- 2026-10-30 P0 passed: contract written. P1 started.
- 2026-10-30 Decision (made before computing any forecast errors): model = AR(1) with intercept, OLS, estimation sample 1985Q1..last quarter of the origin vintage, with 2020Q2 and 2020Q3 (pandemic outliers, -28/+35% SAAR) set to missing so that no regression pair uses them; iterated forecasts for h=2. Rationale: post-Great-Moderation sample; pandemic swings would otherwise dominate OLS. No tuning on evaluation results.
- 2026-10-30 P1 passed (audit written; hashes match). P2 baseline (naive) and P3 AR(1) computed in out/analysis.py. AR(1) cross-checked with statsmodels OLS (last origin c=1.7716, phi=0.3347, 159 pairs).
- 2026-10-30 Exclusion ledger changed to single chain for last origin (gate requires chained rows); per-origin ledger in exclusions_by_origin.csv.
- 2026-10-30 P4 passed (gate receipt 20261030T042026). P5: report and claims written; final gate run.
- 2026-10-30 P5 passed; final gate PASSED.