How Does College Production Affect Draft Position and NFL Success? RB Edition
I began research for this piece with a certain goal in mind: To find out if college production for running backs has a noticeable effect on draft position and/or on future NFL production. I originally had intended for my research to find how strong correlations were between stats like rushing yards/year or rushing yards/game in college and future draft position or future NFL rushing yards/year and game. I began looking at the last 10 draft classes of running backs, specifically those taken in the first 2 rounds of the NFL draft. In doing so, I found there were 45 running backs taken in the first 2 rounds of the past 10 NFL drafts from 2008-2017. Then I began noting each runners college and NFL production, including rushing yards, receiving yards, total yards and total TD’s scored. I then marked down each RB’s draft position in their respective draft classes, and assigned a draft score to each player. As I only was looking at the first 2 rounds of the draft, the scores are out of 64. A Draft Score of 63 indicates the player was drafted #1 overall, thus is stronger, while a Draft Score of 0 indicates the player was drafted last in the 2nd round, or 64th overall, in general. Here is the table of the 45 RB’s taken in the first two rounds of the Draft next to their respective draft class and draft score:
This draft score system allowed me to reward players for being drafted higher in the draft, which in turn allowed me to run correlations on college production and draft results. Ill link the table of my findings here below before going into my analysis of the data, and what I was able to learn about the NFL’s feelings towards running backs.
As the correlation becomes a deeper green, the more positive the relationship between the two categories, while the redder, the more negative the correlation.
Considering what each mean, you find most of the correlations are rather weak. To find some of the more interesting results, I considered the college touchdowns column on the far right. In the entire 10-year period of this study, there is a very slightly negative correlation between college touchdowns scored and draft position, as well as future NFL TD’s scored. Once I noted that, I decided to break the 10-year period into 2 distinct parts to find out if this trend is consistent or if there was a divergence somewhere in between. As it turns out, the periods of 2008-2013 show one pattern, while the 2014-2017 draft classes have shown a completely new, or stronger, pattern in nearly every data set. The 2014-2017 draft classes showed a slightly positive relationship between touchdowns scored in college and draft position, and a more positive relationship between college scores and future NFL scores. On the flip side, the 2008-2013 draft classes showed noticeable negative correlations between both relationships. When the numbers are combined in the total 10-year period, we see a relationship that is essentially random, with almost no strong pull in either direction.
Next, take a look at the only column we see the 2008-2013 period show a stronger positive correlation than the 2014-2017 period: College Receiving yards and draft position. The 08-13 classes showed a slightly more positive relationship between the two, while actually showing a slightly lower positive correlation between those same college receiving yards and future NFL receiving yards. To decipher that, it appears that NFL teams weighed receiving production from RB’s more in the 2008-2013 years, however ended up missing on evaluation more often than the 2014-2017 classes did. With a .53 correlation between college receiving yards and NFL receiving yards the 2014-2017 classes showed the strongest correlation out of the entire data set, meaning NFL GM’s chose backs who were capable receivers in school and were able to continue utilizing those skill sets at the highest level.
Taking in all of the data at once, the logical conclusion follows the general feeling in and around NFL media: Running backs are regarded less highly now than they were 10 years ago. The 2008 class saw 7 runners taken in the first 2 rounds, and the 2008-2013 classes averaged over 5 running backs taken in rounds 1 and 2. The 2014-2017 draft classes averaged just over 3 RB’s taken in those first 2 rounds, with 2015 and 2017 having 4 runners taken each.
Over the 10-year span of this data set, there is a noticeable decline in running backs taken with first picks, which may help explain away some of the variance we see in the correlations. With fewer players contributing to the data sets in those years, there is expected to be a higher overall variance in the numbers. Also worth considering is the smaller career data sets for the players taken in the 2014-2017 classes, as many of them only have a few years under their belt (the 17 class only played this past season). While there are many players in the 08-13 classes that didn’t pan out and only played a handful of seasons, the data set is affected by the smaller data sets in the newer draft classes.
My conclusions from this research are the following: I can see a noticeable decline in running backs taken early in drafts in recent years, while many of the select early rounders produce close to their college levels in the NFL. The higher and stronger correlations between draft position and yards compared to TD’s and draft position is what I expected, as teams will put a premium on overall production over a lack of scoring in most college situations. The one thing this data set fails to incorporate is the pedigree of running backs going into the draft. Often superior stats don’t always mean higher draft status, as we see with the lower strength correlations. Big name recruits from the high school and college levels will carry automatic heavier weights in their draft classes, and that is much more difficult to account for quantitatively. This helps explain why successful players in college like Montee Ball, who averaged about 1,400 yards and 20 TD’s per season in college, were still drafted 58th overall as the 3rd runner in his class (behind Giovanni Bernard and LeVeon Bell).
Hopefully you can find some conclusions in this data set that I might have missed, and you enjoyed reading through some of my findings!
This will be a recurring series of sorts as I break down previous draft classes and try to evaluate what is important at the college level to produce a truly successful pro, all culminating in a final mock draft as we get closer to the 2018 NFL Draft!
All stats that were used for this project were found viaESPN.
I am a 22 year old college graduate from Connecticut who has been a lifelong sports fan following the Minnesota Vikings as a die-hard fan.
Firm “Numbers Never Lie” believer. Tweet me your hot takes!