1# Verification checks (P4)23- Clean re-run (outputs deleted, analysis.py re-run): results.json and forecasts.csv byte-identical: PASS4- Leakage test: all values after 2024-12-01 set to 9.9 in a temp copy; forecasts for origins <= 2024-12-01 identical, later ones change: PASS5- 144 ledger rows: PASS6- unique (origin,h,model): PASS7- target=origin+h months: PASS8- actuals match data: PASS9- n_eval = non-missing actuals: PASS10- naive = last obs <= origin: PASS11- rmse/mae naive h1 recomputed from ledger: PASS12- rmse/mae model h1 recomputed from ledger: PASS13- rel_improvement h1 identity: PASS14- rmse/mae naive h3 recomputed from ledger: PASS15- rmse/mae model h3 recomputed from ledger: PASS16- rel_improvement h3 identity: PASS17- RMSE>=MAE>0: PASS18- exclusion ledger: 945-1-3=941=n_analysed: PASS19- first_origin<last_origin; last target 2026-09 within data: PASS20- naive RMSE h3 > h1 (plausible magnitude): PASS21- Descriptive: UNRATE min 3.5 / max 4.5 over 2023-07..2026-09 (non-missing): computed, cited in report Uncertainty.22