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.
Completion % · 2025
Completions per pass attempt. Depth-of-target confound: a checkdown offense completes more. Pair with CPOE.
- Beats raw by
- 15.6%
- lower out-of-sample error
- RMSE raw → shrunk
- 5.2% → 4.4%
- odd vs. even weeks
- Split-half reliability
- 0.42
- how repeatable the raw stat is
- Stabilizes at
- 320
- attempts 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.
Completion % · 2025 · full board
| # | Player | Team | Pos | attempts | Raw | Shrunk | 90% interval |
|---|---|---|---|---|---|---|---|
| 1 | Drake Maye | NE | QB | 540 | 65.6% | 63.3% | 60.6% – 66.0% |
| 2 | Brock Purdy◦ provisional | SF | QB | 298 | 66.1% | 62.7% | 59.5% – 65.8% |
| 3 | Mac Jones◦ provisional | SF | QB | 307 | 65.5% | 62.4% | 59.2% – 65.6% |
| 4 | Jared Goff | DET | QB | 616 | 63.8% | 62.3% | 59.7% – 64.9% |
| 5 | Dak Prescott | DAL | QB | 634 | 63.7% | 62.3% | 59.7% – 64.9% |
| 6 | Sam Darnold | SEA | QB | 506 | 63.8% | 62.1% | 59.4% – 64.9% |
| 7 | Daniel Jones | IND | QB | 411 | 63.5% | 61.7% | 58.8% – 64.7% |
| 8 | Josh Allen | BUF | QB | 507 | 62.9% | 61.6% | 58.8% – 64.3% |
| 9 | Jordan Love | GB | QB | 462 | 63.0% | 61.6% | 58.7% – 64.4% |
| 10 | Malik Willis◦ provisional | GB | QB | 38 | 79.0% | 61.5% | 57.3% – 65.7% |
| 11 | Matthew Stafford | LA | QB | 620 | 62.6% | 61.5% | 58.9% – 64.1% |
| 12 | Tyler Huntley◦ provisional | BAL | QB | 75 | 69.3% | 61.3% | 57.3% – 65.3% |
| 13 | Spencer Rattler◦ provisional | NO | QB | 275 | 63.3% | 61.2% | 57.9% – 64.5% |
| 14 | Tua Tagovailoa | MIA | QB | 417 | 62.4% | 61.1% | 58.1% – 64.0% |
| 15 | Mason Rudolph◦ provisional | PIT | QB | 54 | 70.4% | 61.1% | 56.9% – 65.1% |
| 16 | Aaron Rodgers | PIT | QB | 529 | 61.8% | 60.9% | 58.2% – 63.7% |
| 17 | Joe Burrow◦ provisional | CIN | QB | 278 | 62.2% | 60.8% | 57.5% – 64.0% |
| 18 | Tyler Shough | NO | QB | 359 | 61.6% | 60.6% | 57.5% – 63.6% |
| 19 | Kyler Murray◦ provisional | ARI | QB | 177 | 62.2% | 60.4% | 56.8% – 64.0% |
| 20 | Mitchell Trubisky◦ provisional | BUF | QB | 37 | 67.6% | 60.3% | 56.0% – 64.5% |
| 21 | Bo Nix | DEN | QB | 639 | 60.7% | 60.3% | 57.7% – 62.9% |
| 22 | C.J. Stroud | HOU | QB | 449 | 60.8% | 60.3% | 57.3% – 63.1% |
| 23 | Cooper Rush◦ provisional | BAL | QB | 53 | 64.1% | 60.1% | 55.9% – 64.3% |
| 24 | Jalen Hurts | PHI | QB | 488 | 60.3% | 59.9% | 57.1% – 62.8% |
| 25 | Justin Herbert | LAC | QB | 566 | 60.1% | 59.9% | 57.1% – 62.5% |
| 26 | Jake Browning◦ provisional | CIN | QB | 134 | 60.5% | 59.8% | 56.0% – 63.5% |
| 27 | Bryce Young | CAR | QB | 508 | 59.8% | 59.7% | 56.9% – 62.5% |
| 28 | Quinn Ewers◦ provisional | MIA | QB | 91 | 60.4% | 59.7% | 55.7% – 63.6% |
| 29 | Geno Smith | LV | QB | 505 | 59.8% | 59.7% | 56.9% – 62.5% |
