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 · 2022
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
- 146
- 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 · 2022 · full board
| # | Player | Team | Pos | targets | Raw | Shrunk | 90% interval |
|---|---|---|---|---|---|---|---|
| 1 | Dallas Goedert◦ provisional | PHI | TE | 69 | 72.5% | 58.3% | 52.8% – 63.8% |
| 2 | Travis Kelce | KC | TE | 153 | 60.8% | 56.3% | 51.6% – 61.0% |
| 3 | Will Dissly◦ provisional | SEA | TE | 38 | 73.7% | 56.2% | 50.2% – 62.2% |
| 4 | Evan Engram◦ provisional | JAX | TE | 101 | 59.4% | 54.8% | 49.6% – 60.0% |
| 5 | Dawson Knox◦ provisional | BUF | TE | 65 | 60.0% | 54.2% | 48.6% – 59.8% |
| 6 | Austin Hooper◦ provisional | TEN | TE | 60 | 60.0% | 54.1% | 48.4% – 59.8% |
| 7 | Donald Parham◦ provisional | LAC | TE | 12 | 83.3% | 54.1% | 47.5% – 60.5% |
| 8 | Mark Andrews◦ provisional | BAL | TE | 114 | 57.0% | 54.0% | 48.9% – 59.1% |
| 9 | David Njoku◦ provisional | CLE | TE | 80 | 57.5% | 53.7% | 48.3% – 59.2% |
| 10 | Noah Gray◦ provisional | KC | TE | 34 | 61.8% | 53.6% | 47.4% – 59.7% |
| 11 | Daniel Bellinger◦ provisional | NYG | TE | 36 | 61.1% | 53.5% | 47.4% – 59.6% |
| 12 | Mitchell Wilcox◦ provisional | CIN | TE | 18 | 66.7% | 53.3% | 46.9% – 59.7% |
| 13 | Adam Trautman◦ provisional | NO | TE | 22 | 63.6% | 53.2% | 46.9% – 59.5% |
| 14 | Brock Wright◦ provisional | DET | TE | 24 | 62.5% | 53.2% | 46.9% – 59.5% |
| 15 | Hayden Hurst◦ provisional | CIN | TE | 68 | 55.9% | 53.0% | 47.4% – 58.6% |
| 16 | Cole Kmet◦ provisional | CHI | TE | 70 | 55.7% | 53.0% | 47.4% – 58.5% |
| 17 | George Kittle◦ provisional | SF | TE | 86 | 54.6% | 52.8% | 47.4% – 58.1% |
| 18 | Jack Stoll◦ provisional | PHI | TE | 14 | 64.3% | 52.8% | 46.3% – 59.2% |
| 19 | Chig Okonkwo◦ provisional | TEN | TE | 47 | 55.3% | 52.5% | 46.6% – 58.4% |
| 20 | Eric Tomlinson◦ provisional | DEN | TE | 13 | 61.5% | 52.5% | 45.9% – 59.0% |
| 21 | Dan Arnold◦ provisional | JAX | TE | 13 | 61.5% | 52.5% | 45.9% – 59.0% |
| 22 | Joseph Fortson◦ provisional | KC | TE | 13 | 61.5% | 52.5% | 45.9% – 59.0% |
| 23 | Colby Parkinson◦ provisional | SEA | TE | 34 | 55.9% | 52.4% | 46.3% – 58.6% |
| 24 | Connor Heyward◦ provisional | PIT | TE | 17 | 58.8% | 52.4% | 46.0% – 58.8% |
| 25 | MyCole Pruitt◦ provisional | ATL | TE | 21 | 57.1% | 52.3% | 46.0% – 58.7% |
| 26 | Johnny Mundt◦ provisional | MIN | TE | 21 | 57.1% | 52.3% | 46.0% – 58.7% |
| 27 | Juwan Johnson◦ provisional | NO | TE | 65 | 53.8% | 52.3% | 46.7% – 58.0% |
| 28 | C.J. Uzomah◦ provisional | NYJ | TE | 27 | 55.6% | 52.3% | 46.0% – 58.5% |
| 29 | Jake Ferguson◦ provisional | DAL | TE | 22 | 54.5% | 52.0% | 45.7% – 58.4% |
| 30 | Noah Fant◦ provisional | SEA | TE | 63 | 52.4% | 51.9% | 46.2% – 57.5% |
| 31 | Zach Ertz◦ provisional | ARI | TE | 71 | 52.1% | 51.8% | 46.2% – 57.4% |
| 32 | Tanner Hudson◦ provisional | NYG | TE | 15 | 53.3% | 51.8% | 45.3% – 58.3% |
| 33 | Josiah Deguara◦ provisional | GB | TE | 15 | 53.3% | 51.8% | 45.3% – 58.3% |
| 34 | Pat Freiermuth◦ provisional | PIT | TE | 98 | 52.0% | 51.8% | 46.5% – 57.0% |
| 35 | Foster Moreau◦ provisional | LV | TE | 54 | 51.8% | 51.7% | 45.9% – 57.5% |
| 36 | Jordan Akins◦ provisional | HOU | TE | 56 | 51.8% | 51.7% | 45.9% – 57.5% |
| 37 | Dalton Schultz◦ provisional | DAL | TE | 89 | 51.7% | 51.7% | 46.3% – 57.0% |
