NFL · Empirical Bayes
True-Talent Leaderboards
A raw single-season rate is a noisy guess at a player’s real ability, and the noise is worst for the smallest samples — so a naïve leaderboard is topped by whoever got lucky in the fewest tries. This one shrinks every rate toward its position-group prior by how much the sample can be trusted, and shows a 90% credible interval instead of a point. Toggle raw vs. shrunk to watch the flukes fall back to the pack.
Rush success rate · 2024
Share of carries with positive EPA. Blocking & scheme are large, un-controlled inputs.
- Beats raw by
- 19.4%
- lower out-of-sample error
- RMSE raw → shrunk
- 9.1% → 7.3%
- odd vs. even weeks
- Split-half reliability
- 0.20
- how repeatable the raw stat is
- Stabilizes at
- 153
- carrys to trust the number
Each row is a player: the solid dot is the shrunk estimate, the hollow dot the raw rate, joined by the red pull of regression; the grey bar is the 90% credible interval. A hollow shrunk dot means the sample is below the stabilization line — the number is mostly the position prior. Switch to Raw rank and watch the small-sample names climb.
Rush success rate · 2024 · full board
| # | Player | Team | Pos | carrys | Raw | Shrunk | 90% interval |
|---|---|---|---|---|---|---|---|
| 1 | Bijan Robinson | ATL | RB | 305 | 49.5% | 46.1% | 42.3% – 50.0% |
| 2 | Derrick Henry | BAL | RB | 325 | 47.7% | 45.0% | 41.3% – 48.8% |
| 3 | Tyler Allgeier◦ provisional | ATL | RB | 138 | 49.3% | 44.0% | 39.3% – 48.9% |
| 4 | David Montgomery | DET | RB | 186 | 47.9% | 44.0% | 39.6% – 48.5% |
| 5 | Emanuel Wilson◦ provisional | GB | RB | 103 | 50.5% | 43.8% | 38.8% – 48.9% |
| 6 | Saquon Barkley | PHI | RB | 348 | 44.5% | 43.0% | 39.3% – 46.6% |
| 7 | Jeremy McNichols◦ provisional | WAS | RB | 56 | 51.8% | 42.7% | 37.1% – 48.3% |
| 8 | Chris Rodriguez Jr.◦ provisional | WAS | RB | 35 | 57.1% | 42.6% | 36.8% – 48.6% |
| 9 | Chuba Hubbard | CAR | RB | 251 | 44.6% | 42.6% | 38.6% – 46.7% |
| 10 | Bucky Irving | TB | RB | 207 | 44.9% | 42.5% | 38.3% – 46.9% |
| 11 | Jahmyr Gibbs | DET | RB | 250 | 44.4% | 42.5% | 38.5% – 46.5% |
| 12 | Kyren Williams | LA | RB | 316 | 44.0% | 42.5% | 38.7% – 46.2% |
| 13 | Ty Johnson◦ provisional | BUF | RB | 41 | 53.7% | 42.4% | 36.6% – 48.2% |
| 14 | Rico Dowdle | DAL | RB | 235 | 44.3% | 42.3% | 38.2% – 46.5% |
| 15 | Rachaad White◦ provisional | TB | RB | 144 | 45.1% | 42.1% | 37.5% – 46.9% |
| 16 | Kareem Hunt | KC | RB | 200 | 44.0% | 42.0% | 37.7% – 46.3% |
| 17 | Sincere McCormick◦ provisional | LV | RB | 39 | 51.3% | 41.8% | 36.0% – 47.6% |
| 18 | Roschon Johnson◦ provisional | CHI | RB | 55 | 47.3% | 41.4% | 35.9% – 47.1% |
