ST Model Accuracy

Out-of-sample backtests of the ST Model — projections vs. what actually happened, and the game model vs. real outcomes
Pitcher projections
Beat the Marcel baseline by 6% on full-season accuracy · p10–p90 coverage 82% (2025 backtest, 1,127 preds)
Game model vs results
Brier 0.2448 vs home-field base 0.2489 · picks the winner 56% (17,881 games)
Player Accuracy
per 700 BF vs Marcel · 1,127 predsskill 6%p10–p90 cov 82%
StatNModelBaseSkillCov
Rate Stats
ERA1610.780.83+6%71%
WHIP1610.110.13+11%91%
Strikeouts & Command
Strikeouts16117.2218.07+5%84%
Walks1618.38.12-2%91%
Run Prevention
Earned Runs16112.5913.47+7%78%
Hits Allowed16111.8413.62+13%80%
HR Allowed1615.15.18+2%80%
Game Model — Out-of-Sample Backtest
backtest
SeasonGamesBrierBaseWin%Margin MAETotal MAEVegas MAEBeat%
20252,4250.24450.248256%3.553.7
20242,4270.24550.249555%3.473.6
20232,4290.24560.249656%3.483.7
20222,4240.24300.248957%3.363.9
20212,4260.24550.248656%3.473.8
20208950.24490.247556%3.613.8
20192,4270.24430.249156%3.563.9
20182,4280.24490.249255%3.454.0
Overall17,8810.24480.248956%3.483.8
The model's pre-game win probability is scored by Brier against every completed game's outcome; a Brier below the home-field Basemeans it adds signal beyond "home teams win." Margin/Total MAE measure the projected score. Vegas MAE / Beat% read — for this sport: the free odds feed carries no historical closing lines yet, so those columns populate as line history accrues.