Nineteen of thirty-six, and where the errors route
Nineteen of thirty-six answers carry at least one provenance error. The thirty-six entries behind them split twenty-two routed to a cited document against fourteen not established: the marked row is the floor the essay refuses to call a measurement.
Reference & Evidence
Source: The pair's replication package: coding.json, 36 counted answers across three questions, recounted 2026-07-28 and recomputed against the shipped package 2026-08-19. Six of the fourteen not-established entries are Gemini rows whose citation URLs the collector does not retain, so the split is a floor, not a measurement.
Each line is a claim from the register this journal publishes against, resolved from the register at build time.
- A Retrieval audit 2026-07: of 36 counted answers across three questions, 19 carry at least one provenance error; the 36 error entries split 22 routed to a cited document against 14 not established, six of the fourteen on Gemini rows dataset-retrieval-audit-2026-07: registered 2026-08-19 (D-B165 pattern), recomputed against the shipped package that day ·
RAUD26-C1
Grades: A, verified against the printed page of the primary source · B, primary source, text layer only · C, authoritative secondary · D, reported.
Related exhibits
-
One hundred twenty-one citations, zero primary sources
From the essay A vendor page supplied one tested source path for the 5x retention rule
-
The chain, and what is actually found at each link
From the essay A vendor page supplied one tested source path for the 5x retention rule
-
The measurement-error indicator audit
From the essay What is measurement error? The gap between a construct and its indicator