🐻⬇️🏀

2026 WNBA Model Performance Analysis

Scope
All Games Ranked Games

All scored games in the selected league and season. AP Poll is excluded here.

Season
2026 2025

Comparing prediction accuracy across 45 games using multiple rating models.

Model Catalog

7-day holdout coverage: 16/17 models .

Rolling Holdout Curves

Each point is a strict weekly holdout: train on all games before that week, test on that week. This first version uses a 21-day warmup, then 7-day holdouts stepped forward weekly.

Log Loss Brier AUC Accuracy

Weekly strict holdout log loss. Lower is better. Showing 16 models across 3 windows. Click legend items to hide/show series.

Recent Window Winners

Holdout Best Log Loss Runner-up Models
Jun 5 - Jun 8 Adjusted Efficiency 0.465 Log Adjusted (0.465) 16
May 29 - Jun 4 Dynamic Bradley-Terry 0.589 Elo (0.624) 16
May 22 - May 28 Bradley-Terry Recency 0.660 Bradley-Terry (0.662) 16

Model Performance Leaderboard

Models ranked by strict holdout AUC when available (fallback: full-season AUC). Hover over column headers for explanations.

# Model 7d Split AUC Acc Brier LogLoss n AUC 7d Acc 7d Brier 7d n 7d
1 Pythagorean Pythagorean Pythagorean win expectation from raw points scored and allowed. More → STRICT
20g
 - - - - 0 0.945 85.0% 0.168 20
2 Avg Margin Baseline Avg Margin Baseline Predict from simple average scoring margin in the training window. More → STRICT
20g
 - - - - 0 0.912 85.0% 0.165 20
3 Bradley-Terry Bradley-Terry Static logistic paired-comparison model with one team strength parameter. More → STRICT
20g
 - - - - 0 0.879 75.0% 0.188 20
4 Elo Elo Streaming paired-comparison rating with recency baked into sequential updates. More → STRICT
20g
 - - - - 0 0.868 70.0% 0.204 20
5 Bradley-Terry Recency Bradley-Terry Recency Static Bradley-Terry with exponential recency weights on newer games. More → STRICT
20g
 - - - - 0 0.857 75.0% 0.201 20
6 Adjusted Efficiency Adjusted Efficiency Opponent-adjusted efficiency model with separate offensive and defensive components. More → STRICT
20g
 - - - - 0 0.857 75.0% 0.149 20
7 Log Adjusted Log Adjusted Log-scale adjusted efficiency model that downweights blowout leverage. More → STRICT
20g
 - - - - 0 0.857 75.0% 0.149 20
8 Core Ensemble Core Ensemble Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and points off/def. More → STRICT
20g
 - - - - 0 0.857 65.0% 0.174 20
9 Recency Ensemble Recency Ensemble Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and recency points off/def. More → STRICT
20g
 - - - - 0 0.857 65.0% 0.174 20
10 Dynamic Bradley-Terry Dynamic Bradley-Terry Time-evolving paired-comparison model with latent team strength drift. More → STRICT
20g
 - - - - 0 0.857 75.0% 0.180 20
11 Points Off/Def Recency Points Off/Def Recency Off/def points regression with exponential recency weights. More → STRICT
20g
 - - - - 0 0.846 70.0% 0.183 20
12 Adjusted Context Blend Adjusted Context Blend Experimental context-heavy win model blending strong team components with rest and venue context. More → STRICT
20g
 - - - - 0 0.846 60.0% 0.193 20
13 Margin Recency Margin Recency Margin regression with exponential recency weights on newer games. More → STRICT
20g
 - - - - 0 0.824 70.0% 0.185 20
14 Margin Margin Linear team-strength model fit on point differential instead of binary wins. More → STRICT
20g
 - - - - 0 0.813 70.0% 0.186 20
15 Points Off/Def Points Off/Def Raw points regression with separate offensive and defensive team parameters. More → STRICT
20g
 - - - - 0 0.813 70.0% 0.185 20
16 Home Team Baseline Home Team Baseline Always favor the home team with a fixed prior. More → STRICT
20g
 - - - - 0 0.736 70.0% 0.220 20
- Efficiency Efficiency Tempo-adjusted efficiency version of Pythagorean ratings. More → FULL
no 7d
 - - - - 0 - - - 0

Methodology

ELO / Bradley-Terry

  • ELO: Iterative updates, K=64, HCA=100
  • BT: Static logistic regression on all games
  • Both model win probability, not margin
  • ELO updates after each game; BT fits once

Pythagorean Models

  • Raw: Classic points scored/allowed formula
  • Efficiency: Pace-adjusted (pts per possession)
  • Adjusted: Opponent-adjusted efficiency
  • Log: Log-linear multiplicative scale

Margin Regression

  • Team-level ridge regression on point margin
  • Linear Bradley-Terry (margin target)
  • Alpha=0.05 (CV-tuned)
  • Learns home advantage from data (~6 pts)

Baselines

  • Home Team: Always predict home wins (60%)
  • Avg Margin: Higher average margin wins
  • Models should beat these to add value