1[2 {"id":"C1","text":"The classical one-way ANOVA does not reject equal weekday mean changes (F=0.263, p=0.90).","results_keys":["anova_f_stat","anova_p_value","n_analysed"],"status":"supported","caveat":"Assumes independent, equal-variance changes; volatility clustering makes the p-value approximate."},3 {"id":"C2","text":"The Monday contrast is +0.050 bp with a 95% Welch CI of [-0.319, 0.420] bp, which includes zero.","results_keys":["effect_estimate","effect_ci_low","effect_ci_high"],"status":"supported","caveat":"Monday changes span weekends; effects below ~0.4 bp cannot be excluded."},4 {"id":"C3","text":"Out of sample (2010-2019), weekday-specific means from 2000-2009 do not beat the zero-change forecast (RMSE 4.884 vs 4.876 bp).","results_keys":["model_metric","baseline_metric","rmse_ratio_model_over_baseline","n_train","n_test"],"status":"supported","caveat":"No formal Diebold-Mariano test; the difference is tiny."},5 {"id":"C4","text":"Overall conclusion: no detectable day-of-week effect in daily 10-year yield changes.","results_keys":["conclusion","anova_p_value"],"status":"supported","caveat":"Absence of evidence within the precision of the data, not proof of zero effect."}6]7