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.
EPA per target · 2024
Mean EPA per target, shrunk within position.
- Beats raw by
- 24.4%
- lower out-of-sample error
- RMSE raw → shrunk
- 0.38 → 0.29
- odd vs. even weeks
- Split-half reliability
- 0.13
- how repeatable the raw stat is
- Stabilizes at
- 164
- targets 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.
EPA per target · 2024 · full board
| # | Player | Team | Pos | targets | Raw | Shrunk | 90% interval |
|---|---|---|---|---|---|---|---|
| 1 | George Kittle◦ provisional | SF | TE | 94 | +0.72 | +0.42 | +0.27 – +0.57 |
| 2 | Mark Andrews◦ provisional | BAL | TE | 69 | +0.73 | +0.39 | +0.24 – +0.55 |
| 3 | Jonnu Smith◦ provisional | MIA | TE | 111 | +0.54 | +0.37 | +0.22 – +0.51 |
| 4 | Tucker Kraft◦ provisional | GB | TE | 71 | +0.52 | +0.33 | +0.18 – +0.49 |
| 5 | Pat Freiermuth◦ provisional | PIT | TE | 78 | +0.46 | +0.32 | +0.17 – +0.47 |
| 6 | Isaiah Likely◦ provisional | BAL | TE | 58 | +0.48 | +0.31 | +0.15 – +0.47 |
| 7 | Mike Gesicki◦ provisional | CIN | TE | 83 | +0.41 | +0.30 | +0.15 – +0.46 |
| 8 | Josh Oliver◦ provisional | MIN | TE | 28 | +0.61 | +0.30 | +0.13 – +0.48 |
| 9 | Noah Gray◦ provisional | KC | TE | 49 | +0.47 | +0.30 | +0.14 – +0.47 |
| 10 | Cole Kmet◦ provisional | CHI | TE | 59 | +0.44 | +0.30 | +0.14 – +0.46 |
| 11 | Foster Moreau◦ provisional | NO | TE | 43 | +0.46 | +0.30 | +0.13 – +0.46 |
| 12 | Grant Calcaterra◦ provisional | PHI | TE | 30 | +0.53 | +0.30 | +0.12 – +0.47 |
| 13 | Sam LaPorta◦ provisional | DET | TE | 83 | +0.38 | +0.30 | +0.14 – +0.45 |
| 14 | Payne Durham◦ provisional | TB | TE | 14 | +0.74 | +0.29 | +0.11 – +0.47 |
| 15 | Trey McBride◦ provisional | ARI | TE | 147 | +0.33 | +0.29 | +0.15 – +0.43 |
| 16 | Nate Adkins◦ provisional | DEN | TE | 15 | +0.61 | +0.28 | +0.10 – +0.46 |
| 17 | Zach Ertz◦ provisional | WAS | TE | 93 | +0.33 | +0.28 | +0.13 – +0.43 |
| 18 | Dallas Goedert◦ provisional | PHI | TE | 52 | +0.37 | +0.28 | +0.12 – +0.45 |
| 19 | Brenton Strange◦ provisional | JAX | TE | 53 | +0.36 | +0.28 | +0.12 – +0.44 |
| 20 | Elijah Higgins◦ provisional | ARI | TE | 24 | +0.40 | +0.27 | +0.10 – +0.45 |
| 21 | Stone Smartt◦ provisional | LAC | TE | 19 | +0.39 | +0.27 | +0.09 – +0.44 |
| 22 | Darnell Washington◦ provisional | PIT | TE | 25 | +0.34 | +0.26 | +0.09 – +0.44 |
| 23 | Tyler Higbee◦ provisional | LA | TE | 12 | +0.41 | +0.26 | +0.08 – +0.44 |
| 24 | Dawson Knox◦ provisional | BUF | TE | 33 | +0.31 | +0.26 | +0.09 – +0.43 |
| 25 | AJ Barner◦ provisional | SEA | TE | 38 | +0.29 | +0.26 | +0.09 – +0.43 |
| 26 | Mo Alie-Cox◦ provisional | IND | TE | 23 | +0.30 | +0.26 | +0.08 – +0.43 |
| 27 | Daniel Bellinger◦ provisional | NYG | TE | 17 | +0.31 | +0.26 | +0.08 – +0.44 |
| 28 | Ja'Tavion Sanders◦ provisional | CAR | TE | 43 | +0.27 | +0.26 | +0.09 – +0.42 |
| 29 | Adam Trautman◦ provisional | DEN | TE | 22 | +0.27 | +0.25 | +0.08 – +0.43 |
| 30 | Luke Schoonmaker◦ provisional | DAL | TE | 36 | +0.26 | +0.25 | +0.08 – +0.42 |
| 31 | Hunter Henry◦ provisional | NE | TE | 97 | +0.25 | +0.25 | +0.10 – +0.40 |
| 32 | Juwan Johnson◦ provisional | NO | TE | 68 | +0.25 | +0.25 | +0.09 – +0.41 |
| 33 | Erick All◦ provisional | CIN | TE | 22 | +0.24 | +0.25 | +0.07 – +0.43 |
| 34 | Josh Whyle◦ provisional | TEN | TE | 38 | +0.24 | +0.25 | +0.08 – +0.42 |
| 35 | Will Dissly◦ provisional | LAC | TE | 64 | +0.24 | +0.25 | +0.09 – +0.41 |
| 36 | Tanner Hudson◦ provisional | CIN | TE | 24 | +0.22 | +0.25 | +0.07 – +0.42 |
