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 · 2016

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
146
carrys to trust the number
shrunkrawthe shrink90% CI
30%35%40%45%50%avg1Mike GillisleeBUF · 10146.1%2Devonta FreemanATL · 22743.1%3Le'Veon BellPIT · 26142.9%4Rex BurkheadCIN · 7442.6%5Robert TurbinIND · 4742.4%6Bilal PowellNYJ · 13142.1%7Ezekiel ElliottDAL · 32242.0%8LeSean McCoyBUF · 23541.9%9Ryan MathewsPHI · 15741.8%10Aaron RipkowskiGB · 3441.5%11John KuhnNO · 1841.3%12David JohnsonARI · 29341.2%13Kenneth DixonBAL · 8840.9%14Mark IngramNO · 20840.9%15Dion LewisNE · 6440.8%16Chris ThompsonWAS · 6940.8%17Darren SprolesPHI · 9440.7%18DeMarco MurrayTEN · 29340.7%19Andre WilliamsLAC · 1840.7%20Jalen RichardLV · 8340.5%21Alfred MorrisDAL · 6939.9%22Jay AjayiMIA · 26039.8%23DeAndre WashingtonLV · 8739.8%24Spencer WareKC · 21439.6%25Jacquizz RodgersTB · 12939.6%26Matt JonesWAS · 9939.5%27Kerwynn WilliamsARI · 1839.5%28Ameer AbdullahDET · 1839.5%29Jordan HowardCHI · 25239.4%30Derrick HenryTEN · 11039.4%31Danny WoodheadLAC · 1939.2%32Kenyan DrakeMIA · 3339.0%33Benny CunninghamLA · 2138.8%34Darren McFaddenDAL · 2438.7%35Bobby RaineyNYG · 1738.5%

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 · 2016 · full board

Rush success rate leaderboard for the 2016 NFL season, shrunk by empirical Bayes with 90% credible intervals, ranked by the shrunk estimate.
#PlayerTeamPoscarrysRawShrunk90% interval
1Mike Gillislee◦ provisionalBUFRB10157.4%46.1%40.9%51.3%
2Devonta FreemanATLRB22746.3%43.1%38.9%47.3%
3Le'Veon BellPITRB26145.6%42.9%38.9%47.0%
4Rex Burkhead◦ provisionalCINRB7451.3%42.6%37.2%48.1%
5Robert Turbin◦ provisionalINDRB4755.3%42.4%36.6%48.2%
6Bilal Powell◦ provisionalNYJRB13146.6%42.1%37.3%47.0%
7Ezekiel ElliottDALRB32243.8%42.0%38.3%45.8%
8LeSean McCoyBUFRB23544.3%41.9%37.8%46.1%
9Ryan MathewsPHIRB15745.2%41.8%37.2%46.5%
10Aaron Ripkowski◦ provisionalGBRB3455.9%41.5%35.5%47.6%
11John Kuhn◦ provisionalNORB1866.7%41.3%35.1%47.7%
12David JohnsonARIRB29342.7%41.2%37.3%45.1%
13Kenneth Dixon◦ provisionalBALRB8845.5%40.9%35.7%46.2%
14Mark IngramNORB20842.8%40.9%36.6%45.2%
15Dion Lewis◦ provisionalNERB6446.9%40.8%35.3%46.5%
16Chris Thompson◦ provisionalWASRB6946.4%40.8%35.4%46.4%
17Darren Sproles◦ provisionalPHIRB9444.7%40.7%35.6%46.0%
18DeMarco MurrayTENRB29342.0%40.7%36.9%44.6%
19Andre Williams◦ provisionalLACRB1861.1%40.7%34.5%47.1%
