# Log

• 2026-10-29 P0 passed: contract written.
• 2026-10-29 P1 passed: data audited; one missing value (2025-10-01). Decision: model = AR(2) on first differences with intercept, OLS, full sample from 1948, fixed a priori (no out-of-sample selection); missing in-window level filled by the model's own forecast from earlier data for forecasting only; OLS uses complete cases.
• 2026-10-29 P2/P3 passed: baseline + AR2_diff computed; DM diagnostics added. P4 passed: reproducibility and 12 checks. P5 passed: report, claims written.