Performance
Every model, judged the same honest way across all sports. The headline metrics below are the +EV subset — only the plays the card actually flags. CLV is the long-run judge: does the model beat the closing line on the games it picks.
+EV picks (graded)
295
across all sports
Record
152-143
52%
P/L
-5.57u
flat 1u stakes
ROI
-1.9%
outcome-based
Avg CLV
0.1%
beat the close
| Sport | +EV picks | Record | Win% | Avg odds | P/L (u) | ROI | CLV | ML | Spread | Total |
|---|---|---|---|---|---|---|---|---|---|---|
 MLB | 253 | 133-120 | 53% | -110 | -1.58u | -0.6% | 0.1% | 281-218 | 7-6 | 63-66 |
 NFL | 24 | 9-15 | 38% | +100 | -5.72u | -23.8% | 0.4% | — | 8-8 | 6-10 |
 NCAAF | 18 | 10-8 | 56% | -102 | +1.72u | +9.6% | 0.2% | — | 10-10 | — |
 NBA | returns October | |||||||||
 NHL | returns October | |||||||||
+EV columns cover only the picks that cleared the threshold (flat 1-unit stakes). ML / Spread / Total are each model's raw record on every graded pick. Record, ROI and P/L are outcome-based and noisy at this sample — CLV is the honest judge, and it needs hundreds of picks to trust. Aggregate CLV is in win-probability terms (MLB totals, judged in runs, are excluded from the blended CLV number). Click a sport for the full by-model breakdown.