# Contract — F3-conf-02
**Question.** How accurately could US quarterly real GDP growth (A191RL1Q225SBEA, % q/q SAAR) be forecast 1 and 2 quarters ahead in real time at 8 origins (vintage dates 2024-01-31 … 2025-12-31), and does a simple time-series model beat the naive (last value) forecast?

**Estimands.** Per horizon h∈{1,2}: out-of-sample RMSE and MAE over the 8 origins of (a) naive forecast (last value in the origin vintage) and (b) one model (chosen: AR(1) by OLS, fixed a priori), errors measured against the first-release value of the target quarter; relative improvement 1−RMSE_model/RMSE_naive; mean |current − first release| over the 8 h=1 target quarters. Meaningful only if each forecast uses solely the origin's vintage file and actuals are first releases. With n=8 per horizon, uncertainty is large; no formal significance is claimed.

**Outputs.** out/forecasts.csv (32 rows; columns origin_date,target_date,horizon,forecast,actual,model,vintage_date,train_last_date,train_last_value,n_train); out/results.json (keys task_id, model_name, model_description, n_origins, first_origin, last_origin, rmse/mae naive & model h1/h2, rel_improvement_rmse_h1/h2, mean_abs_revision_h1_targets, feasibility, feasibility_reason, data_issues, n_input_rows, n_analysed, random_seed); out/report.md (sections Question; Data and exclusions; Method; Results; Uncertainty; Limitations and open issues); out/analysis.py; out/claims.json.

**Information rules.** At origin d only vintage file d is used (values and sample end). current_vintage.csv only for the revision statistic. Targets: last quarter in vintage +1, +2. Actuals: first_release.csv. Naive = last value of vintage.

**Invalidity.** Using current vintage for estimation; missing first releases for any target; a vintage whose last quarter is inconsistent with origin; first-release vintage date before origin (would mean target was known at origin).