| 37 | Drew Sample◦ provisional | CIN | TE | 22 | +0.22 | +0.25 | +0.07 – +0.42 |
| 38 | Cade Stover◦ provisional | HOU | TE | 22 | +0.21 | +0.25 | +0.07 – +0.42 |
| 39 | Brock Wright◦ provisional | DET | TE | 16 | +0.19 | +0.25 | +0.07 – +0.43 |
| 40 | Lucas Krull◦ provisional | DEN | TE | 23 | +0.20 | +0.25 | +0.07 – +0.42 |
| 41 | Eric Saubert◦ provisional | SF | TE | 14 | +0.14 | +0.24 | +0.06 – +0.42 |
| 42 | Julian Hill◦ provisional | MIA | TE | 20 | +0.17 | +0.24 | +0.07 – +0.42 |
| 43 | Austin Hooper◦ provisional | NE | TE | 59 | +0.22 | +0.24 | +0.08 – +0.40 |
| 44 | Johnny Mundt◦ provisional | MIN | TE | 27 | +0.18 | +0.24 | +0.07 – +0.42 |
| 45 | Brevyn Spann-Ford◦ provisional | DAL | TE | 14 | +0.10 | +0.24 | +0.06 – +0.42 |
| 46 | Andrew Ogletree◦ provisional | IND | TE | 14 | +0.09 | +0.24 | +0.06 – +0.42 |
| 47 | Tommy Tremble◦ provisional | CAR | TE | 32 | +0.16 | +0.24 | +0.07 – +0.41 |
| 48 | Jordan Akins◦ provisional | CLE | TE | 58 | +0.19 | +0.24 | +0.07 – +0.40 |
| 49 | Dalton Kincaid◦ provisional | BUF | TE | 75 | +0.20 | +0.24 | +0.08 – +0.39 |
| 50 | Nick Vannett◦ provisional | TEN | TE | 20 | +0.09 | +0.24 | +0.06 – +0.41 |
| 51 | Harrison Bryant◦ provisional | LV | TE | 12 | -0.02 | +0.23 | +0.05 – +0.41 |
| 52 | John Bates◦ provisional | WAS | TE | 13 | -0.05 | +0.23 | +0.05 – +0.41 |
| 53 | Gerald Everett◦ provisional | CHI | TE | 13 | -0.11 | +0.23 | +0.05 – +0.41 |
| 54 | T.J. Hockenson◦ provisional | MIN | TE | 62 | +0.15 | +0.23 | +0.07 – +0.39 |
| 55 | Theo Johnson◦ provisional | NYG | TE | 43 | +0.11 | +0.22 | +0.06 – +0.39 |
| 56 | Tyler Conklin◦ provisional | NYJ | TE | 73 | +0.15 | +0.22 | +0.07 – +0.38 |
| 57 | Taysom Hill◦ provisional | NO | TE | 31 | +0.04 | +0.22 | +0.05 – +0.39 |
| 58 | Brock Bowers◦ provisional | LV | TE | 153 | +0.18 | +0.22 | +0.08 – +0.35 |
| 59 | Cade Otton◦ provisional | TB | TE | 87 | +0.14 | +0.22 | +0.06 – +0.37 |
| 60 | Davis Allen◦ provisional | LA | TE | 13 | -0.31 | +0.21 | +0.03 – +0.39 |
| 61 | Michael Mayer◦ provisional | LV | TE | 32 | -0.00 | +0.21 | +0.04 – +0.38 |
| 62 | Noah Fant◦ provisional | SEA | TE | 64 | +0.10 | +0.21 | +0.05 – +0.37 |
| 63 | Colby Parkinson◦ provisional | LA | TE | 49 | +0.07 | +0.21 | +0.05 – +0.37 |
| 64 | Durham Smythe◦ provisional | MIA | TE | 17 | -0.21 | +0.21 | +0.03 – +0.39 |
| 65 | Travis Kelce◦ provisional | KC | TE | 134 | +0.15 | +0.21 | +0.07 – +0.35 |
| 66 | Jeremy Ruckert◦ provisional | NYJ | TE | 28 | -0.06 | +0.21 | +0.03 – +0.38 |
| 67 | Hayden Hurst◦ provisional | LAC | TE | 13 | -0.37 | +0.21 | +0.03 – +0.39 |
| 68 | Luke Farrell◦ provisional | JAX | TE | 17 | -0.27 | +0.20 | +0.02 – +0.38 |
| 69 | Kyle Pitts◦ provisional | ATL | TE | 74 | +0.09 | +0.20 | +0.05 – +0.36 |
| 70 | Charlie Woerner◦ provisional | ATL | TE | 12 | -0.48 | +0.20 | +0.02 – +0.38 |
| 71 | Pharaoh Brown◦ provisional | SEA | TE | 12 | -0.54 | +0.20 | +0.02 – +0.38 |
| 72 | Greg Dulcich◦ provisional | NYG | TE | 12 | -0.63 | +0.19 | +0.01 – +0.37 |
| 73 | Dalton Schultz◦ provisional | HOU | TE | 85 | +0.04 | +0.18 | +0.03 – +0.33 |
| 74 | Evan Engram◦ provisional | JAX | TE | 67 | -0.01 | +0.18 | +0.02 – +0.33 |
| 75 | Kylen Granson◦ provisional | IND | TE | 31 | -0.27 | +0.17 | -0.00 – +0.34 |
| 76 | Chig Okonkwo◦ provisional | TEN | TE | 71 | -0.10 | +0.15 | -0.01 – +0.30 |
| 77 | David Njoku◦ provisional | CLE | TE | 99 | -0.05 | +0.14 | -0.01 – +0.29 |
| 78 | Jake Ferguson◦ provisional | DAL | TE | 86 | -0.15 | +0.11 | -0.04 – +0.27 |
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).