Preflight (A0): 2026-07-10 17:25 โ load1 3.25/32c (falling vs load15 4.18) ยท no heavy nasimubd jobs ยท 36 GiB avail, si/so=0 ยท all four services active โ ALL PASS. Sync: main up-to-date, worktree 0 behind. One capped run (0.35 min), readonly=2 loader, block bootstrap of real values (03c).
| Leg | Reading | Ruling |
| E1 โ certain legs (the ฮต-doctrine, applied correctly this time) | Substitution profiles bit-equal per env ยท self-pair |ฯโ1| โค 4ยทeps everywhere โ no exactness self-falsification | PASS |
| E2 โ DECISIVE (the sibling's trap): is the centerโtail contrast resolvable above its own block-bootstrap noise? | covid C=0.095, CI [0.060, 0.155] resolves ยท gold C=โ0.217, CI [โ0.346, โ0.096] resolves ยท midcycle straddles โ 2/3 + covid โ the measure escaped the max-over-grid variance trap | PASS |
| E3 โ the boundary role (THE KILL) | |C_covid โ C_midcycle| = 0.210 vs block re-split null 97.5th = 0.405 โ the between-regime difference drowns even as a scalar | FAIL |
| Family verdict | LGC 0-for-2: test (row 83) and measure (row 87) both blind at the boundary โ region-resolved dependence describes regime interiors but cannot certify between-regime differences above dependent-data noise at this scale | FAMILY CLOSED |
| Verdict (row 87) | EXCLUDE(between-regime contrast unresolvable above the block null โ blind at the boundary role) โ KILL #19, BONEYARD #34 |
Iteration 76 ยท 2026-07-10 ยท VERDICT (row 87) โ EXCLUDE ยท kill #19 ยท boneyard #34 ยท capped (0.35 min) ยท readonly=2 ยท zero generated values ยท zero code reuse ยท append-only ยท evidence: lgc_measure_exam.py ยท lgc_measure_exam_results.json