| 19 | Jaleel McLaughlin◦ provisional | DEN | RB | 113 | 44.3% | 41.4% | 36.5% – 46.4% |
| 20 | James Conner | ARI | RB | 239 | 42.7% | 41.4% | 37.3% – 45.5% |
| 21 | Khalil Herbert◦ provisional | CIN | RB | 36 | 50.0% | 41.4% | 35.5% – 47.3% |
| 22 | Chris Brooks◦ provisional | GB | RB | 36 | 50.0% | 41.4% | 35.5% – 47.3% |
| 23 | Braelon Allen◦ provisional | NYJ | RB | 92 | 44.6% | 41.3% | 36.2% – 46.5% |
| 24 | Carson Steele◦ provisional | KC | RB | 56 | 46.4% | 41.2% | 35.7% – 46.9% |
| 25 | James Cook | BUF | RB | 207 | 42.5% | 41.2% | 36.9% – 45.5% |
| 26 | Austin Ekeler◦ provisional | WAS | RB | 77 | 44.2% | 40.9% | 35.7% – 46.3% |
| 27 | Brian Robinson | WAS | RB | 187 | 42.3% | 40.9% | 36.6% – 45.4% |
| 28 | Emari Demercado◦ provisional | ARI | RB | 24 | 50.0% | 40.8% | 34.8% – 46.9% |
| 29 | De'Von Achane | MIA | RB | 203 | 41.9% | 40.8% | 36.5% – 45.1% |
| 30 | Kendre Miller◦ provisional | NO | RB | 39 | 46.2% | 40.7% | 35.0% – 46.6% |
| 31 | JaMycal Hasty◦ provisional | NE | RB | 20 | 50.0% | 40.6% | 34.5% – 46.8% |
| 32 | Isaiah Davis◦ provisional | NYJ | RB | 30 | 46.7% | 40.5% | 34.6% – 46.6% |
| 33 | Patrick Taylor◦ provisional | SF | RB | 40 | 45.0% | 40.5% | 34.8% – 46.4% |
| 34 | Sean Tucker◦ provisional | TB | RB | 50 | 44.0% | 40.5% | 34.9% – 46.2% |
| 35 | Kenneth Walker III◦ provisional | SEA | RB | 153 | 41.2% | 40.3% | 35.7% – 44.9% |
| 36 | Hassan Haskins◦ provisional | LAC | RB | 34 | 44.1% | 40.2% | 34.4% – 46.2% |
| 37 | Ronnie Rivers◦ provisional | LA | RB | 22 | 45.5% | 40.1% | 34.1% – 46.3% |
| 38 | Julius Chestnut◦ provisional | TEN | RB | 22 | 45.5% | 40.1% | 34.1% – 46.3% |
| 39 | Justice Hill◦ provisional | BAL | RB | 47 | 42.5% | 40.1% | 34.5% – 45.8% |
| 40 | Aaron Jones | MIN | RB | 255 | 40.4% | 40.0% | 36.0% – 44.0% |
| 41 | Samaje Perine◦ provisional | KC | RB | 20 | 45.0% | 40.0% | 34.0% – 46.2% |
| 42 | Kenny McIntosh◦ provisional | SEA | RB | 31 | 41.9% | 39.8% | 33.9% – 45.8% |
| 43 | Jeff Wilson◦ provisional | MIA | RB | 16 | 43.8% | 39.8% | 33.7% – 46.0% |
| 44 | Alvin Kamara | NO | RB | 228 | 39.9% | 39.7% | 35.6% – 43.8% |
| 45 | Gus Edwards◦ provisional | LAC | RB | 102 | 40.2% | 39.7% | 34.7% – 44.8% |
| 46 | Jordan Mason◦ provisional | SF | RB | 153 | 39.9% | 39.6% | 35.0% – 44.2% |
| 47 | Cordarrelle Patterson◦ provisional | PIT | RB | 32 | 40.6% | 39.6% | 33.7% – 45.5% |
| 48 | Miles Sanders◦ provisional | CAR | RB | 55 | 40.0% | 39.5% | 34.0% – 45.1% |
| 49 | Christian McCaffrey◦ provisional | SF | RB | 50 | 40.0% | 39.5% | 33.9% – 45.2% |
| 50 | Michael Carter◦ provisional | ARI | RB | 35 | 40.0% | 39.5% | 33.7% – 45.4% |
