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.

Rush success rate · 2025

Share of carries with positive EPA. Blocking & scheme are large, un-controlled inputs.

Beats raw by
19.4%
lower out-of-sample error
RMSE raw → shrunk
9.1% → 7.3%
odd vs. even weeks
Split-half reliability
0.20
how repeatable the raw stat is
Stabilizes at
272
carrys to trust the number
shrunkrawthe shrink90% CI
35%40%45%avg1Kyren WilliamsLA · 25945.4%2Blake CorumLA · 14544.4%3Samaje PerineCIN · 8444.2%4D'Andre SwiftCHI · 22344.1%5Kareem HuntKC · 16343.5%6James CookBUF · 30943.4%7Rachaad WhiteTB · 13343.3%8Javonte WilliamsDAL · 25243.2%9Jordan MasonMIN · 15943.0%10Kyle MonangaiCHI · 17142.9%11Jonathan TaylorIND · 32442.8%12J.K. DobbinsDEN · 15342.6%13Chris Rodriguez J…WAS · 11242.5%14Ty JohnsonBUF · 5042.3%15Kenneth GainwellPIT · 11442.3%16Bhayshul TutenJAX · 8442.2%17Jaylen WrightMIA · 7042.2%18Tank BigsbyPHI · 6342.1%19Josh JacobsGB · 23442.1%20Jaleel McLaughlinDEN · 3742.1%21Emanuel WilsonGB · 12541.9%22David MontgomeryDET · 15941.8%23Hunter LuepkeDAL · 1641.7%24Jaylen WarrenPIT · 21141.7%25Aaron JonesMIN · 13241.6%26Jacory Croskey-Me…WAS · 17641.6%27Brian RobinsonSF · 9241.5%28Rico DowdleCAR · 23741.5%29Derrick HenryBAL · 30741.3%30Michael CarterARI · 9241.3%31Audric EstiméNO · 4641.3%32British BrooksHOU · 1741.3%33Najee HarrisLAC · 1541.2%34Miles SandersDAL · 2041.2%35Antonio GibsonNE · 2541.1%

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.

