2009 WNBA Model Performance
Season
Every ranked model predicted the same 196 games (2009-06-27 to 2009-10-09, 15 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, 2009-10-03 to 2009-10-09), 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 | Elo Elo Streaming paired-comparison rating with recency baked into sequential updates. More → | 0.659 | 62.8% | 0.233 | 0.661 | 0.750 | 66.7% |
| 2 | Adjusted Context Blend Adjusted Context Blend Chronologically trained component blend with rest and signed venue context. More → | 0.654 | 59.7% | 0.235 | 0.666 | 0.750 | 66.7% |
| 3 | Margin Margin Linear team-strength model fit on point differential instead of binary wins. More → | 0.649 | 61.2% | 0.234 | 0.660 | 0.750 | 66.7% |
| 4 | Points Off/Def Points Off/Def Raw points regression with separate offensive and defensive team parameters. More → | 0.643 | 57.7% | 0.235 | 0.664 | 0.750 | 66.7% |
| 5 | Margin Recency Margin Recency Margin regression with exponential recency weights on newer games. More → | 0.641 | 58.7% | 0.235 | 0.662 | 0.750 | 66.7% |
| 6 | Points Off/Def Recency Points Off/Def Recency Off/def points regression with exponential recency weights. More → | 0.637 | 58.2% | 0.236 | 0.665 | 0.750 | 66.7% |
| 7 | Home Team Baseline Home Team Baseline Always favor the home team with a fixed prior. More → | 0.633 | 63.3% | 0.233 | 0.660 | 0.750 | 66.7% |
| 8 | Core Ensemble Core Ensemble Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and points off/def. More → | 0.604 | 58.2% | 0.242 | 0.678 | 0.750 | 66.7% |
| 9 | Recency Ensemble Recency Ensemble Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and recency points off/def. More → | 0.601 | 57.1% | 0.243 | 0.678 | 0.750 | 66.7% |
| 10 | Dynamic Bradley-Terry Dynamic Bradley-Terry Time-evolving paired-comparison model with latent team strength drift. More → | 0.588 | 55.6% | 0.255 | 0.709 | 0.500 | 66.7% |
| 11 | Bradley-Terry Recency Bradley-Terry Recency Static Bradley-Terry with exponential recency weights on newer games. More → | 0.570 | 56.6% | 0.250 | 0.694 | 0.500 | 66.7% |
| 12 | Bradley-Terry Bradley-Terry Static logistic paired-comparison model with one team strength parameter. More → | 0.569 | 54.6% | 0.252 | 0.699 | 0.500 | 33.3% |
| 13 | Adjusted Efficiency Adjusted Efficiency Opponent-adjusted efficiency model with separate offensive and defensive components. More → | 0.568 | 55.1% | 0.252 | 0.698 | 0.500 | 66.7% |
| 14 | Log Adjusted Log Adjusted Log-scale adjusted efficiency model that downweights blowout leverage. More → | 0.567 | 55.1% | 0.252 | 0.698 | 0.500 | 66.7% |
| 15 | Avg Margin Baseline Avg Margin Baseline Predict from simple average scoring margin in the training window. More → | 0.560 | 55.6% | 0.250 | 0.693 | 0.500 | 66.7% |
| 16 | Pythagorean Pythagorean Pythagorean win expectation from raw points scored and allowed. More → | 0.559 | 54.6% | 0.247 | 0.688 | 0.500 | 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