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 · 2021

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
169
carrys to trust the number
shrunkrawthe shrink90% CI
30%35%40%45%avg1AJ DillonGB · 18744.2%2Leonard FournetteTB · 18044.0%3Jordan HowardPHI · 8643.7%4Latavius MurrayBAL · 11943.5%5Jonathan TaylorIND · 33243.4%6Damien HarrisNE · 20243.2%7Antonio GibsonWAS · 25943.1%8Darrel WilliamsKC · 14442.9%9Rhamondre Stevens…NE · 13342.8%10Miles SandersPHI · 13842.8%11Kenneth GainwellPHI · 6942.6%12Darrell HendersonLA · 14942.6%13Austin EkelerLAC · 20842.5%14Tevin ColemanNYJ · 8442.4%15Ronald JonesTB · 10142.4%16Melvin GordonDEN · 20342.3%17Tony PollardDAL · 13042.3%18DeeJay DallasSEA · 3342.3%19D'Ernest JohnsonCLE · 10042.1%20Jonathan WilliamsWAS · 1742.1%21Aaron JonesGB · 17141.9%22James RobinsonJAX · 16441.9%23Derrick GoreKC · 5141.5%24Jermar JeffersonDET · 1541.5%25Chris CarsonSEA · 5441.4%26Ezekiel ElliottDAL · 24041.4%27James ConnerARI · 20241.3%28Patrick TaylorGB · 2341.3%29Justin JacksonLAC · 6941.3%30Devonta FreemanBAL · 13341.2%31Ryquell ArmsteadJAX · 1541.0%32Damien WilliamsCHI · 4040.9%33Brandon BoldenNE · 4540.8%34Travis HomerSEA · 2140.7%35Matt BreidaBUF · 2640.7%36Trayveon WilliamsCIN · 1540.4%

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 · 2021 · full board

Rush success rate leaderboard for the 2021 NFL season, shrunk by empirical Bayes with 90% credible intervals, ranked by the shrunk estimate.
#PlayerTeamPoscarrysRawShrunk90% interval
1AJ DillonGBRB18748.1%44.2%39.9%48.5%
2Leonard FournetteTBRB18047.8%44.0%39.6%48.3%
3Jordan Howard◦ provisionalPHIRB8651.2%43.7%38.6%48.8%
4Latavius Murray◦ provisionalBALRB11948.7%43.5%38.8%48.4%
5Jonathan TaylorINDRB33245.2%43.4%39.8%47.0%
6Damien HarrisNERB20246.0%43.2%39.0%47.5%
7Antonio GibsonWASRB25945.2%43.1%39.2%47.0%
8Darrel Williams◦ provisionalKCRB14446.5%42.9%38.4%47.5%
9Rhamondre Stevenson◦ provisionalNERB13346.6%42.8%38.2%47.5%
10Miles Sanders◦ provisionalPHIRB13846.4%42.8%38.2%47.5%
11Kenneth Gainwell◦ provisionalPHIRB6949.3%42.6%37.4%47.9%
12Darrell Henderson◦ provisionalLARB14945.6%42.6%38.0%47.1%
13Austin EkelerLACRB20844.7%42.5%38.4%46.8%
14Tevin Coleman◦ provisionalNYJRB8447.6%42.4%37.4%47.6%
15Ronald Jones◦ provisionalTBRB10146.5%42.4%37.5%47.3%
16Melvin GordonDENRB20344.3%42.3%38.1%46.5%
17Tony Pollard◦ provisionalDALRB13045.4%42.3%37.6%47.0%
18DeeJay Dallas◦ provisionalSEARB3354.5%42.3%36.6%48.0%
19D'Ernest Johnson◦ provisionalCLERB10046.0%42.1%37.2%47.1%
