# Contract — F6-conf-01
**Question:** Does weekday predict the one-day change in the 10y CMT Treasury yield (DGS10), 2000–2019?
**Estimands:** (i) one-way equal-variance ANOVA F-test of equal weekday means of Δy_t (bp) over all changes; (ii) Monday contrast = mean(Δy | Mon) − mean(Δy | Tue–Fri), all changes, 95% CI (Welch two-sample chosen); (iii) out-of-sample RMSE (2010–2019) of weekday-mean forecast (means estimated on changes dated ≤2009-12-31) vs zero-change forecast.
Assumptions: changes approximately independent across days (ANOVA/Welch validity); ANOVA additionally assumes equal variances and approx normality (large n mitigates).
**Info rules:** drop empty DGS10 rows first; Δy_t = 100(y_t − y_{t−1}) between consecutive available obs, dated t, weekday of t; first obs no change. Train = changes dated ≤ 2009-12-31 (incl. the change on the first 2010 day? No: that is dated 2010 → test). Weekday means for forecasting from train only. Only data/DGS10.csv.
**Outputs:** out/results.json (keys 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); out/figure1.png; out/table1.csv (weekday,n,mean_bp,se_bp); out/report.md (6 sections); out/analysis.py; out/claims.json.
**Conclusion rule:** ANOVA p≥0.05 → no_detectable_effect; else sign from Monday-contrast CI.
**Invalidating conditions:** unparseable / missing data such that changes can't be formed; using test-period info in forecasts.