| 51 | Audric Estimé◦ provisional | DEN | RB | 76 | 39.5% | 39.4% | 34.1% – 44.7% |
| 52 | Jerome Ford◦ provisional | CLE | RB | 104 | 39.4% | 39.4% | 34.4% – 44.4% |
| 53 | Ameer Abdullah◦ provisional | LV | RB | 66 | 39.4% | 39.4% | 34.0% – 44.8% |
| 54 | Ray Davis◦ provisional | BUF | RB | 113 | 38.9% | 39.2% | 34.3% – 44.1% |
| 55 | Josh Jacobs | GB | RB | 303 | 38.9% | 39.1% | 35.4% – 42.9% |
| 56 | Dameon Pierce◦ provisional | HOU | RB | 40 | 37.5% | 39.0% | 33.3% – 44.8% |
| 57 | Chase Brown | CIN | RB | 230 | 38.7% | 39.0% | 34.9% – 43.1% |
| 58 | Tyjae Spears◦ provisional | TEN | RB | 84 | 38.1% | 38.9% | 33.8% – 44.1% |
| 59 | Ty Chandler◦ provisional | MIN | RB | 56 | 37.5% | 38.9% | 33.4% – 44.4% |
| 60 | Raheem Blackshear◦ provisional | CAR | RB | 15 | 33.3% | 38.8% | 32.7% – 45.0% |
| 61 | Craig Reynolds◦ provisional | DET | RB | 31 | 35.5% | 38.7% | 32.9% – 44.6% |
| 62 | Tank Bigsby | JAX | RB | 168 | 38.1% | 38.7% | 34.3% – 43.2% |
| 63 | Pierre Strong◦ provisional | CLE | RB | 26 | 34.6% | 38.6% | 32.8% – 44.7% |
| 64 | Isiah Pacheco◦ provisional | KC | RB | 83 | 37.4% | 38.6% | 33.5% – 43.9% |
| 65 | Trey Benson◦ provisional | ARI | RB | 63 | 36.5% | 38.5% | 33.1% – 44.0% |
| 66 | D'Ernest Johnson◦ provisional | JAX | RB | 32 | 34.4% | 38.5% | 32.7% – 44.4% |
| 67 | Tyler Goodson◦ provisional | IND | RB | 32 | 34.4% | 38.5% | 32.7% – 44.4% |
| 68 | Isaac Guerendo◦ provisional | SF | RB | 84 | 36.9% | 38.5% | 33.3% – 43.7% |
| 69 | Jamaal Williams◦ provisional | NO | RB | 48 | 35.4% | 38.4% | 32.8% – 44.1% |
| 70 | Travis Etienne◦ provisional | JAX | RB | 150 | 37.3% | 38.4% | 33.8% – 43.0% |
| 71 | Jordan Mims◦ provisional | NO | RB | 20 | 30.0% | 38.3% | 32.3% – 44.4% |
| 72 | Zach Charbonnet◦ provisional | SEA | RB | 135 | 37.0% | 38.3% | 33.6% – 43.0% |
| 73 | Jaylen Wright◦ provisional | MIA | RB | 68 | 35.3% | 38.1% | 32.8% – 43.5% |
| 74 | Devin Singletary◦ provisional | NYG | RB | 113 | 36.3% | 38.0% | 33.2% – 43.0% |
| 75 | Blake Corum◦ provisional | LA | RB | 58 | 34.5% | 38.0% | 32.6% – 43.5% |
| 76 | Dare Ogunbowale◦ provisional | HOU | RB | 30 | 30.0% | 37.8% | 32.0% – 43.8% |
| 77 | Jaylen Warren◦ provisional | PIT | RB | 120 | 35.8% | 37.8% | 33.0% – 42.7% |
| 78 | Keaton Mitchell◦ provisional | BAL | RB | 15 | 20.0% | 37.6% | 31.6% – 43.8% |
| 79 | Breece Hall | NYJ | RB | 212 | 36.3% | 37.6% | 33.5% – 41.8% |
| 80 | Ezekiel Elliott◦ provisional | DAL | RB | 74 | 33.8% | 37.5% | 32.3% – 42.9% |
| 81 | Trey Sermon◦ provisional | IND | RB | 56 | 32.1% | 37.4% | 32.0% – 43.0% |
| 82 | Javonte Williams◦ provisional | DEN | RB | 139 | 35.3% | 37.4% | 32.8% – 42.1% |
