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.
Target success rate · 2024
Share of targets with positive EPA, shrunk within position.
- Beats raw by
- 21.9%
- lower out-of-sample error
- RMSE raw → shrunk
- 11.7% → 9.1%
- odd vs. even weeks
- Split-half reliability
- 0.19
- how repeatable the raw stat is
- Stabilizes at
- 52
- 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.
Target success rate · 2024 · full board
| # | Player | Team | Pos | targets | Raw | Shrunk | 90% interval |
|---|---|---|---|---|---|---|---|
| 1 | George Kittle | SF | TE | 94 | 74.5% | 67.4% | 60.9% – 73.6% |
| 2 | Mark Andrews | BAL | TE | 69 | 71.0% | 63.9% | 56.7% – 70.9% |
| 3 | Noah Gray◦ provisional | KC | TE | 49 | 69.4% | 61.8% | 53.7% – 69.5% |
| 4 | Jonnu Smith | MIA | TE | 111 | 64.9% | 61.6% | 55.3% – 67.7% |
| 5 | Trey McBride | ARI | TE | 147 | 62.6% | 60.5% | 54.8% – 66.1% |
| 6 | Mike Gesicki | CIN | TE | 83 | 63.9% | 60.3% | 53.3% – 67.1% |
| 7 | Cole Kmet | CHI | TE | 59 | 64.4% | 59.8% | 52.1% – 67.3% |
| 8 | Isaiah Likely | BAL | TE | 58 | 63.8% | 59.4% | 51.7% – 67.0% |
| 9 | Grant Calcaterra◦ provisional | PHI | TE | 30 | 66.7% | 59.0% | 50.0% – 67.7% |
| 10 | Zach Ertz | WAS | TE | 93 | 61.3% | 58.9% | 52.1% – 65.5% |
| 11 | Nate Adkins◦ provisional | DEN | TE | 15 | 73.3% | 58.8% | 48.8% – 68.4% |
| 12 | Elijah Higgins◦ provisional | ARI | TE | 24 | 66.7% | 58.4% | 49.0% – 67.5% |
| 13 | Foster Moreau◦ provisional | NO | TE | 43 | 62.8% | 58.3% | 49.9% – 66.5% |
| 14 | Payne Durham◦ provisional | TB | TE | 14 | 71.4% | 58.1% | 48.1% – 67.9% |
| 15 | Tucker Kraft | GB | TE | 71 | 60.6% | 58.0% | 50.7% – 65.2% |
| 16 | Josh Oliver◦ provisional | MIN | TE | 28 | 64.3% | 58.0% | 48.9% – 66.9% |
| 17 | Brock Wright◦ provisional | DET | TE | 16 | 68.8% | 57.9% | 48.0% – 67.5% |
| 18 | Nick Vannett◦ provisional | TEN | TE | 20 | 65.0% | 57.5% | 47.9% – 66.8% |
| 19 | Austin Hooper | NE | TE | 59 | 59.3% | 57.1% | 49.4% – 64.7% |
| 20 | Stone Smartt◦ provisional | LAC | TE | 19 | 63.2% | 56.9% | 47.2% – 66.3% |
| 21 | Eric Saubert◦ provisional | SF | TE | 14 | 64.3% | 56.6% | 46.6% – 66.5% |
| 22 | Brenton Strange | JAX | TE | 53 | 58.5% | 56.6% | 48.6% – 64.4% |
| 23 | Pat Freiermuth | PIT | TE | 78 | 57.7% | 56.5% | 49.3% – 63.5% |
| 24 | Darnell Washington◦ provisional | PIT | TE | 25 | 60.0% | 56.4% | 47.0% – 65.5% |
| 25 | Brock Bowers | LV | TE | 153 | 56.9% | 56.3% | 50.6% – 61.9% |
| 26 | Dallas Goedert◦ provisional | PHI | TE | 52 | 57.7% | 56.1% | 48.1% – 64.0% |
| 27 | Luke Schoonmaker◦ provisional | DAL | TE | 36 | 58.3% | 56.1% | 47.4% – 64.7% |
| 28 | AJ Barner◦ provisional | SEA | TE | 38 | 57.9% | 56.0% | 47.4% – 64.5% |
| 29 | Drew Sample◦ provisional | CIN | TE | 22 | 59.1% | 55.9% | 46.4% – 65.3% |
| 30 | Tanner Hudson◦ provisional | CIN | TE | 24 | 58.3% | 55.8% | 46.4% – 65.0% |
| 31 | Daniel Bellinger◦ provisional | NYG | TE | 17 | 58.8% | 55.6% | 45.8% – 65.3% |
| 32 | Travis Kelce | KC | TE | 134 | 56.0% | 55.6% | 49.6% – 61.5% |
| 33 | Noah Fant | SEA | TE | 64 | 56.3% | 55.5% | 47.9% – 63.0% |
| 34 | Harrison Bryant◦ provisional | LV | TE | 12 | 58.3% | 55.3% | 45.1% – 65.3% |
| 35 | Brevyn Spann-Ford◦ provisional | DAL | TE | 14 | 57.1% | 55.1% | 45.1% – 65.0% |
| 36 | Sam LaPorta | DET | TE | 83 | 55.4% | 55.1% | 48.0% – 62.1% |
