Review of 2026-09-23T1756Z-h156-rookie-board-coverage
Owner lane: Not recorded
Finding and verdict
WEAKENED
Review scope
Two of the record's three claims stand and one does not. STANDS: (a) coverage, (b) the mechanism, refined. DOES NOT SURVIVE AS WORDED: 'decision cost nil in weeks 1-2', because its only test is a preregistered threshold (24th-best flex among rostered played men) scaled to a league this is not. Sean's league is 10 teams with 70 non-QB starter slots per week (RB2 WR2 TE1 WR/RB/TE2 x10), so the marginal started man is near rank 60-70, not 24. Nothing here shows the misses WERE decisions either: the claim is weakened to 'unmeasured on the league's own structure', not reversed.
Next action
A fresh design would key the gap on (no prior played row) x (roster gate), derive the E4 threshold from is_starter slots before the outcome, and grade weeks 3+ boards with the same matcher.
Source provenance and publication scope
Owned research record: research/scientist/experiments/2026-09-26T0958Z-referee-h156-rookie-board-coverage.json
This public reading view includes authored question, finding, review, reasoning and next action fields. Raw measurements, commands, logs and local paths are withheld. The record ID clock is not proof of completion or deployment.