| 83 | Nick Chubb◦ provisional | CLE | RB | 102 | 34.3% | 37.3% | 32.4% – 42.4% |
| 84 | Deuce Vaughn◦ provisional | DAL | RB | 17 | 17.6% | 37.2% | 31.2% – 43.3% |
| 85 | J.K. Dobbins | LAC | RB | 195 | 35.4% | 37.1% | 32.9% – 41.4% |
| 86 | Antonio Gibson◦ provisional | NE | RB | 120 | 34.2% | 37.1% | 32.3% – 41.9% |
| 87 | Rhamondre Stevenson | NE | RB | 207 | 35.3% | 37.0% | 32.9% – 41.2% |
| 88 | Jonathan Taylor | IND | RB | 305 | 35.7% | 36.9% | 33.3% – 40.7% |
| 89 | Kimani Vidal◦ provisional | LAC | RB | 43 | 27.9% | 36.8% | 31.3% – 42.6% |
| 90 | Najee Harris | PIT | RB | 263 | 35.4% | 36.8% | 33.0% – 40.8% |
| 91 | Tyrone Tracy Jr. | NYG | RB | 192 | 33.9% | 36.3% | 32.1% – 40.6% |
| 92 | Tony Pollard | TEN | RB | 261 | 34.5% | 36.3% | 32.4% – 40.2% |
| 93 | D'Onta Foreman◦ provisional | CLE | RB | 71 | 29.6% | 36.3% | 31.0% – 41.6% |
| 94 | Raheem Mostert◦ provisional | MIA | RB | 85 | 30.6% | 36.2% | 31.2% – 41.4% |
| 95 | Will Shipley◦ provisional | PHI | RB | 30 | 20.0% | 36.2% | 30.4% – 42.1% |
| 96 | Joe Mixon | HOU | RB | 247 | 34.0% | 36.0% | 32.1% – 40.0% |
| 97 | D'Andre Swift | CHI | RB | 253 | 34.0% | 36.0% | 32.1% – 40.0% |
| 98 | Cam Akers◦ provisional | MIN | RB | 104 | 30.8% | 35.9% | 31.0% – 40.8% |
| 99 | Alexander Mattison◦ provisional | LV | RB | 134 | 31.3% | 35.6% | 31.0% – 40.3% |
| 100 | Zack Moss◦ provisional | CIN | RB | 74 | 25.7% | 34.9% | 29.8% – 40.2% |
| 101 | Kenneth Gainwell◦ provisional | PHI | RB | 75 | 22.7% | 33.9% | 28.8% – 39.1% |
| 102 | Zamir White◦ provisional | LV | RB | 65 | 16.9% | 32.7% | 27.5% – 38.0% |
How the shrinkage works
Two estimators
Rate stats (completion %, success rate, catch rate) use a beta-binomial model: a Beta(α, β) prior fit by marginal likelihood over each position group, then a Beta posterior per player. Per-play averages (EPA, CPOE, yards) use a normal-normal model with DerSimonian–Laird between-player variance. Both pull each player toward their group by exactly how thin their sample is.
Does it help? & the fine print
The trust panel’s numbers come from a leakage-free odd/even-week holdout: fit on odd weeks, predict even-week raw. Shrinkage lowers out-of-sample error for every stat. Priors are fit per season and per position (WR and TE separately), so a TE’s baseline isn’t a skill. Regular season only, 2016–2025. Rushing and receiving efficiency are heavily scheme-driven — read the wide bands as the honesty they are.
Source: nflverse play-by-play. Counted and computed deterministically — never modeled by a language model. Built by build_nfl_leaderboards.py (byte-reproducible).