๐Ÿปโฌ‡๏ธ๐Ÿ€

1999 WNBA Model Performance

Every ranked model predicted the same 138 games (1999-07-05 to 1999-09-05, 9 weekly windows) out of sample: each week's games were predicted by models fit only on the season's earlier games.

Model Catalog · Player models

Model Performance Leaderboard

Ranked by full-season AUC. The final week is the season's last 7 played days (5 games, 1999-08-30 to 1999-09-05), 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 Points Off/Def Points Off/Def Raw points regression with separate offensive and defensive team parameters. More โ†’ 0.717 63.0% 0.218 0.626 0.583 80.0%
2 Points Off/Def Recency Points Off/Def Recency Off/def points regression with exponential recency weights. More โ†’ 0.715 63.8% 0.219 0.628 0.583 80.0%
3 Margin Margin Linear team-strength model fit on point differential instead of binary wins. More โ†’ 0.702 63.0% 0.222 0.633 0.583 80.0%
4 Margin Recency Margin Recency Margin regression with exponential recency weights on newer games. More โ†’ 0.699 64.5% 0.223 0.636 0.583 80.0%
5 Elo Elo Streaming paired-comparison rating with recency baked into sequential updates. More โ†’ 0.687 61.6% 0.225 0.641 0.583 80.0%
6 Adjusted Context Blend Adjusted Context Blend Chronologically trained component blend with rest and signed venue context. More โ†’ 0.679 65.2% 0.231 0.656 0.583 80.0%
7 Core Ensemble Core Ensemble Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and points off/def. More โ†’ 0.670 60.9% 0.231 0.653 0.583 80.0%
7 Recency Ensemble Recency Ensemble Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and recency points off/def. More โ†’ 0.670 60.9% 0.231 0.653 0.583 80.0%
9 Dynamic Bradley-Terry Dynamic Bradley-Terry Time-evolving paired-comparison model with latent team strength drift. More โ†’ 0.646 63.8% 0.239 0.673 0.500 80.0%
10 Home Team Baseline Home Team Baseline Always favor the home team with a fixed prior. More โ†’ 0.644 64.5% 0.231 0.655 0.583 60.0%
11 Log Adjusted Log Adjusted Log-scale adjusted efficiency model that downweights blowout leverage. More โ†’ 0.641 57.2% 0.246 0.690 0.667 80.0%
12 Adjusted Efficiency Adjusted Efficiency Opponent-adjusted efficiency model with separate offensive and defensive components. More โ†’ 0.641 57.2% 0.246 0.690 0.667 80.0%
13 Bradley-Terry Bradley-Terry Static logistic paired-comparison model with one team strength parameter. More โ†’ 0.636 60.9% 0.243 0.683 0.667 80.0%
14 Avg Margin Baseline Avg Margin Baseline Predict from simple average scoring margin in the training window. More โ†’ 0.627 55.8% 0.239 0.669 0.667 80.0%
15 Bradley-Terry Recency Bradley-Terry Recency Static Bradley-Terry with exponential recency weights on newer games. More โ†’ 0.626 60.9% 0.243 0.680 0.667 80.0%
16 Pythagorean Pythagorean Pythagorean win expectation from raw points scored and allowed. More โ†’ 0.611 56.5% 0.241 0.675 0.667 80.0%

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