NBA • Tracking

Model Performance

This page tracks Dr Cover's NBA Model – Version 1. The model was built using machine learning and decades of historical NBA data to evaluate performance over time. This is the baseline release, and future versions will include additional enhancements, expanded features, and continued calibration as the OddsRX platform evolves.

📄 Read: NBA Model Notes

Data: Live
Last updated: 3/1/2026, 11:18 PM
Rows: 44

Sides (Winners) 4,879 games

Split Games Correct Accuracy
Overall 838 503 60.0%
Pred not B2B 735 442 60.1%
Mkt agree 603 408 67.7%
Book fav (>55%) 564 387 68.6%
Conf 50–55% 554 303 54.7%
Home picks 517 318 61.5%
Away picks 321 185 57.6%
Conf 56–65% 277 194 70.0%
Mkt disagree 214 83 38.8%
Dog (<45%) 146 48 32.9%
Pred on B2B 103 61 59.2%
Conf 66–75% 6 5 83.3%
Conf >75% 1 1 100.0%

Spreads (ATS) 5,712 games

Split Games Correct Accuracy
Overall 816 421 51.6%
Picking dog 710 367 51.7%
Pick not B2B 689 348 50.5%
WinConf 50–55% 537 289 53.8%
Disagrees w/ winner 492 259 52.6%
Away picks 438 228 52.1%
Home picks 378 193 51.1%
Agrees w/ winner 324 162 50.0%
Spread >7.5 285 142 49.8%
Spread 4–7.5 278 144 51.8%
WinConf 56–65% 273 127 46.5%
Spread ≤3.5 253 135 53.4%
Pick on B2B 127 73 57.5%
Picking favorite 106 54 50.9%
WinConf 66–75% 5 4 80.0%
WinConf >75% 1 1 100.0%

Expected Value 1,512 games

Split Games Correct Accuracy
Negative EV 580 392 67.6%
Positive EV 241 100 41.5%
EV >5% 203 82 40.4%
EV >10% 181 67 37.0%
EV >15% 163 60 36.8%
EV >20% 144 53 36.8%

Totals (O/U) 1,983 games

Split Games Correct Accuracy
Overall 805 398 49.4%
Pred overs 407 190 46.7%
Pred unders 378 189 50.0%
Overs Δ≥5 218 103 47.2%
Unders Δ≤-5 175 84 48.0%

Variance / Pace 1,610 games

Split Games Correct Accuracy
Low SD ≤25.6 403 192 47.6%
Slow ≤98.3 403 206 51.1%
High SD >25.6 402 206 51.2%
Fast >98.3 402 192 47.8%