Samples thousands of latent performances per runner — informed by AngelScore, style-based pace sensitivity, and a shared pace shock per trial — then tallies win and top-3 rates. Fully reproducible: same seed, same result.
Ready.
Simulated outcomes are modeled distributions built from demo data — not forecasts with guaranteed accuracy. Seed 42 and scenario assumptions are shown above for reproducibility, exactly as the blueprint's governance section requires.