Skip to content
← Back to Rankings

Model Accuracy — WNBA

How well do ranking models predict regular-season game outcomes?

Pick Accuracy

Fraction of games where the predicted favourite won

higher is better

Brier Score

Mean squared error of probability predictions (lower is better)

lower is better

Log Loss

Cross-entropy loss of probability predictions (lower is better)

lower is better

Results by Season

Season Games Pick Accuracy Brier Score Log Loss
2026 373 65.7% 0.2216 0.6347
2025 287 63.1% 0.2259 0.6434
2024 228 65.8% 0.2202 0.6316
2023 241 63.1% 0.2289 0.6492
2022 217 64.1% 0.2209 0.6327
2021 179 64.8% 0.2207 0.6321
2020 132 68.9% 0.2221 0.6360
2019 204 61.3% 0.2239 0.6395
2018 203 57.6% 0.2381 0.6692
2017 204 67.2% 0.2182 0.6269
2016 201 57.7% 0.2364 0.6652

Calibration (All Seasons)

Each point is a 5% probability bucket. On the diagonal = perfectly calibrated. Above = model underestimates; below = overestimates.