decisions for a 0.5 pt half-width · CLV compares the recorded bet price with the market close. This is separate from pipeline health.
Checked Sep 23 · 10:50 PM MT · Details ↗Source timestampForecast snapshot
Forecast data published Sep 22 · 12:49 PM MT. A fresh site check does not mean a new forecast was generated.
Checked Sep 23 · 11:06 PM MT · Details ↗CAPACITY
Where the work runs ↗How much usage is left
Provider allowances are account-wide, not individual model token balances. These readings use account-status requests, with saved receipts as fallback: no model calls.
Claude
Weekly2% remaining
Read Sep 23 · 11:06 PM MT
Resets Sep 24 · 9:00 PM MT
Account-wide allowance; includes work outside Playerweek.
Codex
Weekly77% remaining
Read Sep 23 · 11:06 PM MT
Resets Sep 30 · 12:10 PM MT
Account-wide Codex allowance. Direct account-status read; no model request.
Gemini CLI
Declared hold
CLI quota only. Gemini web / Deep Think allowance is not measured here.
Expected reset Sep 24 · 7:12 PM MT
UP NEXT
Full plan & evidence ↗The next work in the queue
Recorded priority order · 135 queued. Assignment can change as workers finish. Board states are claims, not deployment proof.
PW-93Eight lanes starve the write lock: weather lost 8 hours of captures+
The watchdog caught it: 'weather capture — no work for 481 min (budget 480)'.
Owner: Claude
PW-74The register is the logjam: production scaled 4x tonight, promotion did not+
Sean, 2026-09-22: "the top half is green, but everything is going to logjam at the register level?" Measured the moment he asked, and he is right.
Owner: Claude
PW-68Turn the density-ratio classifier into the loss function (GAIL loop)+
Gemini's re-ranked #1 and its explicit WHAT TO DO FIRST, after Sean pushed it off the coward's consensus.
Owner: Claude
PW-64The nine stub tabs ARE the nine desks — V-04 is a staffing gap+
Commission & Portfolio, World Model Lab, Football Intelligence, Behavior & Qualitative, Data & Provenance, Markets & Portfolio, Experience Studio, Operations & Learning, Forecast Accuracy Directorate — an exact match to D01-D08 plus the accuracy desk, 37 chartered jobs between them. The institution was chartered and never staffed, and nine 200-character pages are the visible shape of that. Work V-04 with this in mind: each tab shows what its desk knows, or the desk is struck. See §4.
Owner: Claude
Recent code changes Commits are not deployment receipts
4863fd29C-13: claim_author carries a clock -- attribution.build stamps ingested_at (ADD COLUMN IF NOT EX
2 minutes ago66a86283A-02: the hurdle's own pre-kickoff number is now filed and scored -- avail_hurdle --forecast app
4 minutes ago1f11db0cArchive podcast audio to verified Drive storage while transcription is paused
5 minutes ago705af54aC-13: an event time is not a clock -- the survey counted 8 tables as clocked on kickoff/lock_utc
5 minutes ago9a65092cReferee H-053 (QB-change exit): WEAKENED -- registered refutation stands, mechanism narrative re
5 minutes ago265ab265Q-01: a flush may not queue for the write lock past the pass's own kill; per-stage pass log
7 minutes agodd15b731E-10: unstick the d192_cells grind -- it sat at 1/18 for ten hours because per_run_s 300 is shor
8 minutes ago70eec369X-05: the 285-game key-number run finished -- gate FAILS, and the failure is one-sided: sim marg
10 minutes agoe159ae8aQ-01: second post-fix launchd pass read -- the 60/hour reading did not hold
12 minutes agob1147521Pace scheduled Codex work against live weekly allowance and supervise runs
18 minutes agoRefreshes automatically after the scheduled collection and publication, normally every five minutes. Source readings have their own timestamps. Missing or stale evidence is never a pass.