The receipts behind the claims.

Replication — the Crucible

• 21 reproduced · 12 failed · 20 not computable (53 verdicts). Every verdict ships a runnable probe and a measured number.
• The 12 failed replications — folklore that did not survive its smallest honest model — are the point. Browse the ledger →
• Each failed verdict that becomes a public post passes the full gate before it ships: an independent multi-perspective storm briefing, an adversarial stress-claim red-team (skeptic, prior-art hunter, method/confound auditor), and citation verification against primary sources.
• The ledger was itself audited this cycle (73 raw → 53 honest): duplicate textbook re-runs removed, and “by-construction” reproductions that baked in their own answer demoted. We audit a REPRODUCED with the same suspicion as a FAILED.

Forecasting

• 20 forecasts on record, each with a public falsifier and a resolution window.
• We log the calls before they resolve — the windows are still open, so there is no resolved Brier score to report yet. When they resolve, the scores land here whichever way they fall.

Self-challenge — we retract our own overclaims

• An adversarial audit runs on our own published posts and claims; several have been reframed or corrected under it.
• Most recent: we prior-art-checked one of our own “novel” findings within an hour of having it, found it textbook, and killed the headline — keeping it only as an honestly-scoped case study.

Product

• inspeximus (inspeximus, open-source on PyPI) — the recall + consolidation core. Its design rules are measured, not assumed.

A track record with no failures is a track record hiding its tests.