| 38 | Peyton Hendershot◦ provisional | DAL | TE | 16 | 50.0% | 51.5% | 45.0% – 57.9% |
| 39 | Durham Smythe◦ provisional | MIA | TE | 20 | 50.0% | 51.4% | 45.1% – 57.8% |
| 40 | Lawrence Cager◦ provisional | NYG | TE | 20 | 50.0% | 51.4% | 45.1% – 57.8% |
| 41 | Greg Dulcich◦ provisional | DEN | TE | 55 | 50.9% | 51.4% | 45.6% – 57.2% |
| 42 | Robert Tonyan◦ provisional | GB | TE | 67 | 50.7% | 51.4% | 45.7% – 57.0% |
| 43 | Ian Thomas◦ provisional | CAR | TE | 30 | 50.0% | 51.4% | 45.2% – 57.6% |
| 44 | Kylen Granson◦ provisional | IND | TE | 40 | 50.0% | 51.3% | 45.3% – 57.3% |
| 45 | Taysom Hill◦ provisional | NO | TE | 13 | 46.2% | 51.2% | 44.7% – 57.7% |
| 46 | Anthony Firkser◦ provisional | ATL | TE | 13 | 46.2% | 51.2% | 44.7% – 57.7% |
| 47 | Trey McBride◦ provisional | ARI | TE | 39 | 48.7% | 51.0% | 45.0% – 57.1% |
| 48 | Harrison Bryant◦ provisional | CLE | TE | 43 | 48.8% | 51.0% | 45.0% – 57.0% |
| 49 | Teagan Quitoriano◦ provisional | HOU | TE | 14 | 42.9% | 50.9% | 44.4% – 57.4% |
| 50 | Albert Okwuegbunam◦ provisional | DEN | TE | 18 | 44.4% | 50.8% | 44.4% – 57.3% |
| 51 | John Bates◦ provisional | WAS | TE | 22 | 45.5% | 50.8% | 44.5% – 57.2% |
| 52 | Mo Alie-Cox◦ provisional | IND | TE | 28 | 46.4% | 50.8% | 44.6% – 57.0% |
| 53 | Irv Smith◦ provisional | MIN | TE | 36 | 47.2% | 50.8% | 44.7% – 56.9% |
| 54 | Jelani Woods◦ provisional | IND | TE | 40 | 47.5% | 50.7% | 44.7% – 56.8% |
| 55 | Gerald Everett◦ provisional | LAC | TE | 88 | 48.9% | 50.6% | 45.2% – 56.0% |
| 56 | Shane Zylstra◦ provisional | DET | TE | 15 | 40.0% | 50.6% | 44.1% – 57.0% |
| 57 | Tyler Higbee◦ provisional | LA | TE | 108 | 49.1% | 50.5% | 45.4% – 55.7% |
| 58 | Josh Oliver◦ provisional | BAL | TE | 25 | 44.0% | 50.5% | 44.2% – 56.8% |
| 59 | Hunter Henry◦ provisional | NE | TE | 59 | 47.5% | 50.4% | 44.7% – 56.2% |
| 60 | Cade Otton◦ provisional | TB | TE | 65 | 47.7% | 50.4% | 44.8% – 56.1% |
| 61 | Geoff Swaim◦ provisional | TEN | TE | 16 | 37.5% | 50.2% | 43.8% – 56.7% |
| 62 | Darren Waller◦ provisional | LV | TE | 44 | 45.5% | 50.2% | 44.3% – 56.2% |
| 63 | T.J. Hockenson◦ provisional | MIN | TE | 131 | 48.1% | 50.0% | 45.0% – 54.9% |
| 64 | Mike Gesicki◦ provisional | MIA | TE | 53 | 45.3% | 50.0% | 44.1% – 55.8% |
| 65 | Jonnu Smith◦ provisional | NE | TE | 38 | 42.1% | 49.7% | 43.6% – 55.7% |
| 66 | Eric Saubert◦ provisional | DEN | TE | 24 | 37.5% | 49.6% | 43.3% – 56.0% |
| 67 | Pharaoh Brown◦ provisional | CLE | TE | 20 | 35.0% | 49.6% | 43.3% – 56.0% |
| 68 | O.J. Howard◦ provisional | HOU | TE | 23 | 34.8% | 49.3% | 43.0% – 55.7% |
| 69 | Zach Gentry◦ provisional | PIT | TE | 23 | 34.8% | 49.3% | 43.0% – 55.7% |
| 70 | Tommy Tremble◦ provisional | CAR | TE | 32 | 37.5% | 49.1% | 42.9% – 55.3% |
| 71 | Brevin Jordan◦ provisional | HOU | TE | 28 | 35.7% | 49.1% | 42.9% – 55.3% |
| 72 | Logan Thomas◦ provisional | WAS | TE | 61 | 42.6% | 49.0% | 43.3% – 54.7% |
| 73 | Kyle Pitts◦ provisional | ATL | TE | 59 | 42.4% | 49.0% | 43.2% – 54.7% |
| 74 | Tre' McKitty◦ provisional | LAC | TE | 18 | 22.2% | 48.4% | 42.0% – 54.8% |
| 75 | Tyler Conklin◦ provisional | NYJ | TE | 87 | 42.5% | 48.2% | 42.9% – 53.6% |
| 76 | Cameron Brate◦ provisional | TB | TE | 38 | 34.2% | 48.0% | 42.0% – 54.1% |
| 77 | Isaiah Likely◦ provisional | BAL | TE | 60 | 35.0% | 46.8% | 41.1% – 52.5% |
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).