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.

Target success rate · 2025

Share of targets with positive EPA, shrunk within position.

Beats raw by
21.9%
lower out-of-sample error
RMSE raw → shrunk
11.7% → 9.1%
odd vs. even weeks
Split-half reliability
0.19
how repeatable the raw stat is
Stabilizes at
91
targets to trust the number
shrunkrawthe shrink90% CI
45%50%55%60%65%TE avg1George KittleSF · 6960.5%2Sam LaPortaDET · 4959.1%3AJ BarnerSEA · 6859.0%4Hunter HenryNE · 8758.9%5Austin HooperNE · 2658.8%6Daniel BellingerNYG · 2658.8%7Oronde Gadsden IILAC · 6958.6%8Colby ParkinsonLA · 5658.3%9Trey McBrideARI · 17058.1%10Brenton StrangeJAX · 6058.1%11Dawson KnoxBUF · 5058.0%12Jackson HawesBUF · 1958.0%13Tucker KraftGB · 4457.6%14Mitchell EvansCAR · 2557.6%15Ian ThomasLV · 1357.5%16Dalton KincaidBUF · 5057.3%17Travis KelceKC · 10956.9%18Darnell WashingtonPIT · 4456.9%19Adam TrautmanDEN · 2356.8%20Jake TongesSF · 4656.8%21Colston LovelandCHI · 8356.8%22Dallas GoedertPHI · 8256.5%23Dalton SchultzHOU · 10756.4%24Charlie KolarBAL · 1556.4%25Darren WallerMIA · 3556.2%26Mason TaylorNYJ · 6655.9%27John BatesWAS · 1655.9%28Greg DulcichMIA · 3455.8%29Tyler HigbeeLA · 3655.7%30Mo Alie-CoxIND · 2055.6%31Josh OliverMIN · 2055.6%32Ben SinnottWAS · 1355.5%33Hunter LongJAX · 1755.3%

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.

