Model validation /backtest
stats after R28 · #11
As-of-round backtestno look-ahead · vs 1/3 uniform baseline
BEATS BASELINE
2025/26: PASS — beats uniform on all three scores over 280 matches (Brier 0.525 vs 0.667; RPS 0.189 vs 0.278; log-loss 0.898 vs 1.099).
ScoreModelUniformSkill
Brier0.5250.66721.2%
RPS0.1890.27832.1%
Log-loss0.8981.09918.3%
scope: 2025/26 · 280 matches · ratings frozen at the eve of each match · lower is better for all three scores.
also runnable: dotnet run --project src/Ohms.PremierLeague.Ingest -- backtest-pl 2024
SYSTEM HALTED — Runtime fault detected. RELOAD 🗙

Rejoining the server...

Rejoin failed... trying again in seconds.

Failed to rejoin.
Please retry or reload the page.

The session has been paused by the server.

Failed to resume the session.
Please retry or reload the page.