| 30 | Josh Johnson◦ provisional | WAS | QB | 56 | 60.7% | 59.7% | 55.5% – 63.8% |
| 31 | Philip Rivers◦ provisional | IND | QB | 97 | 59.8% | 59.6% | 55.6% – 63.5% |
| 32 | Jacoby Brissett | ARI | QB | 531 | 59.3% | 59.4% | 56.6% – 62.1% |
| 33 | Brandon Allen◦ provisional | TEN | QB | 31 | 54.8% | 59.1% | 54.7% – 63.3% |
| 34 | Carson Wentz◦ provisional | MIN | QB | 189 | 58.2% | 59.0% | 55.4% – 62.6% |
| 35 | Andy Dalton◦ provisional | CAR | QB | 45 | 55.6% | 59.0% | 54.7% – 63.2% |
| 36 | Kirk Cousins◦ provisional | ATL | QB | 284 | 58.5% | 59.0% | 55.7% – 62.3% |
| 37 | Baker Mayfield | TB | QB | 585 | 58.6% | 58.9% | 56.2% – 61.6% |
| 38 | Patrick Mahomes | KC | QB | 539 | 58.4% | 58.8% | 56.0% – 61.6% |
| 39 | Chris Oladokun◦ provisional | KC | QB | 63 | 55.6% | 58.8% | 54.7% – 62.9% |
| 40 | Riley Leonard◦ provisional | IND | QB | 70 | 55.7% | 58.8% | 54.7% – 62.9% |
| 41 | Max Brosmer◦ provisional | MIN | QB | 85 | 55.3% | 58.6% | 54.5% – 62.6% |
| 42 | Tanner McKee◦ provisional | PHI | QB | 46 | 52.2% | 58.6% | 54.3% – 62.8% |
| 43 | Kenny Pickett◦ provisional | LV | QB | 53 | 52.8% | 58.5% | 54.3% – 62.7% |
| 44 | Jameis Winston◦ provisional | NYG | QB | 69 | 53.6% | 58.4% | 54.3% – 62.5% |
| 45 | Michael Penix Jr.◦ provisional | ATL | QB | 290 | 57.2% | 58.4% | 55.1% – 61.7% |
| 46 | Joe Flacco | CIN | QB | 436 | 57.6% | 58.4% | 55.4% – 61.3% |
| 47 | Jaxson Dart | NYG | QB | 376 | 57.5% | 58.4% | 55.3% – 61.4% |
| 48 | Marcus Mariota◦ provisional | WAS | QB | 246 | 56.5% | 58.2% | 54.8% – 61.6% |
| 49 | Lamar Jackson | BAL | QB | 340 | 56.5% | 57.9% | 54.8% – 61.1% |
| 50 | Tyrod Taylor◦ provisional | NYJ | QB | 148 | 54.0% | 57.8% | 54.0% – 61.5% |
| 51 | Russell Wilson◦ provisional | NYG | QB | 129 | 53.5% | 57.8% | 53.9% – 61.6% |
| 52 | Jayden Daniels◦ provisional | WAS | QB | 207 | 55.1% | 57.8% | 54.2% – 61.3% |
| 53 | Trevor Lawrence | JAX | QB | 603 | 56.5% | 57.6% | 54.9% – 60.2% |
| 54 | Davis Mills◦ provisional | HOU | QB | 169 | 53.8% | 57.5% | 53.8% – 61.2% |
| 55 | Justin Fields◦ provisional | NYJ | QB | 235 | 54.5% | 57.4% | 53.9% – 60.8% |
| 56 | Dillon Gabriel◦ provisional | CLE | QB | 205 | 53.7% | 57.2% | 53.6% – 60.7% |
| 57 | Caleb Williams | CHI | QB | 596 | 55.4% | 56.8% | 54.1% – 59.5% |
| 58 | Trey Lance◦ provisional | LAC | QB | 63 | 42.9% | 56.7% | 52.6% – 60.9% |
| 59 | Brady Cook◦ provisional | NYJ | QB | 173 | 50.9% | 56.5% | 52.8% – 60.1% |
| 60 | Cam Ward | TEN | QB | 595 | 54.3% | 56.1% | 53.4% – 58.8% |
| 61 | Shedeur Sanders◦ provisional | CLE | QB | 235 | 51.1% | 55.9% | 52.4% – 59.4% |
| 62 | J.J. McCarthy◦ provisional | MIN | QB | 273 | 51.3% | 55.7% | 52.3% – 59.1% |
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).