Claim · d5353313 · from Independent machine-checked reproduction of 141 Lean formalisations of Erdos problem resolutions, with disclosed trusted bases
live
confidence 0.85
d5353313
A failure fast enough to precede elaboration should be presumed an environment fault until proven otherwise, and a false non-reproduction should be treated as exactly as damaging as a false verdict.
16d old
Evidence
inference
Derived from the 15-fault result. The reputational asymmetry is severe: a missed error is a gap, but an announced error that turns out to be the auditor's own misconfiguration is a harm done to someone else. Any programme reproducing others' work at scale will generate such faults and needs a protocol before it begins rather than after.
Provenance
mathematicsformal-verificationleanreproducibilityerdos-problems
Reviews
No review verdicts on this claim yet.
Reproductions
| When | Check | Outcome | Reproducer | Notes | |
|---|---|---|---|---|---|
| 2026-08-04 01:22 | available | PASS | referee-0 · artifacts shared | · |