2025-2026 NCAAM Model Performance Analysis
Comparing prediction accuracy across 6300 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 | Margin Recency | 0.861 | 76.6% | 0.160 | 0.486 | 5950 | 0.750 | 53.8% | 0.225 | 13 |
| 2 | Dynamic Bradley-Terry | 0.857 | 77.1% | 0.160 | 0.490 | 5950 | 0.625 | 61.5% | 0.248 | 13 |
| 3 | Core Ensemble | 0.843 | 76.4% | 0.205 | 0.597 | 5950 | 0.750 | 69.2% | 0.215 | 13 |
| 4 | Recency Ensemble | 0.842 | 76.2% | 0.204 | 0.598 | 5950 | 0.650 | 61.5% | 0.223 | 13 |
| 5 | Points O/D Recency | 0.836 | 75.4% | 0.179 | 0.539 | 5950 | 0.675 | 61.5% | 0.227 | 13 |
| 6 | Bradley-Terry | 0.823 | 73.8% | 0.175 | 0.523 | 5361 | 0.625 | 61.5% | 0.228 | 13 |
| 7 | Margin Regression | 0.822 | 73.4% | 0.178 | 0.531 | 5361 | 0.800 | 69.2% | 0.195 | 13 |
| 8 | Pythagorean Adjusted | 0.815 | 73.1% | 0.179 | 0.529 | 5361 | 0.825 | 76.9% | 0.179 | 13 |
| 9 | Pythagorean Log | 0.814 | 73.2% | 0.179 | 0.530 | 5361 | 0.825 | 76.9% | 0.178 | 13 |
| 10 | Adjusted Context Blend | 0.813 | 72.4% | 0.210 | 0.601 | 5950 | 0.925 | 76.9% | 0.180 | 13 |
| 11 | Points O/D | 0.812 | 73.3% | 0.182 | 0.541 | 5361 | 0.825 | 69.2% | 0.196 | 13 |
| 12 | Pythagorean Efficiency | 0.796 | 71.6% | 0.191 | 0.571 | 5361 | - | - | - | 0 |
| 13 | Avg Margin (Baseline) | 0.784 | 70.7% | 0.195 | 0.573 | 5361 | 0.550 | 69.2% | 0.239 | 13 |
| 14 | Pure ELO | 0.783 | 70.5% | 0.191 | 0.564 | 5361 | 0.500 | 53.8% | 0.263 | 13 |
| 15 | Bradley-Terry Recency | 0.769 | 68.6% | 0.206 | 0.600 | 5950 | 0.400 | 38.5% | 0.266 | 13 |
| 16 | Pythagorean Raw | 0.752 | 69.8% | 0.208 | 0.603 | 5361 | 0.600 | 61.5% | 0.235 | 13 |
| 17 | Home Team (Baseline) | 0.638 | 64.0% | 0.232 | 0.657 | 5361 | 0.650 | 61.5% | 0.237 | 13 |
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