2025-2026 GLEAGUE Model Performance Analysis
Comparing prediction accuracy across 543 games using multiple rating models.
Model Performance Leaderboard
Models ranked by AUC on 2025-2026 games. Hover over column headers for explanations.
| # | Model | AUC | Acc | Brier | LogLoss | n | AUC 7d | Acc 7d | Brier 7d | n 7d |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Avg Margin (Baseline) | 0.686 | 64.3% | 0.223 | 0.637 | 756 | 0.667 | 75.0% | 0.231 | 4 |
| 2 | Pure ELO | 0.544 | 52.4% | 0.252 | 0.698 | 756 | - | - | 0.375 | 4 |
| 3 | Pythagorean Log | 0.541 | 53.0% | 0.301 | 0.861 | 756 | 0.667 | 75.0% | 0.227 | 4 |
| 4 | Pythagorean Adjusted | 0.541 | 53.0% | 0.301 | 0.861 | 756 | 0.667 | 75.0% | 0.227 | 4 |
| 5 | Home Team (Baseline) | 0.539 | 54.0% | 0.252 | 0.697 | 756 | 0.167 | 25.0% | 0.310 | 4 |
| 6 | Pythagorean Raw | 0.532 | 53.4% | 0.265 | 0.731 | 756 | 0.667 | 75.0% | 0.243 | 4 |
| 7 | Points O/D | 0.521 | 52.2% | 0.291 | 0.802 | 756 | 0.333 | 50.0% | 0.249 | 4 |
| 8 | Pythagorean Efficiency | 0.518 | 53.6% | 0.350 | 1.121 | 661 | - | - | - | 0 |
| 9 | Bradley-Terry | 0.486 | 48.4% | 0.273 | 0.745 | 756 | - | 25.0% | 0.272 | 4 |
| 10 | Margin Regression | 0.474 | 47.9% | 0.286 | 0.775 | 756 | 0.333 | 50.0% | 0.249 | 4 |
| 11 | Bradley-Terry Recency | - | - | - | - | 0 | - | - | 0.367 | 4 |
| 12 | Margin Recency | - | - | - | - | 0 | - | 25.0% | 0.296 | 4 |
| 13 | Points O/D Recency | - | - | - | - | 0 | - | 25.0% | 0.293 | 4 |
| 14 | Core Ensemble | - | - | - | - | 0 | - | 25.0% | 0.299 | 4 |
| 15 | Recency Ensemble | - | - | - | - | 0 | - | 25.0% | 0.309 | 4 |
| 16 | Adjusted Context Blend | - | - | - | - | 0 | - | 25.0% | 0.299 | 4 |
| 17 | Dynamic Bradley-Terry | - | - | - | - | 0 | - | - | 0.332 | 4 |
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: Bill James formula (pts scored/allowed)
- Efficiency: Pace-adjusted (pts per possession)
- Adjusted: KenPom-style opponent adjustment
- 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