20Jalen Richard◦ provisionalLVRB8344.6%40.5%35.2%45.9%
21Alfred Morris◦ provisionalDALRB6943.5%39.9%34.4%45.4%
22Jay AjayiMIARB26040.8%39.8%35.9%43.9%
23DeAndre Washington◦ provisionalLVRB8742.5%39.8%34.6%45.1%
24Spencer WareKCRB21440.6%39.6%35.4%43.9%
25Jacquizz Rodgers◦ provisionalTBRB12941.1%39.6%34.8%44.4%
26Matt Jones◦ provisionalWASRB9941.4%39.5%34.4%44.7%
27Kerwynn Williams◦ provisionalARIRB1850.0%39.5%33.3%45.8%
28Ameer Abdullah◦ provisionalDETRB1850.0%39.5%33.3%45.8%
29Jordan HowardCHIRB25240.1%39.4%35.4%43.4%
30Derrick Henry◦ provisionalTENRB11040.9%39.4%34.4%44.4%
31Danny Woodhead◦ provisionalLACRB1947.4%39.2%33.1%45.6%
32Theo Riddick◦ provisionalDETRB9240.2%39.0%33.8%44.2%
33Kenyan Drake◦ provisionalMIARB3342.4%39.0%33.1%45.0%
34Latavius MurrayLVRB19539.5%38.9%34.6%43.3%
35Wendell Smallwood◦ provisionalPHIRB7740.3%38.9%33.6%44.3%
36Jeremy Langford◦ provisionalCHIRB6240.3%38.8%33.3%44.4%
37Fozzy Whittaker◦ provisionalCARRB5740.4%38.8%33.2%44.5%
38James White◦ provisionalNERB3941.0%38.8%33.0%44.7%
39Benny Cunningham◦ provisionalLARB2142.9%38.8%32.7%45.0%
40Darren McFadden◦ provisionalDALRB2441.7%38.7%32.6%44.9%
41Carlos HydeSFRB21839.0%38.7%34.5%42.9%
42Tim Hightower◦ provisionalNORB13339.1%38.6%33.9%43.5%
43Kenjon Barner◦ provisionalPHIRB2740.7%38.6%32.6%44.7%
44Jonathan Williams◦ provisionalBUFRB2740.7%38.6%32.6%44.7%
45Bobby Rainey◦ provisionalNYGRB1741.2%38.5%32.3%44.8%
46C.J. Prosise◦ provisionalSEARB3040.0%38.5%32.6%44.6%
47Shane Vereen◦ provisionalNYGRB3339.4%38.4%32.5%44.5%
48Akeem Hunt◦ provisionalHOURB2040.0%38.4%32.3%44.7%
49Zach Zenner◦ provisionalDETRB8838.6%38.4%33.2%43.6%
50Paul Perkins◦ provisionalNYGRB11238.4%38.3%33.4%43.3%
51Tevin Coleman◦ provisionalATLRB11838.1%38.2%33.3%43.1%
52Kapri Bibbs◦ provisionalDENRB2937.9%38.1%32.2%44.2%
53Eddie Lacy◦ provisionalGBRB7138.0%38.1%32.8%43.6%
54Matt Asiata◦ provisionalMINRB12138.0%38.1%33.3%43.0%
55Alfred Blue◦ provisionalHOURB10038.0%38.1%33.1%43.3%
56DeAngelo Williams◦ provisionalPITRB9837.8%38.0%33.0%43.2%
57Christine MichaelGBRB14837.8%38.0%33.4%42.7%
58Jamize Olawale◦ provisionalLVRB1735.3%37.9%31.7%44.2%
59Giovani Bernard◦ provisionalCINRB9137.4%37.9%32.8%43.1%
60Duke Johnson◦ provisionalCLERB7337.0%37.8%32.5%43.2%
61Andre Ellington◦ provisionalARIRB3435.3%37.6%31.8%43.6%
62Ka'Deem Carey◦ provisionalCHIRB3234.4%37.5%31.6%43.5%
63Corey Grant◦ provisionalJAXRB3234.4%37.5%31.6%43.5%