Target success rate · 2025 · full board

Target success rate leaderboard for the 2025 NFL season, shrunk by empirical Bayes with 90% credible intervals, ranked by the shrunk estimate.
#PlayerTeamPostargetsRawShrunk90% interval
1George Kittle◦ provisionalSFTE6968.1%60.5%54.1%66.7%
2Sam LaPorta◦ provisionalDETTE4967.3%59.1%52.2%65.9%
3AJ Barner◦ provisionalSEATE6864.7%59.0%52.5%65.3%
4Hunter Henry◦ provisionalNETE8763.2%58.9%52.7%64.8%
5Austin Hooper◦ provisionalNETE2673.1%58.8%51.2%66.1%
6Daniel Bellinger◦ provisionalNYGTE2673.1%58.8%51.2%66.1%
7Oronde Gadsden II◦ provisionalLACTE6963.8%58.6%52.1%64.9%
8Colby Parkinson◦ provisionalLATE5664.3%58.3%51.6%65.0%
9Trey McBrideARITE17060.0%58.1%53.1%63.1%
10Brenton Strange◦ provisionalJAXTE6063.3%58.1%51.5%64.6%
11Dawson Knox◦ provisionalBUFTE5064.0%58.0%51.1%64.7%
12Jackson Hawes◦ provisionalBUFTE1973.7%58.0%50.2%65.6%
13Tucker Kraft◦ provisionalGBTE4463.6%57.6%50.6%64.5%
14Mitchell Evans◦ provisionalCARTE2568.0%57.6%50.0%65.0%
15Ian Thomas◦ provisionalLVTE1376.9%57.5%49.4%65.3%
16Dalton Kincaid◦ provisionalBUFTE5062.0%57.3%50.4%64.0%
17Travis KelceKCTE10958.7%56.9%51.1%62.6%
18Darnell Washington◦ provisionalPITTE4461.4%56.9%49.8%63.8%
19Adam Trautman◦ provisionalDENTE2365.2%56.8%49.1%64.3%
20Jake Tonges◦ provisionalSFTE4660.9%56.8%49.8%63.6%
21Colston Loveland◦ provisionalCHITE8359.0%56.8%50.5%62.9%
22Dallas Goedert◦ provisionalPHITE8258.5%56.5%50.3%62.6%
23Dalton SchultzHOUTE10757.9%56.4%50.6%62.2%
24Charlie Kolar◦ provisionalBALTE1566.7%56.4%48.4%64.2%
25Darren Waller◦ provisionalMIATE3560.0%56.2%48.9%63.3%
26Mason Taylor◦ provisionalNYJTE6657.6%55.9%49.4%62.4%
27John Bates◦ provisionalWASTE1662.5%55.9%47.9%63.6%
28Greg Dulcich◦ provisionalMIATE3458.8%55.8%48.5%63.0%
29Tyler Higbee◦ provisionalLATE3658.3%55.7%48.4%62.9%
30Mo Alie-Cox◦ provisionalINDTE2060.0%55.6%47.9%63.3%
31Josh Oliver◦ provisionalMINTE2060.0%55.6%47.9%63.3%
32Luke Musgrave◦ provisionalGBTE3158.1%55.5%48.1%62.9%
33Ben Sinnott◦ provisionalWASTE1361.5%55.5%47.5%63.5%
34Brock Bowers◦ provisionalLVTE8756.3%55.5%49.3%61.6%
35Hunter Long◦ provisionalJAXTE1758.8%55.3%47.4%63.1%
36Juwan JohnsonNOTE10255.9%55.3%49.4%61.2%
37Tommy Tremble◦ provisionalCARTE3756.8%55.3%48.0%62.4%
38Noah Fant◦ provisionalCINTE4156.1%55.1%48.0%62.2%
39Kyle PittsATLTE11955.5%55.1%49.5%60.7%
40Mike Gesicki◦ provisionalCINTE4355.8%55.0%47.9%62.1%
41Luke Farrell◦ provisionalSFTE1457.1%55.0%47.0%62.9%
42Pat Freiermuth◦ provisionalPITTE5455.6%55.0%48.2%61.7%
43Davis Allen◦ provisionalLATE3354.5%54.6%47.3%61.9%
44Tanner Hudson◦ provisionalCINTE2454.2%54.6%46.9%62.1%
45Grant Calcaterra◦ provisionalPHITE1353.8%54.6%46.5%62.5%
46Brevyn Spann-Ford◦ provisionalDALTE1353.8%54.6%46.5%62.5%
47John FitzPatrick◦ provisionalGBTE1553.3%54.5%46.5%62.4%
48Julian Hill◦ provisionalMIATE2152.4%54.2%46.5%61.9%
49Isaiah Likely◦ provisionalBALTE3652.8%54.1%46.9%61.4%
50Cade Stover◦ provisionalHOUTE1650.0%54.0%46.0%61.8%
51Brock Wright◦ provisionalDETTE2250.0%53.8%46.0%61.4%
52Elijah Higgins◦ provisionalARITE3751.3%53.7%46.5%60.9%
53Johnny Mundt◦ provisionalJAXTE1947.4%53.4%45.6%61.2%
54Zach Ertz◦ provisionalWASTE7350.7%52.9%46.5%59.3%
55T.J. Hockenson◦ provisionalMINTE6650.0%52.7%46.2%59.2%
56Cole Kmet◦ provisionalCHITE4949.0%52.7%45.7%59.6%
57Mark Andrews◦ provisionalBALTE7050.0%52.6%46.2%59.1%
58Tanner Conner◦ provisionalMIATE1540.0%52.6%44.6%60.5%
59Ja'Tavion Sanders◦ provisionalCARTE3447.1%52.6%45.3%59.9%
60Luke Schoonmaker◦ provisionalDALTE2343.5%52.4%44.7%60.1%
61David Njoku◦ provisionalCLETE4847.9%52.3%45.4%59.3%
62Jeremy Ruckert◦ provisionalNYJTE2944.8%52.3%44.8%59.8%
63Taysom Hill◦ provisionalNOTE1637.5%52.1%44.2%60.0%
64Will Dissly◦ provisionalLACTE1637.5%52.1%44.2%60.0%
65Drew Sample◦ provisionalCINTE2040.0%52.0%44.2%59.8%
66Elijah Arroyo◦ provisionalSEATE2642.3%51.9%44.3%59.5%
67Cade Otton◦ provisionalTBTE8248.8%51.9%45.6%58.1%
68Michael Mayer◦ provisionalLVTE5046.0%51.6%44.7%58.5%
69Jonnu Smith◦ provisionalPITTE5545.5%51.2%44.4%58.0%
70Gunnar Helm◦ provisionalTENTE5545.5%51.2%44.4%58.0%
71Tyler WarrenINDTE11448.3%51.1%45.4%56.8%
72Jake FergusonDALTE10347.6%50.9%45.0%56.8%
73Terrance Ferguson◦ provisionalLATE2536.0%50.6%43.0%58.3%
74Chig Okonkwo◦ provisionalTENTE7944.3%49.9%43.6%56.1%
75Noah Gray◦ provisionalKCTE3834.2%48.6%41.4%55.9%
76Theo Johnson◦ provisionalNYGTE7440.5%48.3%42.0%54.7%
77Harold Fannin Jr.CLETE10841.7%47.6%41.8%53.4%
78Evan Engram◦ provisionalDENTE7739.0%47.5%41.2%53.8%

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).