🐻⬇️🏀

2025-2026 NCAAW Model Performance Analysis

Scope
All Games Ranked Games

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

Season
2025-2026

Comparing prediction accuracy across 2698 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 22 windows. Click legend items to hide/show series.

Recent Window Winners

Holdout Best Log Loss Runner-up Models
Apr 1 - Apr 5 Elo 0.673 Pythagorean (0.687) 16
Mar 25 - Mar 31 Adjusted Context Blend 0.474 Points Off/Def (0.500) 16
Mar 18 - Mar 24 Adjusted Efficiency 0.431 Log Adjusted (0.431) 16
Mar 11 - Mar 17 Margin Recency 0.603 Margin (0.611) 16
Mar 4 - Mar 10 Recency Ensemble 0.500 Core Ensemble (0.501) 16
Feb 25 - Mar 3 Recency Ensemble 0.484 Core Ensemble (0.484) 16
Feb 18 - Feb 24 Recency Ensemble 0.515 Core Ensemble (0.516) 16
Feb 11 - Feb 17 Core Ensemble 0.485 Recency Ensemble (0.485) 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 Efficiency Efficiency Tempo-adjusted efficiency version of Pythagorean ratings. More → FULL
no 7d
 0.840 75.4% 0.170 0.528 5048 - - - 0
2 Bradley-Terry Bradley-Terry Static logistic paired-comparison model with one team strength parameter. More → STRICT
14g
 0.876 78.8% 0.149 0.462 5048 0.800 57.1% 0.212 14
3 Margin Recency Margin Recency Margin regression with exponential recency weights on newer games. More → STRICT
14g
 - - - - 0 0.800 71.4% 0.180 14
4 Points Off/Def Recency Points Off/Def Recency Off/def points regression with exponential recency weights. More → STRICT
14g
 - - - - 0 0.756 78.6% 0.180 14
5 Core Ensemble Core Ensemble Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and points off/def. More → STRICT
14g
 - - - - 0 0.756 71.4% 0.194 14
6 Recency Ensemble Recency Ensemble Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and recency points off/def. More → STRICT
14g
 - - - - 0 0.756 71.4% 0.191 14
7 Dynamic Bradley-Terry Dynamic Bradley-Terry Time-evolving paired-comparison model with latent team strength drift. More → STRICT
14g
 - - - - 0 0.733 78.6% 0.204 14
8 Margin Margin Linear team-strength model fit on point differential instead of binary wins. More → STRICT
14g
 0.878 79.4% 0.148 0.455 5048 0.711 71.4% 0.202 14
9 Adjusted Efficiency Adjusted Efficiency Opponent-adjusted efficiency model with separate offensive and defensive components. More → STRICT
14g
 0.871 78.0% 0.151 0.466 5048 0.711 71.4% 0.208 14
10 Log Adjusted Log Adjusted Log-scale adjusted efficiency model that downweights blowout leverage. More → STRICT
14g
 0.870 78.0% 0.151 0.468 5048 0.711 71.4% 0.204 14
11 Points Off/Def Points Off/Def Raw points regression with separate offensive and defensive team parameters. More → STRICT
14g
 0.870 78.1% 0.153 0.469 5048 0.711 78.6% 0.197 14
12 Elo Elo Streaming paired-comparison rating with recency baked into sequential updates. More → STRICT
14g
 0.841 75.4% 0.168 0.511 5048 0.689 64.3% 0.215 14
13 Bradley-Terry Recency Bradley-Terry Recency Static Bradley-Terry with exponential recency weights on newer games. More → STRICT
14g
 - - - - 0 0.689 57.1% 0.224 14
14 Adjusted Context Blend Adjusted Context Blend Experimental context-heavy win model blending strong team components with rest and venue context. More → STRICT
14g
 - - - - 0 0.667 64.3% 0.216 14
15 Avg Margin Baseline Avg Margin Baseline Predict from simple average scoring margin in the training window. More → STRICT
14g
 0.843 75.9% 0.166 0.502 5048 0.644 50.0% 0.226 14
16 Home Team Baseline Home Team Baseline Always favor the home team with a fixed prior. More → STRICT
14g
 0.595 59.6% 0.241 0.675 5048 0.633 64.3% 0.231 14
17 Pythagorean Pythagorean Pythagorean win expectation from raw points scored and allowed. More → STRICT
14g
 0.794 73.6% 0.189 0.563 5048 0.600 50.0% 0.226 14

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