About NFL Quant
NFL Quant is a data-driven NFL analytics platform that combines an Elo-based quantitative model with real-time market lines to surface edges against the Pinnacle closing line. The goal is straightforward: find games where the model disagrees with the market and quantify the expected value of that disagreement.
Backtested across 4,630 regular-season games (2009–2025) using Elo-derived EV and cover probability at the opening Pinnacle line. Each row shows the result of picking the side with the highest EV when both thresholds are met.
| Min EV | Min Cover % | n | ATS % | ROI |
|---|---|---|---|---|
| 0% (all EV+) | — | 3,019 | 52.3% | −0.2% |
| 5% | 50% | 1,667 | 52.8% | +0.8% |
| 5% | 52% | 1,272 | 53.7% | +2.5% |
| 10% | 50% | 860 | 54.8% | +4.5% |
| 10% | 52% | 794 | 54.6% | +4.1% |
| 15% | 55% | 344 | 54.6% | +4.1% |
| 20% | 52% | 179 | 57.9% | +10.4% |
| 20% | 60% | 107 | 62.3% | +18.7% |
The EV threshold is the primary driver — cover probability adds little on its own. EV ≥ 10% + Cover ≥ 50% is the recommended default: 860 bets over 17 seasons gives a statistically meaningful sample. Higher EV thresholds (20%+) show stronger results but produce only ~10 bets per season. Defaults on the Quant Picks page are set to EV ≥ 10%.
The lean signals shown in the matchup detail panel are derived from a backtest across 354 completed NFL games (2023–2026) using public betting data — tickets percentage (how many bets are on each side) and money percentage (how much dollar volume is on each side). The gap between the two reveals sharp money: when a small number of large bets (money%) move away from where the crowd is betting (tickets%), that's a sign of professional action.
The core finding: sharp money on the under is the most reliable signal in the dataset. Sharp over signals have no historical edge. On the spread, sharp money on the away team is consistently profitable; sharp money on the home team has only a marginal edge and is shown as a lean rather than a play.
Totals Signals
| Condition | Pick | n | Hit % | ROI | Label shown |
|---|---|---|---|---|---|
| Public >60% on over and sharp money fades (gap ≤ −15) | Under | 38 | 63.2% | +20.6% | ▼ STRONG UNDER |
| Sharp money heavily on under (gap ≤ −25) | Under | 23 | 60.9% | +16.2% | ▼ Sharp UNDER |
| Only <30% of tickets on over (>70% crowd on under) | Under | 33 | 60.6% | +15.7% | ▼ UNDER lean |
| Sharp money on under (gap ≤ −15) | Under | 65 | 58.5% | +11.6% | ▼ Sharp UNDER |
| Sharp money on over (gap ≥ +15) | — | 32 | 46.9% | −10.5% | No edge (flagged) |
| Public >60% tickets on over (no sharp signal) | — | 202 | 50.0% | −4.5% | No signal |
Why under signals dominate: sportsbooks shade totals slightly high to account for public over-betting bias. Sharp bettors exploit this by taking the under, especially when the public is heavily on the over and the line hasn't moved down.
Spread Signals
| Condition | Pick | n | ATS % | ROI | Label shown |
|---|---|---|---|---|---|
| Sharp money heavily on away team (gap ≤ −20) | Away | 18 | 66.7% | +27.3% | BET [away] |
| Sharp money on away team (gap ≤ −15) | Away | 33 | 63.6% | +21.5% | BET [away] |
| Any disagreement ≥ 15pt → bet sharp side | Sharp side | 86 | 58.1% | +11.0% | BET [away] or Lean home |
| Sharp money on home team (gap ≥ +15) | Home | 53 | 54.7% | +4.5% | Lean home |
| Public >65% tickets on home | — | 40 | 50.0% | −4.5% | No signal |
| Both tickets+money >60% on away team | — | 105 | 43.8% home | −16.4% | No signal (public away picks fade home) |
Why sharp away signals are strongest: the public systematically over-bets home favorites. When sharp money disagrees and takes the away side, the line moves against the crowd and the away team covers at a high rate. Home sharp signals are weaker — sharps occasionally reinforce public home favorites, but the crowd is also sometimes right.
Disclaimer: NFL Quant is for informational and research purposes only. Past model performance does not guarantee future results. Sports betting involves significant financial risk. Always bet responsibly and within your means.