64Byron Marshall◦ provisionalPHIRB1931.6%37.4%31.3%43.7%
65C.J. Anderson◦ provisionalDENRB11036.4%37.4%32.5%42.4%
66Orleans Darkwa◦ provisionalNYGRB3033.3%37.4%31.4%43.4%
67Frank GoreINDRB26336.9%37.4%33.5%41.3%
68Chris Johnson◦ provisionalARIRB2532.0%37.3%31.3%43.4%
69Kenneth Farrow◦ provisionalLACRB6035.0%37.3%31.8%42.9%
70Alex Collins◦ provisionalSEARB3132.3%37.1%31.3%43.2%
71DuJuan Harris◦ provisionalSFRB3831.6%36.8%31.1%42.7%
72Mike Davis◦ provisionalSFRB1926.3%36.8%30.8%43.1%
73Denard Robinson◦ provisionalJAXRB4131.7%36.8%31.1%42.6%
74Cameron Artis-Payne◦ provisionalCARRB3630.6%36.7%30.9%42.6%
75Terron Ward◦ provisionalATLRB3129.0%36.6%30.7%42.6%
76Isaiah CrowellCLERB19835.4%36.5%32.3%40.9%
77Jonathan Grimes◦ provisionalHOURB2326.1%36.5%30.6%42.7%
78Melvin GordonLACRB25435.4%36.4%32.5%40.4%
79Malcolm Brown◦ provisionalLARB1822.2%36.4%30.4%42.7%
80Chris Ivory◦ provisionalJAXRB11734.2%36.4%31.6%41.3%
81Jeremy HillCINRB22235.1%36.3%32.3%40.5%
82Ronnie Hillman◦ provisionalLACRB4129.3%36.2%30.5%42.1%
83Rob KelleyWASRB16834.5%36.2%31.8%40.7%
84Peyton Barber◦ provisionalTBRB5530.9%36.2%30.7%41.8%
85Devontae BookerDENRB17434.5%36.2%31.8%40.6%
86Arian Foster◦ provisionalMIARB2222.7%36.2%30.2%42.3%
87Doug Martin◦ provisionalTBRB14434.0%36.1%31.5%40.8%
88Lamar MillerHOURB26834.7%35.9%32.1%39.8%
89Jonathan StewartCARRB21834.4%35.9%31.8%40.1%
90LeGarrette BlountNERB29934.8%35.9%32.2%39.7%
91Damien Williams◦ provisionalMIARB3525.7%35.8%30.0%41.7%
92Matt ForteNYJRB21833.5%35.4%31.3%39.5%
93Shaun Draughn◦ provisionalSFRB7429.7%35.3%30.1%40.7%
94Josh Ferguson◦ provisionalINDRB156.7%35.3%29.2%41.5%
95Knile Davis◦ provisionalKCRB1811.1%35.2%29.2%41.4%
96T.J. Yeldon◦ provisionalJAXRB13031.5%35.0%30.4%39.8%
97Charles Sims◦ provisionalTBRB5125.5%34.9%29.4%40.6%
98Todd GurleyLARB27833.1%34.8%31.1%38.7%
99Terrance WestBALRB19332.1%34.7%30.5%39.0%
100Mike Tolbert◦ provisionalCARRB3520.0%34.7%29.0%40.6%
101Charcandrick West◦ provisionalKCRB8828.4%34.5%29.5%39.7%
102Adrian Peterson◦ provisionalMINRB3718.9%34.3%28.6%40.1%
103James Starks◦ provisionalGBRB6425.0%34.2%28.9%39.6%
104Thomas Rawls◦ provisionalSEARB10928.4%34.0%29.2%39.0%
105Rashad JenningsNYGRB18030.6%34.0%29.7%38.3%
106Dwayne Washington◦ provisionalDETRB9026.7%33.8%28.8%38.9%
107Jerick McKinnonMINRB16128.6%33.1%28.8%37.6%
108Justin Forsett◦ provisionalDENRB8724.1%32.9%28.0%38.1%

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).