The Lab Analytics|Basketball

 2025-26 SEASON
Today
APR 05
Total Games
2
Model
XGBoost
Features
144
Training Set
25,500+
Season
2025-26

All Games Today

2 games shown
▼ Medium (60-70%) 1 games
Med 67%
No edge48
08:00 PM — CONFLICT
Tulsa
33.5% 29-7
+185
@
Auburn PICK
66.5% 21-16
-225
Predicted Winner Auburn (66.5%)
Spread
Auburn -6.5
▼ Low (<60%) 1 games
Low 51%
+EDGE +11%54
05:30 PM — CLOSE UNDERDOG CONFLICT
Oklahoma
48.9% 21-15
-170
@
West Virginia PICK
51.1% 20-14
+142
Predicted Winner West Virginia (51.1%)
Spread
Virginia +3.5
TIERED RECOMMENDATION SYSTEM
PREMIUM: 80%+ conf, 5+ edge (88.5% win, 69.1% ROI) • GOOD: 75%+ conf, 7+ edge (82.7% win, 57.9% ROI)
Based on 4,181 game analysis

No games in this category.

No games in this category.

▼ Low (<60%) 1 games
Low 51%
+EDGE +11%54
05:30 PM — CLOSE CONFLICT
Oklahoma
48.9% 21-15
-170
@
West Virginia PICK
51.1% 20-14
+142
Predicted Winner West Virginia (51.1%)
Spread
Virginia +3.5
▼ Low (<60%) 1 games
Low 51%
+EDGE +11%54
05:30 PM — UNDERDOG ML +11.5% EDGE CONFLICT
Oklahoma
48.9% 21-15
-170
@
West Virginia PICK
51.1% 20-14
+142
Predicted Winner West Virginia (51.1%)
Spread
Virginia +3.5
▼ Low (<60%) 1 games
Low 51%
+EDGE +11%54
05:30 PM — ML +142 EV +23.6% CONFLICT
Oklahoma
48.9% 21-15
-170
@
West Virginia PICK
51.1% 20-14
+142
Predicted Winner West Virginia (51.1%)
Spread
Virginia +3.5

No games in this category.

No games in this category.

No games in this category.

Model Explainability (SHAP Analysis)

SHAP (SHapley Additive exPlanations) values show which features have the most impact on our model's predictions. The values below represent the mean absolute SHAP value across all 2 games today.

XGBoost Win/Loss Model

Feature importance for win/loss predictions. Higher values indicate stronger influence on the predicted outcome.

Away Team Implied Probability
0.2377
Point Spread
0.2247
Dominance Diff
0.2122
Quality Dominance Diff
0.0694
Srs Diff Rolling
0.0570
Home Cumsum Fouls
0.0471
Efficiency Diff
0.0439
Home Moneyline
0.0317
Away Cumsum Reb Offensive
0.0308
Conference Game True
0.0276

How to interpret: Features at the top have the largest average impact on predictions. The bar width shows relative importance, and the number shows the mean absolute SHAP value.

Technical note: SHAP values are calculated for each game individually, then averaged to show overall feature importance across today's predictions.

NCAA Basketball Prediction Model

Win/Loss: xgb_basketball.pkl | Points: XGB-POINTS (dual-perspective) (xgb_points.joblib)

Generated: Sunday, April 05, 2026 • 2026-04-05T18:00:38.451313

The Lab Analytics · 2026 · Privacy Policy · Disclaimer