Last active
December 2, 2019 13:14
-
-
Save psinger/103a0b0d433e4d61bf12ed218abce89f to your computer and use it in GitHub Desktop.
2018 Yards per Attempt stats for players with more than 50 rushing plays
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Name # plays Actual Expected Ratio Dif | |
Kerryon Johnson 118 5.432 3.755 1.446 1.677 | |
Kenneth Dixon 60 5.550 4.201 1.321 1.349 | |
Saquon Barkley 261 5.008 3.738 1.340 1.270 | |
Nick Chubb 190 5.184 3.931 1.319 1.253 | |
Derrick Henry 200 4.825 3.695 1.306 1.130 | |
Jordan Wilkins 60 5.600 4.558 1.229 1.042 | |
Aaron Jones 132 5.485 4.465 1.228 1.020 | |
Zach Zenner 55 4.818 3.845 1.253 0.973 | |
Phillip Lindsay 189 5.466 4.510 1.212 0.955 | |
Isaiah Crowell 143 4.790 3.919 1.222 0.872 | |
Rashaad Penny 85 4.929 4.071 1.211 0.859 | |
Gus Edwards 137 5.241 4.450 1.178 0.791 | |
Lamar Miller 207 4.638 3.865 1.200 0.772 | |
C.J. Anderson 67 6.015 5.324 1.130 0.691 | |
Joe Mixon 237 4.928 4.245 1.161 0.684 | |
Ezekiel Elliott 297 4.758 4.098 1.161 0.659 | |
Christian McCaffrey 216 4.958 4.351 1.140 0.607 | |
Matt Breida 153 5.320 4.782 1.113 0.539 | |
James Conner 214 4.542 4.004 1.134 0.538 | |
Chris Carson 247 4.660 4.159 1.121 0.501 | |
Frank Gore 155 4.626 4.172 1.109 0.454 | |
Mike Davis 111 4.631 4.230 1.095 0.400 | |
Latavius Murray 139 4.151 3.776 1.099 0.375 | |
Kareem Hunt 181 4.552 4.185 1.088 0.368 | |
Tevin Coleman 167 4.790 4.457 1.075 0.334 | |
Bilal Powell 80 4.287 3.955 1.084 0.332 | |
Mark Ingram 138 4.674 4.349 1.075 0.325 | |
Marlon Mack 195 4.656 4.380 1.063 0.276 | |
Marshawn Lynch 90 4.178 3.926 1.064 0.252 | |
Jalen Richard 55 4.709 4.476 1.052 0.233 | |
Austin Ekeler 105 5.114 4.901 1.044 0.213 | |
Adrian Peterson 251 4.151 3.990 1.041 0.162 | |
Royce Freeman 124 3.927 3.809 1.031 0.119 | |
Kenyan Drake 118 4.398 4.308 1.021 0.091 | |
Todd Gurley 256 4.887 4.804 1.017 0.083 | |
Dalvin Cook 131 4.611 4.604 1.001 0.007 | |
Jaylen Samuels 55 4.636 4.649 0.997 -0.013 | |
Marcus Murphy 51 4.824 4.869 0.991 -0.045 | |
Melvin Gordon 175 5.057 5.103 0.991 -0.046 | |
T.J. Yeldon 102 3.990 4.089 0.976 -0.099 | |
Doug Martin 172 4.203 4.305 0.976 -0.102 | |
Alvin Kamara 191 4.539 4.643 0.978 -0.103 | |
David Johnson 254 3.657 3.770 0.970 -0.112 | |
Leonard Fournette 133 3.301 3.419 0.965 -0.118 | |
Josh Adams 119 4.294 4.445 0.966 -0.150 | |
Peyton Barber 234 3.722 3.891 0.957 -0.169 | |
Jamaal Williams 121 3.835 4.051 0.947 -0.216 | |
Sony Michel 209 4.455 4.700 0.948 -0.245 | |
Alfred Morris 111 3.856 4.118 0.936 -0.262 | |
Jordan Howard 249 3.755 4.027 0.932 -0.272 | |
Alfred Blue 149 3.282 3.570 0.919 -0.288 | |
James White 93 4.559 4.872 0.936 -0.313 | |
Chase Edmonds 59 3.186 3.570 0.892 -0.384 | |
Nyheim Hines 83 3.578 3.985 0.898 -0.406 | |
Dion Lewis 154 3.312 3.728 0.888 -0.416 | |
Alex Collins 112 3.625 4.063 0.892 -0.438 | |
Carlos Hyde 172 3.320 3.803 0.873 -0.483 | |
Elijah McGuire 88 3.045 3.587 0.849 -0.541 | |
Wendell Smallwood 87 4.184 4.731 0.884 -0.547 | |
Tarik Cohen 94 4.362 4.948 0.882 -0.586 | |
Jeff Wilson 66 4.030 4.624 0.872 -0.593 | |
Chris Ivory 114 3.342 3.941 0.848 -0.599 | |
Wayne Gallman 51 3.451 4.201 0.822 -0.750 | |
Ito Smith 89 3.596 4.382 0.821 -0.786 | |
Rex Burkhead 56 3.304 4.198 0.787 -0.894 | |
Corey Clement 68 3.809 4.776 0.798 -0.967 | |
LeGarrette Blount 154 2.714 3.692 0.735 -0.977 | |
LeSean McCoy 157 3.242 4.258 0.761 -1.016 | |
Giovani Bernard 56 3.768 4.845 0.778 -1.078 | |
Expected yards calculated by our model explained in https://www.kaggle.com/c/nfl-big-data-bowl-2020/discussion/119400. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Thanks for sharing this!
Based on the play-level data here...
https://www.kaggle.com/dott1718/nfl-big-data-bowl-expected-yards
I computed a slightly different ranking, using standard errors for the "Diff" column. We can compute the 95% confidence intervals for the "mean yards gained vs. expected" for each player, and rank the players by the lower bounds of those intervals. This ranking basically penalizes players with small sample sizes and/or high variability in their play-level "Diff" values.
This "confidence interval" ranking is pasted below. A couple thoughts:
In 2018, almost zero players had statistically significant estimates for their "mean yards gained vs. expected". This goes along with the recent theory that running backs are practically interchangeable, and that no individual RB adds that much value.
The column 'mean_rank' is the ranking based purely on the point estimate of "Diff", aka "yards gained vs. expected". This gives an idea of what changes when you account for variability in addition to the point estimates.
One major drawback of ranking RBs this way is that their speed at the time of the handoff is explicitly accounted for in the model. So this ranking will penalize RBs with high speed at the time of the handoff (which is a good attribute for an RB to have!). These RBs basically give themselves a higher bar to clear, i.e. a higher value of "expected", so it's not fair to penalize them for not meeting or exceeding that bar.
Anyway, still cool. Seems like Kerryon Johnson really had a good 2018 season: