All Sports Performance

AD TEST MODE

PERFORMANCE HEADER

Mock Placement 728x90

NBA Basketball

basketball • Performance Deep Dive

Trained Never Predicted Never

All-Time Average (154 folds)

Flat ROI

+57.9%

Hit Rate

57.7%

+5.3pp vs b/e

Recent Trend (last 3 folds · Mar 18–Apr 28, 2026 ) ↓ Declining

Flat ROI

+23.7%

-34.1pp vs avg

Hit Rate

44.4%

Out-of-sample Accuracy

58.7%

Avg across 154 folds

AUC-ROC

0.562

0.5 = coin flip · 1.0 = perfect

Log Loss

0.6686

Lower is better

Brier Score

0.2383

Probability calibration

154 validation folds

Nov 21, 2017 – Apr 28, 2026

Final Holdout Test

Held back from all training and fold evaluation — the cleanest out-of-sample signal.

No holdout data available. Re-run prediction-system run --sport nba to generate holdout metrics.

Fold-by-Fold Metrics

Supporter+ required

Supporter access required

Unlock fold-level charts to identify training stability.

Unlock Supporter

Betting Simulation

Supporter+

Recommendation ROI

�

Recommendation Win Rate

�

Universe ROI

�

Universe Win Rate

�

Supporter access required

Unlock Supporter

Model: LightGBMModel

Calibration: platt

All-Time Average (154 folds)

Flat ROI

+29.9%

Hit Rate

67.9%

+15.5pp vs b/e

Recent Trend (last 1 folds · Mar 18–Apr 28, 2026 ) → Stable

Flat ROI

+29.3%

-0.6pp vs avg

Hit Rate

66.7%

Out-of-sample Accuracy

56.3%

Avg across 154 folds

AUC-ROC

0.587

0.5 = coin flip · 1.0 = perfect

Log Loss

0.6750

Lower is better

Brier Score

0.2411

Probability calibration

154 validation folds

Nov 21, 2017 – Apr 28, 2026

Final Holdout Test

Held back from all training and fold evaluation — the cleanest out-of-sample signal.

No holdout data available. Re-run prediction-system run --sport nba to generate holdout metrics.

Fold-by-Fold Metrics

Supporter+ required

Supporter access required

Unlock fold-level charts to identify training stability.

Unlock Supporter

Betting Simulation

Supporter+

Recommendation ROI

�

Recommendation Win Rate

�

Universe ROI

�

Universe Win Rate

�

Supporter access required

Unlock Supporter

Model: XGBoostModel

Calibration: platt

All-Time Average (154 folds)

Flat ROI

+23.6%

Hit Rate

64.7%

+12.3pp vs b/e

Recent Trend (last 1 folds · Mar 18–Apr 28, 2026 ) ↓ Declining

Flat ROI

+15.6%

-8pp vs avg

Hit Rate

60.0%

Out-of-sample Accuracy

55.3%

Avg across 154 folds

AUC-ROC

0.583

0.5 = coin flip · 1.0 = perfect

Log Loss

0.6849

Lower is better

Brier Score

0.2459

Probability calibration

154 validation folds

Nov 21, 2017 – Apr 28, 2026

Final Holdout Test

Held back from all training and fold evaluation — the cleanest out-of-sample signal.

No holdout data available. Re-run prediction-system run --sport nba to generate holdout metrics.

Fold-by-Fold Metrics

Supporter+ required

Supporter access required

Unlock fold-level charts to identify training stability.

Unlock Supporter

Betting Simulation

Supporter+

Recommendation ROI

�

Recommendation Win Rate

�

Universe ROI

�

Universe Win Rate

�

Supporter access required

Unlock Supporter

Model: XGBoostModel

Calibration: platt

Avg RMSE

20.195

Root mean squared error

Avg MAE

15.924

Mean absolute error

R² Score

-0.0365

Explained variance

154 validation folds

Nov 21, 2017 – Apr 28, 2026

Fold-by-Fold Metrics

Supporter+ required

Supporter access required

Unlock fold-level charts to identify training stability.

Unlock Supporter

Model: LightGBMRegressorModel

Calibration: platt