# Contract — F6-conf-01

Question: does weekday predict the one-day change in the 10y CMT Treasury yield (DGS10), 2000–2019? Estimands: (i) equality of the five weekday conditional means of Δy_t (bp), classical one-way ANOVA F-test over all changes; (ii) Monday contrast = mean(Δy | Mon) − mean(Δy | Tue–Fri), with 95% CI (chosen: Welch); (iii) OOS RMSE over changes dated ≥ 2010-01-01 of weekday-mean forecast (means estimated on changes ≤ 2009-12-31) vs zero-change forecast. Assumptions: changes roughly independent with finite variance (ANOVA additionally assumes equal variances; large n makes normality unimportant); stationarity of weekday means across the sample for OOS meaning. Information rules: drop empty DGS10 rows first; Δy between consecutive available obs, dated and weekday-attributed at t; first obs no change; forecast parameters from estimation period only; only data/DGS10.csv. Outputs: results.json (n_missing_rows, n_train, n_test, anova_f_stat, anova_p_value, effect_estimate, effect_ci_low, effect_ci_high, baseline_metric, model_metric, conclusion, task_id, feasibility, feasibility_reason, data_issues, n_input_rows, n_analysed, random_seed), figure1.png, table1.csv, report.md (6 sections), analysis.py, claims.json. Invalidity/impossibility: missing file/columns, non-weekday dates, leakage of test-period data into forecasts. None found.