Model CLV — Aggregate Audit

Per-sport closing-line-value audit on the platform's recommended picks. CLV = closing price − the price we took, raw implied on both sides. The model is not in that formula, which is what makes it a test rather than a restatement of our own opinion. Positive mean CLV means our picks consistently beat the close, the standard proxy for sharp skill.

-3.07pp
Our picks are not beating the closing line.
Mean closing-line value across 7,524 recommended picks. We got a better price than the close on 28.5% of them.

When we like a side, the price tends to move against us before the game starts. That is the earliest sign a pick was wrong, and it shows up long before results do.

Minimum before we sell picks +0.00pp Target +1.00pp Short of target by 4.07pp
Window:
30 days 90 days All time
View:
Per sport By sport + bet type
Methodology: CLV computed at game settlement against closing line (Pinnacle preferred, falls through to best-priced sharp book). Buckets with N<30 hidden until threshold met. Backfilled via OddsAPI historical-markets API; capture rate 84.3% (was 40.5% before 2026-05-10 fix).
Model CLV by sport and bet type.
Sport Bet type N Mean CLV Median CLV % beating close
AMERICANFOOTBALL_NFL (all) 77 -0.12pp +0.00pp 22.1%
BUNDESLIGA (all) 65 -0.80pp -0.08pp 43.1%
CS2 (all) 43 -0.04pp +0.00pp 2.3%
EPL (all) 108 -0.89pp -1.67pp 29.6%
LALIGA (all) 139 -1.62pp -1.19pp 34.5%
LIGUE1 (all) 80 -1.81pp -1.49pp 33.8%
MLB (all) 3,324 -5.48pp -1.22pp 25.2%
MMA_MIXED_MARTIAL_ARTS (all) 599 -4.03pp +0.00pp 35.9%
NBA (all) 60 +0.29pp +0.00pp 45.0%
NCAAF (all) 376 -0.15pp +0.00pp 1.6%
NFL_PRESEASON (all) 147 +0.69pp +0.00pp 26.5%
NHL (all) 54 -1.24pp +0.00pp 44.4%
SERIEA (all) 128 -1.78pp -2.41pp 28.1%
TENNIS_ATP (all) 1,106 -0.91pp +0.00pp 35.3%
TENNIS_WTA (all) 1,110 -0.79pp +0.00pp 33.3%
UCL (all) 60 +2.05pp -2.39pp 28.3%

Cumulative-mean CLV trajectory

cumulative mean CLV (pp) · N=7524-3.07pp

What we are doing about it

A number like the one above is worth little without the response to it, so here is ours, and the figure that will show whether it worked.

  1. Narrow to where we are least behind. CLV is not evenly bad. Moneyline runs far closer to the line than spreads and totals, and one sport carries both our largest sample and our worst number. Coverage follows the evidence rather than the calendar.
  2. Pull the model toward the market where it disagrees most. In the band where our probability departs furthest from the closing price, the market has been the better estimator. That is a calibration problem with a known direction, not a mystery.
  3. Fix the instrument before trusting it. Some of what this page measured was measurement error, found and corrected in September 2026. Numbers here predate and postdate those fixes; the trajectory below is the honest view.

The test: mean CLV on the picks we recommend, non-negative at 95% confidence. Until it clears that bar we are not selling picks — and this page will keep saying so.

See /trust for live calibration buckets across sports and /methodology for the signal pipeline behind each pick.