Rush success rate · 2025 · full board

Rush success rate leaderboard for the 2025 NFL season, shrunk by empirical Bayes with 90% credible intervals, ranked by the shrunk estimate.
#PlayerTeamPoscarrysRawShrunk90% interval
1Kyren Williams◦ provisionalLARB25950.2%45.4%41.9%49.0%
2Blake Corum◦ provisionalLARB14551.0%44.4%40.4%48.4%
3Samaje Perine◦ provisionalCINRB8454.8%44.2%39.9%48.5%
4D'Andre Swift◦ provisionalCHIRB22348.0%44.1%40.4%47.8%
5Kareem Hunt◦ provisionalKCRB16347.9%43.5%39.6%47.4%
6James CookBUFRB30945.6%43.4%40.0%46.8%
7Rachaad White◦ provisionalTBRB13348.1%43.3%39.2%47.3%
8Javonte Williams◦ provisionalDALRB25245.6%43.2%39.6%46.7%
9Jordan Mason◦ provisionalMINRB15946.5%43.0%39.1%46.9%
10Kyle Monangai◦ provisionalCHIRB17146.2%42.9%39.1%46.8%
11Jonathan TaylorINDRB32444.4%42.8%39.5%46.2%
12J.K. Dobbins◦ provisionalDENRB15345.8%42.6%38.7%46.6%
13Chris Rodriguez Jr.◦ provisionalWASRB11246.4%42.5%38.4%46.7%
14Ty Johnson◦ provisionalBUFRB5050.0%42.3%37.8%46.9%
15Kenneth Gainwell◦ provisionalPITRB11445.6%42.3%38.2%46.4%
16Bhayshul Tuten◦ provisionalJAXRB8446.4%42.2%37.9%46.5%
17Jaylen Wright◦ provisionalMIARB7047.1%42.2%37.8%46.6%
18Tank Bigsby◦ provisionalPHIRB6347.6%42.1%37.8%46.6%
19Josh Jacobs◦ provisionalGBRB23443.6%42.1%38.6%45.8%
20Jaleel McLaughlin◦ provisionalDENRB3751.3%42.1%37.5%46.8%
21Emanuel Wilson◦ provisionalGBRB12544.0%41.9%37.8%46.0%
22David Montgomery◦ provisionalDETRB15943.4%41.8%37.9%45.7%
23Hunter Luepke◦ provisionalDALRB1656.3%41.7%37.0%46.6%
24Jaylen Warren◦ provisionalPITRB21142.6%41.7%38.0%45.4%
25Aaron Jones◦ provisionalMINRB13243.2%41.6%37.6%45.7%
26Jacory Croskey-Merritt◦ provisionalWASRB17642.6%41.6%37.8%45.4%
27Brian Robinson◦ provisionalSFRB9243.5%41.5%37.3%45.8%
28Rico Dowdle◦ provisionalCARRB23742.2%41.5%37.9%45.1%
29Derrick HenryBALRB30741.7%41.3%38.0%44.7%
30Michael Carter◦ provisionalARIRB9242.4%41.3%37.0%45.5%
31Audric Estimé◦ provisionalNORB4643.5%41.3%36.8%45.8%
32British Brooks◦ provisionalHOURB1747.1%41.3%36.5%46.0%
33Omarion Hampton◦ provisionalLACRB12441.9%41.2%37.2%45.3%
34TreVeyon Henderson◦ provisionalNERB18041.7%41.2%37.4%45.0%
35Najee Harris◦ provisionalLACRB1546.7%41.2%36.4%46.0%
36Miles Sanders◦ provisionalDALRB2045.0%41.2%36.5%45.9%
37Antonio Gibson◦ provisionalNERB2544.0%41.1%36.5%45.9%
38Bijan RobinsonATLRB28841.3%41.1%37.7%44.5%
39Jeremy McNichols◦ provisionalWASRB4542.2%41.1%36.6%45.6%
40Zach Charbonnet◦ provisionalSEARB18441.3%41.1%37.3%44.9%
41DJ Giddens◦ provisionalINDRB2642.3%41.0%36.4%45.7%
42Ray Davis◦ provisionalBUFRB5841.4%41.0%36.6%45.5%
43Cam Skattebo◦ provisionalNYGRB10241.2%41.0%36.8%45.2%
44Raheim Sanders◦ provisionalCLERB2740.7%40.9%36.2%45.6%
45Zavier Scott◦ provisionalMINRB3240.6%40.9%36.3%45.5%
46Chuba Hubbard◦ provisionalCARRB13540.7%40.8%36.9%44.9%
47Breece Hall◦ provisionalNYJRB24340.7%40.8%37.3%44.4%
48Kendre Miller◦ provisionalNORB4740.4%40.8%36.3%45.4%
49Malik Davis◦ provisionalDALRB5240.4%40.8%36.4%45.3%
