2017 WNBA Model Performance
Season
Every ranked model predicted the same 174 games (2017-06-08 to 2017-10-04, 17 weekly windows) out of sample: each week's games were predicted by models fit only on the season's earlier games.
Model Performance Leaderboard
Ranked by full-season AUC. The final week is the season's last 7 played days (3 games, 2017-09-28 to 2017-10-04), predicted from every earlier game: a small sample that shows the latest form.
| # | Model | AUC | Acc | Brier | LogLoss | AUC final wk | Acc final wk |
|---|---|---|---|---|---|---|---|
| 1 | Margin Margin Linear team-strength model fit on point differential instead of binary wins. More → | 0.734 | 67.8% | 0.208 | 0.604 | 0.750 | 66.7% |
| 2 | Points Off/Def Points Off/Def Raw points regression with separate offensive and defensive team parameters. More → | 0.731 | 66.7% | 0.208 | 0.603 | 0.750 | 66.7% |
| 3 | Margin Recency Margin Recency Margin regression with exponential recency weights on newer games. More → | 0.728 | 66.1% | 0.208 | 0.603 | 0.750 | 66.7% |
| 4 | Points Off/Def Recency Points Off/Def Recency Off/def points regression with exponential recency weights. More → | 0.724 | 66.7% | 0.209 | 0.605 | 0.750 | 66.7% |
| 5 | Core Ensemble Core Ensemble Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and points off/def. More → | 0.719 | 67.2% | 0.211 | 0.610 | 0.750 | 66.7% |
| 6 | Adjusted Context Blend Adjusted Context Blend Chronologically trained component blend with rest and signed venue context. More → | 0.718 | 67.8% | 0.212 | 0.616 | 0.750 | 66.7% |
| 7 | Recency Ensemble Recency Ensemble Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and recency points off/def. More → | 0.718 | 66.7% | 0.211 | 0.610 | 0.750 | 66.7% |
| 8 | Log Adjusted Log Adjusted Log-scale adjusted efficiency model that downweights blowout leverage. More → | 0.718 | 67.2% | 0.211 | 0.612 | 0.500 | 66.7% |
| 9 | Adjusted Efficiency Adjusted Efficiency Opponent-adjusted efficiency model with separate offensive and defensive components. More → | 0.717 | 67.2% | 0.211 | 0.612 | 0.500 | 66.7% |
| 10 | Avg Margin Baseline Avg Margin Baseline Predict from simple average scoring margin in the training window. More → | 0.707 | 66.7% | 0.215 | 0.621 | 0.500 | 66.7% |
| 11 | Pythagorean Pythagorean Pythagorean win expectation from raw points scored and allowed. More → | 0.700 | 67.8% | 0.221 | 0.632 | 0.500 | 66.7% |
| 12 | Elo Elo Streaming paired-comparison rating with recency baked into sequential updates. More → | 0.697 | 63.8% | 0.216 | 0.617 | 0.750 | 66.7% |
| 13 | Bradley-Terry Bradley-Terry Static logistic paired-comparison model with one team strength parameter. More → | 0.690 | 67.2% | 0.219 | 0.631 | 0.500 | 66.7% |
| 14 | Bradley-Terry Recency Bradley-Terry Recency Static Bradley-Terry with exponential recency weights on newer games. More → | 0.686 | 65.5% | 0.218 | 0.627 | 0.500 | 33.3% |
| 15 | Dynamic Bradley-Terry Dynamic Bradley-Terry Time-evolving paired-comparison model with latent team strength drift. More → | 0.679 | 66.1% | 0.224 | 0.644 | 1.000 | 33.3% |
| 16 | Home Team Baseline Home Team Baseline Always favor the home team with a fixed prior. More → | 0.600 | 59.8% | 0.240 | 0.674 | 0.750 | 66.7% |
Methodology
Out-of-sample windows
- The season is split into 7-day windows ending on its last played day
- Each window is predicted by models refit on all earlier games of the season
- The first window starts after 21 days of games
- Every model is scored on the same games; a model that cannot predict most of them is listed as not compared
Elo / Bradley-Terry
- Elo: the published league Elo (preseason priors and carryover where configured), frozen at the window start
- Bradley-Terry: static logistic fit on the earlier games
- Recency variants weight games with a 120-day half-life in NCAA leagues (keeping the non-conference games that link conferences) and a 30-day half-life in pro leagues
Pythagorean and margin models
- Raw, adjusted and log-adjusted Pythagorean from scores; efficiency from box-score possessions
- Margin: ridge regression on point margin with a learned home advantage
- Ensembles average the log-odds of five families
Baselines
- Home Team: the home team wins 60% (50% at neutral sites)
- Avg Margin: the higher average margin so far wins
- Models should beat these to add value