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.
Catch rate · 2023
Receptions per target, shrunk within position (WR vs TE separately). Depth-of-target confound; catchable-target rate not available here.
- Beats raw by
- 18.3%
- lower out-of-sample error
- RMSE raw → shrunk
- 11.1% → 9.1%
- odd vs. even weeks
- Split-half reliability
- 0.35
- how repeatable the raw stat is
- Stabilizes at
- 283
- 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.
Catch rate · 2023 · full board
| # | Player | Team | Pos | targets | Raw | Shrunk | 90% interval |
|---|---|---|---|---|---|---|---|
| 1 | Evan Engram◦ provisional | JAX | TE | 144 | 79.2% | 74.0% | 70.5% – 77.5% |
| 2 | Cole Kmet◦ provisional | CHI | TE | 90 | 81.1% | 73.8% | 70.0% – 77.4% |
| 3 | Dalton Kincaid◦ provisional | BUF | TE | 91 | 80.2% | 73.6% | 69.8% – 77.2% |
| 4 | Travis Kelce◦ provisional | KC | TE | 121 | 76.9% | 73.0% | 69.4% – 76.6% |
| 5 | Daniel Bellinger◦ provisional | NYG | TE | 28 | 89.3% | 73.0% | 68.8% – 77.1% |
| 6 | Taysom Hill◦ provisional | NO | TE | 40 | 82.5% | 72.8% | 68.7% – 76.8% |
| 7 | Trey McBride◦ provisional | ARI | TE | 106 | 76.4% | 72.8% | 69.0% – 76.4% |
| 8 | Durham Smythe◦ provisional | MIA | TE | 43 | 81.4% | 72.7% | 68.6% – 76.7% |
| 9 | T.J. Hockenson◦ provisional | MIN | TE | 127 | 74.8% | 72.5% | 68.8% – 76.0% |
| 10 | Foster Moreau◦ provisional | NO | TE | 25 | 84.0% | 72.4% | 68.2% – 76.5% |
| 11 | Brock Wright◦ provisional | DET | TE | 14 | 92.9% | 72.4% | 68.1% – 76.6% |
| 12 | Tanner Hudson◦ provisional | CIN | TE | 50 | 78.0% | 72.4% | 68.3% – 76.3% |
| 13 | Drew Sample◦ provisional | CIN | TE | 27 | 81.5% | 72.3% | 68.0% – 76.4% |
| 14 | Pharaoh Brown◦ provisional | NE | TE | 15 | 86.7% | 72.2% | 67.8% – 76.4% |
| 15 | Luke Farrell◦ provisional | JAX | TE | 15 | 86.7% | 72.2% | 67.8% – 76.4% |
| 16 | Tucker Kraft◦ provisional | GB | TE | 40 | 77.5% | 72.2% | 68.0% – 76.2% |
| 17 | Austin Hooper◦ provisional | LV | TE | 32 | 78.1% | 72.1% | 67.9% – 76.2% |
| 18 | Brevin Jordan◦ provisional | HOU | TE | 21 | 81.0% | 72.1% | 67.8% – 76.2% |
| 19 | Josh Oliver◦ provisional | MIN | TE | 28 | 78.6% | 72.1% | 67.8% – 76.2% |
| 20 | Isaiah Likely◦ provisional | BAL | TE | 40 | 75.0% | 71.9% | 67.7% – 75.9% |
| 21 | Will Dissly◦ provisional | SEA | TE | 22 | 77.3% | 71.9% | 67.5% – 76.0% |
| 22 | Mark Andrews◦ provisional | BAL | TE | 61 | 73.8% | 71.8% | 67.8% – 75.8% |
| 23 | Noah Fant◦ provisional | SEA | TE | 43 | 74.4% | 71.8% | 67.7% – 75.8% |
| 24 | Luke Musgrave◦ provisional | GB | TE | 46 | 73.9% | 71.8% | 67.6% – 75.8% |
| 25 | Gerald Everett◦ provisional | LAC | TE | 70 | 72.9% | 71.7% | 67.7% – 75.6% |
| 26 | Colby Parkinson◦ provisional | SEA | TE | 34 | 73.5% | 71.7% | 67.4% – 75.7% |
| 27 | George Kittle◦ provisional | SF | TE | 90 | 72.2% | 71.6% | 67.7% – 75.4% |
| 28 | Johnny Mundt◦ provisional | MIN | TE | 23 | 73.9% | 71.6% | 67.3% – 75.8% |
| 29 | MyCole Pruitt◦ provisional | ATL | TE | 12 | 75.0% | 71.6% | 67.2% – 75.8% |
| 30 | Mitchell Wilcox◦ provisional | CIN | TE | 12 | 75.0% | 71.6% | 67.2% – 75.8% |
| 31 | Elijah Higgins◦ provisional | ARI | TE | 19 | 73.7% | 71.6% | 67.2% – 75.7% |
| 32 | Cole Turner◦ provisional | WAS | TE | 15 | 73.3% | 71.5% | 67.2% – 75.7% |
| 33 | Jeremy Ruckert◦ provisional | NYJ | TE | 22 | 72.7% | 71.5% | 67.2% – 75.7% |
| 34 | Tommy Tremble◦ provisional | CAR | TE | 32 | 71.9% | 71.5% | 67.2% – 75.6% |