| 37 | T.J. Hockenson | MIN | TE | 62 | 54.8% | 54.7% | 47.1% – 62.3% |
| 38 | Julian Hill◦ provisional | MIA | TE | 20 | 55.0% | 54.7% | 45.1% – 64.2% |
| 39 | Hunter Henry | NE | TE | 97 | 54.6% | 54.6% | 47.9% – 61.3% |
| 40 | Dawson Knox◦ provisional | BUF | TE | 33 | 54.5% | 54.6% | 45.7% – 63.3% |
| 41 | Theo Johnson◦ provisional | NYG | TE | 43 | 53.5% | 54.1% | 45.7% – 62.4% |
| 42 | Jordan Akins | CLE | TE | 58 | 53.4% | 54.0% | 46.2% – 61.7% |
| 43 | Lucas Krull◦ provisional | DEN | TE | 23 | 52.2% | 53.9% | 44.4% – 63.2% |
| 44 | Will Dissly | LAC | TE | 64 | 53.1% | 53.8% | 46.2% – 61.3% |
| 45 | Tyler Higbee◦ provisional | LA | TE | 12 | 50.0% | 53.8% | 43.5% – 63.8% |
| 46 | Juwan Johnson | NO | TE | 68 | 52.9% | 53.7% | 46.2% – 61.1% |
| 47 | Erick All◦ provisional | CIN | TE | 22 | 50.0% | 53.2% | 43.7% – 62.6% |
| 48 | John Bates◦ provisional | WAS | TE | 13 | 46.2% | 52.9% | 42.8% – 63.0% |
| 49 | Colby Parkinson◦ provisional | LA | TE | 49 | 51.0% | 52.9% | 44.7% – 61.0% |
| 50 | Tommy Tremble◦ provisional | CAR | TE | 32 | 50.0% | 52.9% | 43.9% – 61.7% |
| 51 | Josh Whyle◦ provisional | TEN | TE | 38 | 50.0% | 52.7% | 44.0% – 61.2% |
| 52 | Mo Alie-Cox◦ provisional | IND | TE | 23 | 47.8% | 52.5% | 43.1% – 61.9% |
| 53 | Johnny Mundt◦ provisional | MIN | TE | 27 | 48.1% | 52.4% | 43.2% – 61.5% |
| 54 | Chig Okonkwo | TEN | TE | 71 | 50.7% | 52.4% | 45.0% – 59.7% |
| 55 | Pharaoh Brown◦ provisional | SEA | TE | 12 | 41.7% | 52.2% | 42.0% – 62.3% |
| 56 | Charlie Woerner◦ provisional | ATL | TE | 12 | 41.7% | 52.2% | 42.0% – 62.3% |
| 57 | Andrew Ogletree◦ provisional | IND | TE | 14 | 42.9% | 52.1% | 42.1% – 62.1% |
| 58 | Ja'Tavion Sanders◦ provisional | CAR | TE | 43 | 48.8% | 52.0% | 43.6% – 60.4% |
| 59 | Tyler Conklin | NYJ | TE | 73 | 49.3% | 51.5% | 44.2% – 58.8% |
| 60 | Hayden Hurst◦ provisional | LAC | TE | 13 | 38.5% | 51.4% | 41.3% – 61.5% |
| 61 | Luke Farrell◦ provisional | JAX | TE | 17 | 41.2% | 51.3% | 41.5% – 61.1% |
| 62 | Kyle Pitts | ATL | TE | 74 | 48.6% | 51.1% | 43.8% – 58.4% |
| 63 | Dalton Kincaid | BUF | TE | 75 | 48.0% | 50.7% | 43.5% – 58.0% |
| 64 | Adam Trautman◦ provisional | DEN | TE | 22 | 40.9% | 50.6% | 41.1% – 60.0% |
| 65 | Michael Mayer◦ provisional | LV | TE | 32 | 43.8% | 50.5% | 41.6% – 59.4% |
| 66 | Davis Allen◦ provisional | LA | TE | 13 | 30.8% | 49.9% | 39.8% – 60.0% |
| 67 | Cade Otton | TB | TE | 87 | 46.0% | 49.2% | 42.3% – 56.2% |
| 68 | Cade Stover◦ provisional | HOU | TE | 22 | 36.4% | 49.2% | 39.8% – 58.7% |
| 69 | Taysom Hill◦ provisional | NO | TE | 31 | 38.7% | 48.7% | 39.8% – 57.7% |
| 70 | Durham Smythe◦ provisional | MIA | TE | 17 | 29.4% | 48.4% | 38.7% – 58.3% |
| 71 | Jeremy Ruckert◦ provisional | NYJ | TE | 28 | 35.7% | 48.0% | 38.9% – 57.2% |
| 72 | Dalton Schultz | HOU | TE | 85 | 43.5% | 47.8% | 40.8% – 54.8% |
| 73 | Kylen Granson◦ provisional | IND | TE | 31 | 35.5% | 47.5% | 38.6% – 56.5% |
| 74 | Gerald Everett◦ provisional | CHI | TE | 13 | 15.4% | 46.8% | 36.8% – 57.0% |
| 75 | Greg Dulcich◦ provisional | NYG | TE | 12 | 8.3% | 46.0% | 35.9% – 56.2% |
| 76 | Evan Engram | JAX | TE | 67 | 37.3% | 44.9% | 37.5% – 52.4% |
| 77 | Jake Ferguson | DAL | TE | 86 | 37.2% | 43.8% | 36.9% – 50.8% |
| 78 | David Njoku | CLE | TE | 99 | 37.4% | 43.3% | 36.8% – 50.0% |
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).