Skip to main content

How the model scores

Sharp products prove edge with numbers, not testimonials. Below is the prop model graded against the no-vig market on settled picks - log loss (lower is better) and calibration (predicted vs actual hit rate). Updated daily by the model's own back-test.

Estimator back-test

hitrate_prob0.81820.282262.9% hit · n=140
dist_prob0.92410.293761.5% hit · n=169
dist_prob_unadj0.92410.293761.5% hit · n=169
model_only_prob0.96130.299761.7% hit · n=149
shrunk_prob0.85170.288061.7% hit · n=167
fair_prob0.67270.239860.8% hit · n=74

Closing-line coverage

93% of settled picks have a captured closing line (26/28). CLV is only honest when coverage is high - we publish this so you can check.

By sport (worst-covered first)

  • mlb93%(26/28)

Parlays are excluded (no single closing line). Some sports (player props on smaller leagues) have thinner free-market closing data - we show it rather than hide the gap.

Methodology: every prop pick records its model probability and the no-vig market consensus at fire time. After the game settles, the back-test scores each with log loss and Brier. We do not hide losing stretches - see the public /record.

Must be 21+. Gambling problem? Call 1-800-GAMBLER. NJ/PA: 1-800-GAMBLER · MI: 1-800-270-7117 · VA: 1-888-532-3500 · CO: 1-800-522-4700 · NY: 877-846-7369 (HOPENY).

NuroPicks LLC · 30 N Gould St Ste R, Sheridan, WY 82801