ST Model Accuracy

Out-of-sample backtests of the skater projection engine — projections vs. what actually happened, graded on a full-season (per-82-GP) basis against a simple baseline
Skater projections
Beat the baseline by 8% on out-of-sample accuracy · p10–p90 coverage 81% (2025-26 backtest, 4,968 preds)
Player Accuracy
per 82 GP vs Baseline · 4,968 predsskill 8%p10–p90 cov 81%
StatNModelBaseSkillCov
Scoring
Points6238.528.91+4%80%
Goals6234.534.72+4%75%
Assists6235.946.39+7%80%
Power Play
PP Points6233.253.64+11%82%
Shooting
Shots62320.8922.77+8%83%
Peripheral
Hits61524.6829.12+15%81%
Blocks61512.9415.48+16%81%
PIM62312.1112.18+1%82%
Game Model — Out-of-Sample Backtest
backtest
SeasonGamesBrierBaseWin%Margin MAETotal MAEVegas MAEBeat%
2025-261,3120.24740.249553%2.131.9
2024-251,3120.23550.246160%2.131.9
2023-241,2300.23480.248661%2.131.9
2022-231,2300.23600.249860%2.111.8
2021-221,2300.22930.248364%2.081.9
2020-218120.23390.248760%2.041.9
2019-201,0120.24300.248955%2.051.8
Overall8,1380.23720.248559%2.101.9
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.