| 35 | Jonnu Smith◦ provisional | ATL | TE | 70 | 71.4% | 71.4% | 67.4% – 75.3% |
| 36 | Dallas Goedert◦ provisional | PHI | TE | 83 | 71.1% | 71.4% | 67.4% – 75.2% |
| 37 | Irv Smith◦ provisional | CIN | TE | 26 | 69.2% | 71.2% | 66.9% – 75.4% |
| 38 | Will Mallory◦ provisional | IND | TE | 26 | 69.2% | 71.2% | 66.9% – 75.4% |
| 39 | C.J. Uzomah◦ provisional | NYJ | TE | 12 | 66.7% | 71.2% | 66.8% – 75.5% |
| 40 | Darren Waller◦ provisional | NYG | TE | 74 | 70.3% | 71.2% | 67.2% – 75.0% |
| 41 | Chig Okonkwo◦ provisional | TEN | TE | 77 | 70.1% | 71.2% | 67.2% – 75.0% |
| 42 | Sam LaPorta◦ provisional | DET | TE | 122 | 70.5% | 71.1% | 67.4% – 74.8% |
| 43 | Tyler Conklin◦ provisional | NYJ | TE | 87 | 70.1% | 71.1% | 67.2% – 74.9% |
| 44 | John Bates◦ provisional | WAS | TE | 28 | 67.9% | 71.1% | 66.8% – 75.2% |
| 45 | Robert Tonyan◦ provisional | CHI | TE | 17 | 64.7% | 71.0% | 66.7% – 75.3% |
| 46 | Logan Thomas◦ provisional | WAS | TE | 79 | 69.6% | 71.0% | 67.0% – 74.9% |
| 47 | Noah Gray◦ provisional | KC | TE | 41 | 68.3% | 71.0% | 66.8% – 75.1% |
| 48 | Connor Heyward◦ provisional | PIT | TE | 34 | 67.7% | 71.0% | 66.8% – 75.1% |
| 49 | Harrison Bryant◦ provisional | CLE | TE | 20 | 65.0% | 71.0% | 66.6% – 75.2% |
| 50 | Cade Otton◦ provisional | TB | TE | 68 | 69.1% | 71.0% | 66.9% – 74.9% |
| 51 | Hunter Henry◦ provisional | NE | TE | 61 | 68.8% | 71.0% | 66.9% – 74.9% |
| 52 | Jordan Akins◦ provisional | CLE | TE | 23 | 65.2% | 71.0% | 66.6% – 75.1% |
| 53 | Jake Ferguson◦ provisional | DAL | TE | 102 | 69.6% | 70.9% | 67.1% – 74.7% |
| 54 | Josh Whyle◦ provisional | TEN | TE | 15 | 60.0% | 70.9% | 66.5% – 75.1% |
| 55 | Lucas Krull◦ provisional | DEN | TE | 14 | 57.1% | 70.8% | 66.3% – 75.0% |
| 56 | Pat Freiermuth◦ provisional | PIT | TE | 48 | 66.7% | 70.7% | 66.6% – 74.8% |
| 57 | Donald Parham◦ provisional | LAC | TE | 41 | 65.8% | 70.7% | 66.5% – 74.8% |
| 58 | Michael Mayer◦ provisional | LV | TE | 41 | 65.8% | 70.7% | 66.5% – 74.8% |
| 59 | Tyler Higbee◦ provisional | LA | TE | 70 | 67.1% | 70.6% | 66.5% – 74.5% |
| 60 | Luke Schoonmaker◦ provisional | DAL | TE | 15 | 53.3% | 70.5% | 66.1% – 74.8% |
| 61 | Adam Trautman◦ provisional | DEN | TE | 35 | 62.9% | 70.5% | 66.2% – 74.6% |
| 62 | Mo Alie-Cox◦ provisional | IND | TE | 23 | 56.5% | 70.3% | 65.9% – 74.5% |
| 63 | Zach Ertz◦ provisional | ARI | TE | 43 | 62.8% | 70.3% | 66.1% – 74.4% |
| 64 | Mike Gesicki◦ provisional | NE | TE | 46 | 63.0% | 70.3% | 66.0% – 74.3% |
| 65 | Dalton Schultz◦ provisional | HOU | TE | 89 | 66.3% | 70.2% | 66.2% – 74.0% |
| 66 | Stone Smartt◦ provisional | LAC | TE | 21 | 52.4% | 70.1% | 65.7% – 74.3% |
| 67 | Dawson Knox◦ provisional | BUF | TE | 37 | 59.5% | 70.0% | 65.8% – 74.2% |
| 68 | Juwan Johnson◦ provisional | NO | TE | 59 | 62.7% | 69.9% | 65.8% – 73.9% |
| 69 | Hayden Hurst◦ provisional | CAR | TE | 32 | 56.3% | 69.9% | 65.6% – 74.1% |
| 70 | Stephen Sullivan◦ provisional | CAR | TE | 24 | 50.0% | 69.8% | 65.4% – 74.0% |
| 71 | David Njoku◦ provisional | CLE | TE | 123 | 65.8% | 69.7% | 65.9% – 73.4% |
| 72 | Kylen Granson◦ provisional | IND | TE | 50 | 60.0% | 69.7% | 65.5% – 73.8% |
| 73 | Andrew Ogletree◦ provisional | IND | TE | 21 | 42.9% | 69.5% | 65.0% – 73.7% |
| 74 | Kyle Pitts◦ provisional | ATL | TE | 90 | 58.9% | 68.4% | 64.4% – 72.3% |
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).