2025-2026 NCAAMD2 Model Performance
Season
2026-2027 has not had enough games yet: results start once three weeks of games are in (the first holdout week is predicted from 21 days of earlier games). Showing 2025-2026, the newest season with results.
Every ranked model predicted the same 3578 games (2025-11-24 to 2026-04-05, 19 weekly windows) out of sample: each week's games were predicted by models fit only on the season's earlier games. 164 of 3742 games are left out because a model had no rating for a team yet.
Model Performance Leaderboard
Ranked by full-season AUC. The final week is the season's last 7 played days (1 games, 2026-03-30 to 2026-04-05), 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 Recency Points Off/Def Recency Off/def points regression with exponential recency weights. More → | 0.805 | 72.1% | 0.185 | 0.550 | - | 100.0% |
| 2 | Margin Recency Margin Recency Margin regression with exponential recency weights on newer games. More → | 0.804 | 72.1% | 0.184 | 0.547 | - | 100.0% |
| 3 | Points Off/Def Points Off/Def Raw points regression with separate offensive and defensive team parameters. More → | 0.804 | 72.1% | 0.185 | 0.550 | - | 100.0% |
| 4 | Margin Margin Linear team-strength model fit on point differential instead of binary wins. More → | 0.804 | 71.9% | 0.184 | 0.548 | - | 100.0% |
| 5 | Dynamic Bradley-Terry Dynamic Bradley-Terry Time-evolving paired-comparison model with latent team strength drift. More → | 0.784 | 70.5% | 0.190 | 0.562 | - | 100.0% |
| 6 | Bradley-Terry Bradley-Terry Static logistic paired-comparison model with one team strength parameter. More → | 0.781 | 70.4% | 0.191 | 0.561 | - | 100.0% |
| 7 | Bradley-Terry Recency Bradley-Terry Recency Static Bradley-Terry with exponential recency weights on newer games. More → | 0.780 | 70.4% | 0.191 | 0.563 | - | 100.0% |
| 8 | Elo Elo Streaming paired-comparison rating with recency baked into sequential updates. More → | 0.778 | 70.7% | 0.194 | 0.569 | - | 100.0% |
| 9 | Pythagorean Pythagorean Pythagorean win expectation from raw points scored and allowed. More → | 0.775 | 69.6% | 0.194 | 0.567 | - | 100.0% |
| 10 | Avg Margin Baseline Avg Margin Baseline Predict from simple average scoring margin in the training window. More → | 0.763 | 69.1% | 0.198 | 0.579 | - | 100.0% |
| 11 | Home Team Baseline Home Team Baseline Always favor the home team with a fixed prior. More → | 0.598 | 59.8% | 0.240 | 0.674 | - | 100.0% |
| - |
Adjusted Context Blend
Adjusted Context Blend
Chronologically trained component blend with rest and signed venue context.
More →
not compared: predicted 0 of 3742 games
|
- | - | - | - | - | - |
| - |
Adjusted Efficiency
Adjusted Efficiency
Opponent-adjusted efficiency model with separate offensive and defensive components.
More →
not compared: predicted 0 of 3742 games
|
- | - | - | - | - | - |
| - |
Core Ensemble
Core Ensemble
Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and points off/def.
More →
not compared: predicted 0 of 3742 games
|
- | - | - | - | - | - |
| - |
Log Adjusted
Log Adjusted
Log-scale adjusted efficiency model that downweights blowout leverage.
More →
not compared: predicted 0 of 3742 games
|
- | - | - | - | - | - |
| - |
Recency Ensemble
Recency Ensemble
Equal-logit blend of Elo, recency BT, recency margin, log-adjusted pyth, and recency points off/def.
More →
not compared: predicted 0 of 3742 games
|
- | - | - | - | - | - |
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