Performance · Baseline

FORWARD-ONLY EVALUATION

Know what the model earned.

Accuracy and calibration lead. Betting outcomes follow, with sample size always visible.

Latest test window untouchedNo metric shown from training data

Calibration error31.9%7 settled moneyline forecasts

Score MAE2.357 stored score snapshots

Paper decisions0No settled paper decisions

Line value—Requires an explicitly defined comparison line

Probability calibrationPredicted probability vs observed frequency

ECE 31.9%

Model status by sportPromotion gates remain conservative

MLmlb-sequential-bivariate-poisson-v1moneyline · first_seen · 7 settledresearch0.218Brier

MLmlb-sequential-bivariate-poisson-v1score · first_seen · 7 settledresearch2.35MAE

MLmlb-sequential-bivariate-poisson-v1spread · first_seen · 7 settledresearch0.291Brier

MLmlb-sequential-bivariate-poisson-v1total · first_seen · 7 settledresearch0.254Brier

MLmlb-sequential-bivariate-poisson-v1moneyline · last_observed_prestart · 4 settledresearch0.211Brier

MLmlb-sequential-bivariate-poisson-v1score · last_observed_prestart · 4 settledresearch2.00MAE

MLmlb-sequential-bivariate-poisson-v1spread · last_observed_prestart · 4 settledresearch0.359Brier

MLmlb-sequential-bivariate-poisson-v1total · last_observed_prestart · 4 settledresearch0.226Brier

Benchmark tableModels cannot graduate on ROI alone

ModelMarket / checkpointNMAEBrierCal. errorLine valueStatus

mlb-sequential-bivariate-poisson-v1MLBmoneylinefirst_seen7—0.21831.9%—research

mlb-sequential-bivariate-poisson-v1MLBscorefirst_seen72.35———research

mlb-sequential-bivariate-poisson-v1MLBspreadfirst_seen7—0.29134.3%—research

mlb-sequential-bivariate-poisson-v1MLBtotalfirst_seen7—0.25435.8%—research

mlb-sequential-bivariate-poisson-v1MLBmoneylinelast_observed_prestart4—0.21145.9%—research

mlb-sequential-bivariate-poisson-v1MLBscorelast_observed_prestart42.00———research

mlb-sequential-bivariate-poisson-v1MLBspreadlast_observed_prestart4—0.35957.1%—research