20Jonathan Williams◦ provisionalWASRB1764.7%42.1%36.3%48.1%
21Aaron JonesGBRB17143.9%41.9%37.5%46.3%
22James Robinson◦ provisionalJAXRB16443.9%41.9%37.5%46.3%
23Derrick Gore◦ provisionalKCRB5147.1%41.5%36.1%47.0%
24Jermar Jefferson◦ provisionalDETRB1560.0%41.5%35.6%47.5%
25Chris Carson◦ provisionalSEARB5446.3%41.4%36.1%46.9%
26Ezekiel ElliottDALRB24042.5%41.4%37.4%45.4%
27James ConnerARIRB20242.6%41.3%37.2%45.6%
28Patrick Taylor◦ provisionalGBRB2352.2%41.3%35.6%47.2%
29Justin Jackson◦ provisionalLACRB6944.9%41.3%36.1%46.6%
30Devonta Freeman◦ provisionalBALRB13342.9%41.2%36.6%45.9%
31Christian McCaffrey◦ provisionalCARRB9943.4%41.2%36.3%46.2%
32Clyde Edwards-Helaire◦ provisionalKCRB11942.9%41.1%36.4%45.9%
33Josh JacobsLVRB21741.9%41.0%37.0%45.2%
34Ryquell Armstead◦ provisionalJAXRB1553.3%41.0%35.1%47.0%
35Kareem Hunt◦ provisionalCLERB7943.0%40.9%35.8%46.1%
36Damien Williams◦ provisionalCHIRB4045.0%40.9%35.3%46.5%
37Brandon Bolden◦ provisionalNERB4544.4%40.8%35.4%46.4%
38Jamaal Williams◦ provisionalDETRB15341.8%40.8%36.3%45.3%
39Travis Homer◦ provisionalSEARB2147.6%40.7%34.9%46.6%
40Matt Breida◦ provisionalBUFRB2646.2%40.7%35.0%46.5%
41Trey Sermon◦ provisionalSFRB4143.9%40.7%35.2%46.3%
42Elijah MitchellSFRB20741.1%40.5%36.4%44.7%
43Chase Edmonds◦ provisionalARIRB11641.4%40.5%35.8%45.3%
44Trayveon Williams◦ provisionalCINRB1546.7%40.4%34.6%46.4%
45Boston Scott◦ provisionalPHIRB8741.4%40.4%35.4%45.5%
46Peyton Barber◦ provisionalLVRB5541.8%40.4%35.0%45.8%
47Godwin Igwebuike◦ provisionalDETRB1844.4%40.3%34.5%46.3%
48JaMycal Hasty◦ provisionalSFRB1643.8%40.2%34.4%46.2%
49Dontrell Hilliard◦ provisionalTENRB5641.1%40.2%34.9%45.6%
50J.D. McKissic◦ provisionalWASRB4940.8%40.1%34.7%45.6%
51Dalvin CookMINRB25140.2%40.1%36.2%44.0%
52Joe MixonCINRB29440.1%40.0%36.3%43.8%
53Elijhaa Penny◦ provisionalNYGRB2540.0%39.9%34.2%45.7%
54Nick ChubbCLERB22839.9%39.9%35.9%44.0%
55Khalil Herbert◦ provisionalCHIRB10339.8%39.9%35.0%44.8%
56Zack Moss◦ provisionalBUFRB9639.6%39.8%34.9%44.8%
57Duke Johnson◦ provisionalMIARB7139.4%39.8%34.6%45.0%
58Qadree Ollison◦ provisionalATLRB2138.1%39.7%33.9%45.6%
59Rex Burkhead◦ provisionalHOURB12239.3%39.7%35.0%44.4%
60Sony MichelLARB20839.4%39.6%35.5%43.8%
61Devontae Booker◦ provisionalNYGRB14539.3%39.6%35.1%44.2%
62Carlos Hyde◦ provisionalJAXRB7238.9%39.6%34.5%44.8%
63Craig Reynolds◦ provisionalDETRB5538.2%39.5%34.2%44.9%
64Ty'Son Williams◦ provisionalBALRB3537.1%39.4%33.9%45.1%
65Jaret Patterson◦ provisionalWASRB6838.2%39.4%34.3%44.7%
