1[2 {"id": "C1", "text": "Over 60 origins (2021-09 to 2026-08), the ridge forecast (lambda=50, expanding standardisation) has RMSE 0.561 and MAE 0.451 pp/month, versus 0.883/0.699 for naive and 0.544/0.429 for the expanding mean.",3 "results_keys": ["rmse_ridge","mae_ridge","rmse_naive","mae_naive","rmse_mean","mae_mean"], "status": "supported",4 "caveat": "Current-vintage data; publication lags ignored."},5 {"id": "C2", "text": "Ridge improves on the naive forecast by about 36.5% in RMSE, and the difference is statistically significant (DM p~0.001).",6 "results_keys": ["rel_improvement_vs_naive"], "status": "supported",7 "caveat": "DM test and bootstrap are supplementary (out/uncertainty.json); the naive baseline is weak for a noisy growth series."},8 {"id": "C3", "text": "Ridge does not beat the expanding historical mean: its RMSE is about 3.1% higher (relative improvement -0.031), and the difference is not statistically significant (DM p~0.56-0.59; bootstrap 90% CI [-0.14, 0.04]).",9 "results_keys": ["rel_improvement_vs_mean","rmse_ridge","rmse_mean"], "status": "supported",10 "caveat": "Short 60-month evaluation window in one regime; lambda fixed and not tuned."},11 {"id": "C4", "text": "All standardisation statistics and the target mean were recomputed from each origin's training window only.",12 "results_keys": ["normalisation_window","n_analysed"], "status": "supported",13 "caveat": "None."}14]15