1[2{"id":"C1","text":"Real-time naive forecast RMSE vs first releases was 2.01 pp at h=1 and 2.27 pp at h=2 over 8 origins.","results_keys":["rmse_naive_h1","rmse_naive_h2","mae_naive_h1","mae_naive_h2"],"status":"supported","caveat":"Only 8 origins (2024-2025)."},3{"id":"C2","text":"The AR(1) (1985+, COVID excluded) had lower RMSE than naive: 1.48 pp at h=1 and 1.40 pp at h=2, a relative improvement of 26% and 38%.","results_keys":["rmse_model_h1","rmse_model_h2","mae_model_h1","mae_model_h2","rel_improvement_rmse_h1","rel_improvement_rmse_h2"],"status":"supported","caveat":"Point estimates over 8 origins; window-specific."},4{"id":"C3","text":"The improvement is not statistically significant at 5% (DM p=0.155 at h=1, 0.105 at h=2), although the AR(1) had lower squared error at 5/8 (h=1) and 8/8 (h=2) origins.","results_keys":["supp_dm_stat_h1","supp_dm_pvalue_t7_h1","supp_dm_stat_h2","supp_dm_pvalue_t7_h2","supp_n_origins_model_lower_sqerr_h1","supp_n_origins_model_lower_sqerr_h2"],"status":"supported","caveat":"Small-sample DM test is approximate."},5{"id":"C4","text":"The mean absolute revision between first release and current value of the 8 h=1 target quarters is 0.56 pp.","results_keys":["mean_abs_revision_h1_targets"],"status":"supported","caveat":"Current vintage of 2026-10-30; later revisions possible."}6]7