66Javonte WilliamsDENRB20338.9%39.4%35.2%43.5%
67Nyheim Hines◦ provisionalINDRB5637.5%39.3%34.0%44.7%
68Michael Carter◦ provisionalNYJRB14838.5%39.2%34.8%43.8%
69Austin Walter◦ provisionalNYJRB2634.6%39.2%33.5%45.0%
70Eno Benjamin◦ provisionalARIRB3435.3%39.1%33.6%44.8%
71Kenyan Drake◦ provisionalLVRB6336.5%39.0%33.8%44.3%
72Jeff Wilson◦ provisionalSFRB7936.7%38.9%33.8%44.0%
73Adrian Peterson◦ provisionalSEARB3834.2%38.8%33.4%44.5%
74Royce Freeman◦ provisionalHOURB5635.7%38.8%33.6%44.2%
75Alex Collins◦ provisionalSEARB10837.0%38.8%34.0%43.6%
76Rashaad Penny◦ provisionalSEARB11937.0%38.7%34.0%43.4%
77Samaje Perine◦ provisionalCINRB5534.5%38.6%33.3%44.0%
78Le'Veon Bell◦ provisionalTBRB4032.5%38.5%33.0%44.0%
79Ty Johnson◦ provisionalNYJRB6134.4%38.4%33.2%43.8%
80Chris Evans◦ provisionalCINRB1723.5%38.4%32.6%44.3%
81Corey Clement◦ provisionalDALRB3330.3%38.3%32.8%44.0%
82Joshua Kelley◦ provisionalLACRB3330.3%38.3%32.8%44.0%
83Wayne Gallman◦ provisionalMINRB2828.6%38.3%32.7%44.0%
84Marlon Mack◦ provisionalINDRB2828.6%38.3%32.7%44.0%
85Benny Snell◦ provisionalPITRB3630.6%38.3%32.8%43.9%
86Ke'Shawn Vaughn◦ provisionalTBRB3630.6%38.3%32.8%43.9%
87Chuba HubbardCARRB17236.6%38.2%34.0%42.6%
88Saquon Barkley◦ provisionalNYGRB16236.4%38.2%33.9%42.6%
89Derrick HenryTENRB22036.8%38.1%34.1%42.2%
90Alvin KamaraNORB24136.9%38.1%34.2%42.1%
91Najee HarrisPITRB30737.1%38.1%34.5%41.8%
92Mike Davis◦ provisionalATLRB13835.5%37.9%33.4%42.5%
93David MontgomeryCHIRB22536.4%37.9%33.9%42.0%
94D'Onta Foreman◦ provisionalTENRB13335.3%37.9%33.4%42.5%
95Malcolm Brown◦ provisionalMIARB3327.3%37.8%32.3%43.5%
96Jeremy McNichols◦ provisionalTENRB4129.3%37.8%32.4%43.4%
97Devin SingletaryBUFRB18835.6%37.6%33.5%41.9%
98J.J. Taylor◦ provisionalNERB1915.8%37.5%31.7%43.3%
99Dare Ogunbowale◦ provisionalJAXRB4427.3%37.3%31.9%42.8%
100Larry Rountree III◦ provisionalLACRB3625.0%37.3%31.8%42.9%
101Ameer Abdullah◦ provisionalCARRB5127.5%37.0%31.7%42.4%
102Tony Jones◦ provisionalNORB5427.8%37.0%31.7%42.3%
103Salvon Ahmed◦ provisionalMIARB5425.9%36.5%31.3%41.9%
104Alexander Mattison◦ provisionalMINRB13531.9%36.3%31.8%40.9%
105Myles GaskinMIARB17330.6%35.2%31.0%39.5%
106Cordarrelle Patterson◦ provisionalATLRB15329.4%34.9%30.6%39.3%
107Phillip Lindsay◦ provisionalMIARB8825.0%34.8%30.0%39.7%
108Mark Ingram◦ provisionalNORB16028.7%34.5%30.2%38.8%
109David Johnson◦ provisionalHOURB6719.4%34.1%29.1%39.2%
110D'Andre Swift◦ provisionalDETRB15127.2%33.9%29.6%38.3%

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).