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 % · 2023
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
- 255
- 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 % · 2023 · full board
| # | Player | Team | Pos | attempts | Raw | Shrunk | 90% interval |
|---|---|---|---|---|---|---|---|
| 1 | Tua Tagovailoa | MIA | QB | 591 | 65.6% | 63.8% | 61.0% – 66.5% |
| 2 | Dak Prescott | DAL | QB | 631 | 65.0% | 63.4% | 60.7% – 66.0% |
| 3 | Brock Purdy | SF | QB | 472 | 65.3% | 63.2% | 60.2% – 66.1% |
| 4 | Kirk Cousins | MIN | QB | 328 | 65.8% | 63.0% | 59.7% – 66.3% |
| 5 | Derek Carr | NO | QB | 581 | 64.5% | 63.0% | 60.2% – 65.7% |
| 6 | Patrick Mahomes | KC | QB | 624 | 64.3% | 62.8% | 60.2% – 65.5% |
| 7 | Jared Goff | DET | QB | 640 | 63.6% | 62.4% | 59.7% – 65.0% |
| 8 | Josh Allen | BUF | QB | 607 | 63.4% | 62.2% | 59.5% – 64.9% |
| 9 | Jake Browning | CIN | QB | 268 | 63.8% | 61.7% | 58.1% – 65.1% |
| 10 | Mason Rudolph◦ provisional | PIT | QB | 80 | 68.8% | 61.6% | 57.2% – 66.0% |
| 11 | Joe Burrow | CIN | QB | 389 | 62.7% | 61.4% | 58.2% – 64.5% |
| 12 | Lamar Jackson | BAL | QB | 495 | 62.0% | 61.1% | 58.2% – 64.0% |
| 13 | Trevor Lawrence | JAX | QB | 602 | 61.5% | 60.9% | 58.1% – 63.6% |
| 14 | C.J. Beathard◦ provisional | JAX | QB | 60 | 66.7% | 60.8% | 56.2% – 65.3% |
| 15 | Tyson Bagent◦ provisional | CHI | QB | 149 | 63.1% | 60.8% | 56.7% – 64.7% |
| 16 | Jalen Hurts | PHI | QB | 575 | 61.2% | 60.7% | 57.9% – 63.4% |
| 17 | Nick Mullens◦ provisional | MIN | QB | 160 | 62.5% | 60.6% | 56.6% – 64.5% |
| 18 | Cooper Rush◦ provisional | DAL | QB | 25 | 72.0% | 60.5% | 55.7% – 65.3% |
| 19 | Justin Herbert | LAC | QB | 486 | 61.1% | 60.5% | 57.6% – 63.5% |
| 20 | Jordan Love | GB | QB | 611 | 60.9% | 60.5% | 57.7% – 63.2% |
| 21 | Geno Smith | SEA | QB | 532 | 60.7% | 60.3% | 57.4% – 63.1% |
| 22 | Mac Jones | NE | QB | 368 | 60.9% | 60.3% | 57.0% – 63.5% |
| 23 | Kyler Murray | ARI | QB | 291 | 60.5% | 60.0% | 56.5% – 63.4% |
| 24 | Russell Wilson | DEN | QB | 495 | 60.0% | 59.8% | 56.8% – 62.7% |
| 25 | Carson Wentz◦ provisional | LA | QB | 27 | 63.0% | 59.7% | 54.9% – 64.5% |
| 26 | Baker Mayfield | TB | QB | 610 | 59.7% | 59.6% | 56.8% – 62.3% |
| 27 | Jimmy Garoppolo◦ provisional | LV | QB | 184 | 59.8% | 59.6% | 55.7% – 63.4% |
| 28 | Easton Stick◦ provisional | LAC | QB | 188 | 59.0% | 59.3% | 55.4% – 63.1% |
| 29 | Marcus Mariota◦ provisional | PHI | QB | 26 | 57.7% | 59.2% | 54.4% – 64.0% |
| 30 | Desmond Ridder | ATL | QB | 422 | 59.0% | 59.2% | 56.0% – 62.2% |
| 31 | C.J. Stroud | HOU | QB | 541 | 59.0% | 59.1% | 56.2% – 62.0% |
| 32 | Case Keenum◦ provisional | HOU | QB | 59 | 57.6% | 59.1% | 54.5% – 63.6% |
| 33 | Tyrod Taylor◦ provisional | NYG | QB | 198 | 58.6% | 59.0% | 55.2% – 62.8% |