50Ollie Gordon II◦ provisionalMIARB7040.0%40.7%36.4%45.1%
51Jaret Patterson◦ provisionalLACRB4139.0%40.6%36.1%45.2%
52Chris Brooks◦ provisionalGBRB2737.0%40.5%35.9%45.2%
53Tyler Allgeier◦ provisionalATLRB14339.9%40.5%36.6%44.5%
54Trevor Etienne◦ provisionalCARRB2035.0%40.5%35.8%45.2%
55Chase Brown◦ provisionalCINRB23339.9%40.4%36.9%44.0%
56Jahmyr Gibbs◦ provisionalDETRB24339.9%40.4%36.9%44.0%
57Terrell Jennings◦ provisionalNERB2334.8%40.4%35.8%45.1%
58Khalil Herbert◦ provisionalNYJRB1631.3%40.4%35.6%45.1%
59Braelon Allen◦ provisionalNYJRB1931.6%40.3%35.6%45.0%
60Jerome Ford◦ provisionalCLERB2433.3%40.3%35.6%45.0%
61Raheem Mostert◦ provisionalLVRB2231.8%40.2%35.5%45.0%
62Devin Singletary◦ provisionalNYGRB11938.7%40.2%36.2%44.3%
63James Conner◦ provisionalARIRB3234.4%40.2%35.6%44.9%
64Tony Pollard◦ provisionalTENRB24239.3%40.1%36.6%43.7%
65Justice Hill◦ provisionalBALRB1827.8%40.1%35.4%44.8%
66Isaiah Davis◦ provisionalNYJRB4334.9%40.1%35.6%44.6%
67Bam Knight◦ provisionalARIRB8337.4%40.1%35.8%44.4%
68Christian McCaffreySFRB31139.2%40.0%36.7%43.4%
69Sean Tucker◦ provisionalTBRB8637.2%40.0%35.8%44.3%
70Tahj Brooks◦ provisionalCINRB1625.0%40.0%35.3%44.8%
71Keaton Mitchell◦ provisionalBALRB5935.6%39.9%35.5%44.4%
72Devin Neal◦ provisionalNORB5735.1%39.9%35.5%44.4%
73Rhamondre Stevenson◦ provisionalNERB13137.4%39.8%35.8%43.8%
74LeQuint Allen Jr.◦ provisionalJAXRB2326.1%39.7%35.1%44.5%
75De'Von Achane◦ provisionalMIARB23838.2%39.6%36.1%43.2%
76Emari Demercado◦ provisionalARIRB4431.8%39.6%35.1%44.2%
77Kenneth Walker III◦ provisionalSEARB22138.0%39.6%36.0%43.2%
78Tyjae Spears◦ provisionalTENRB7234.7%39.6%35.3%44.0%
79Trey Benson◦ provisionalARIRB2927.6%39.6%35.0%44.3%
80George Holani◦ provisionalSEARB2222.7%39.5%34.9%44.3%
81Ty Chandler◦ provisionalMINRB1717.6%39.5%34.8%44.3%
82Hassan Haskins◦ provisionalLACRB1717.6%39.5%34.8%44.3%
83Jaydon Blue◦ provisionalDALRB3828.9%39.4%34.9%44.0%
84Kaleb Johnson◦ provisionalPITRB2825.0%39.4%34.8%44.1%
85Tyrone Tracy Jr.◦ provisionalNYGRB17636.9%39.3%35.6%43.1%
86Brashard Smith◦ provisionalKCRB4429.5%39.3%34.8%43.9%
87Saquon BarkleyPHIRB28137.7%39.3%35.9%42.7%
88Evan Hull◦ provisionalNORB1915.8%39.3%34.6%44.0%
89RJ Harvey◦ provisionalDENRB14736.0%39.2%35.3%43.1%
90Kimani Vidal◦ provisionalLACRB15536.1%39.2%35.3%43.1%
91Isiah Pacheco◦ provisionalKCRB11834.8%39.0%35.0%43.1%
92Woody Marks◦ provisionalHOURB19636.2%38.9%35.3%42.7%
93Nick Chubb◦ provisionalHOURB12234.4%38.9%34.9%43.0%
94Jawhar Jordan◦ provisionalHOURB4325.6%38.8%34.3%43.4%
95Bucky Irving◦ provisionalTBRB17334.7%38.5%34.7%42.3%
96Travis Etienne◦ provisionalJAXRB26035.8%38.4%34.9%41.9%
97Dylan Sampson◦ provisionalCLERB6526.2%38.0%33.7%42.4%
98Alvin Kamara◦ provisionalNORB13132.1%38.0%34.1%42.0%
99Quinshon Judkins◦ provisionalCLERB23133.3%37.4%33.9%41.0%
100Ashton Jeanty◦ provisionalLVRB26731.8%36.4%33.0%39.8%

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).