| 34 | Mitchell Trubisky◦ provisional | PIT | QB | 115 | 58.3% | 59.0% | 54.8% – 63.2% |
| 35 | Matthew Stafford | LA | QB | 554 | 58.8% | 59.0% | 56.2% – 61.9% |
| 36 | Drew Lock◦ provisional | SEA | QB | 83 | 57.8% | 59.0% | 54.6% – 63.4% |
| 37 | Andy Dalton◦ provisional | CAR | QB | 61 | 55.7% | 58.7% | 54.1% – 63.2% |
| 38 | Joe Flacco◦ provisional | CLE | QB | 213 | 57.8% | 58.7% | 54.9% – 62.4% |
| 39 | Joshua Dobbs | MIN | QB | 450 | 58.2% | 58.7% | 55.6% – 61.7% |
| 40 | Brian Hoyer◦ provisional | LV | QB | 43 | 53.5% | 58.5% | 53.8% – 63.2% |
| 41 | Aidan O'Connell | LV | QB | 368 | 57.9% | 58.5% | 55.2% – 61.7% |
| 42 | Tim Boyle◦ provisional | NYJ | QB | 86 | 55.8% | 58.5% | 54.1% – 62.8% |
| 43 | Sam Darnold◦ provisional | SF | QB | 52 | 53.8% | 58.5% | 53.8% – 63.0% |
| 44 | Kenny Pickett | PIT | QB | 348 | 57.8% | 58.5% | 55.1% – 61.7% |
| 45 | Jarrett Stidham◦ provisional | DEN | QB | 73 | 54.8% | 58.4% | 53.9% – 62.8% |
| 46 | Gardner Minshew | IND | QB | 527 | 57.9% | 58.4% | 55.5% – 61.3% |
| 47 | Tyler Huntley◦ provisional | BAL | QB | 41 | 51.2% | 58.3% | 53.5% – 62.9% |
| 48 | Blaine Gabbert◦ provisional | KC | QB | 36 | 50.0% | 58.2% | 53.5% – 62.9% |
| 49 | Daniel Jones◦ provisional | NYG | QB | 191 | 56.5% | 58.2% | 54.3% – 62.0% |
| 50 | Ryan Tannehill | TEN | QB | 262 | 56.9% | 58.1% | 54.5% – 61.7% |
| 51 | Jameis Winston◦ provisional | NO | QB | 49 | 51.0% | 58.1% | 53.4% – 62.7% |
| 52 | Jeff Driskel◦ provisional | CLE | QB | 29 | 44.8% | 57.9% | 53.1% – 62.7% |
| 53 | Deshaun Watson◦ provisional | CLE | QB | 188 | 55.9% | 57.9% | 54.0% – 61.7% |
| 54 | Anthony Richardson◦ provisional | IND | QB | 93 | 53.8% | 57.9% | 53.5% – 62.2% |
| 55 | Sam Howell | WAS | QB | 679 | 57.1% | 57.8% | 55.1% – 60.4% |
| 56 | Clayton Tune◦ provisional | ARI | QB | 28 | 42.9% | 57.8% | 52.9% – 62.5% |
| 57 | Brett Rypien◦ provisional | LA | QB | 40 | 45.0% | 57.5% | 52.7% – 62.1% |
| 58 | Trevor Siemian◦ provisional | NYJ | QB | 161 | 53.4% | 57.1% | 53.1% – 61.1% |
| 59 | Davis Mills◦ provisional | HOU | QB | 42 | 42.9% | 57.1% | 52.3% – 61.8% |
| 60 | Bailey Zappe◦ provisional | NE | QB | 237 | 53.6% | 56.6% | 52.9% – 60.3% |
| 61 | Dorian Thompson-Robinson◦ provisional | CLE | QB | 119 | 50.4% | 56.5% | 52.3% – 60.7% |
| 62 | Taylor Heinicke◦ provisional | ATL | QB | 144 | 51.4% | 56.5% | 52.4% – 60.6% |
| 63 | Tommy DeVito◦ provisional | NYG | QB | 215 | 53.0% | 56.5% | 52.7% – 60.2% |
| 64 | Justin Fields | CHI | QB | 416 | 54.6% | 56.4% | 53.3% – 59.5% |
| 65 | Will Levis | TEN | QB | 284 | 52.5% | 55.8% | 52.2% – 59.3% |
| 66 | Zach Wilson | NYJ | QB | 416 | 53.1% | 55.5% | 52.3% – 58.7% |
| 67 | Bryce Young | CAR | QB | 590 | 53.4% | 55.2% | 52.4% – 58.0% |
| 68 | PJ Walker◦ provisional | CLE | QB | 122 | 44.3% | 54.5% | 50.3% – 58.7% |
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).