Do Running Back Rush Attempts Translate From College to the NFL?

Antonio Gibson barely ran the ball in college and then burst onto the scene in his NFL rookie season. How much do college rush attempts matter when projecting NFL workload?

My favorite high school memory came during my junior year. I was in pre-calculus for my math class, and it was a challenge. I understood a lot of what we tackled in that class, but really abstract concepts seemed pointlessly confusing.

One of my classmates must have felt the same because, when our teacher told us to take out our notes to review for the test on “imaginary numbers,” my classmate didn’t move a muscle. The teacher noticed, quirked an eyebrow, and asked him why he wasn’t doing anything.

My classmate responded, without missing a beat, “Well, you see… This was the chapter I took imaginary notes on."

Like my classmate, I spent a lot of time in high school and college wondering, Is this actually going to matter in real life? Will I ever use this in the real world? Most of the time, this thought would crop up in math classes -- an irony that is far from lost on me, as I now write for a stats-based website called “numberFire.”

I’m sure college running backs often feel much the same way. Not only do I bet many of them would like to drop their calculus classes for art history, but I’d bet they also want to know the rushing work they do in school will be worth something in their eventual pro careers. As a fantasy analyst, I want to know about that too. So, that’s what we’re examining today.

Do rushing attempt workloads carry over from college football to the NFL?

Show Your Work

I have previously done a study where we looked at whether running back receptions carried over from college to the NFL. I found a pretty strong relationship, which prompted me to ask the same question about rushing workload.

numberFire has done similar studies to this, such as this one last year that looked at the threshold of college work it took to be an eventual NFL fantasy star. The study from Sam Hoppen uncovered some interesting conclusions, but I wanted to take an even broader approach. I decided to ask if we could not only point out the NFL stars but also find the part-timers and contributors (or even the eventual busts) -- all based on their college production on the ground.

I took a sample of 402 running backs who entered the NFL between 2000 and 2018 and who went to a Division I college program. I then compared their college rushing production to their NFL production from solely their first three years in the league -- about the average career length of an NFL running back.

I also specifically used per-game or market share (percentage of the team total) rates to help control for both availability and scheme. A running back who is injured will obviously play fewer games, so per-game analysis won’t penalize his low counting stats, and being on a team that hardly throws (like some smaller programs) won’t penalize a player due to using market share.

With all of this in place, we are now equipped to dig in: if you rush a ton in college, are you likely to rush in the NFL, as well?

Numbers Don’t Lie

The way I analyzed this data is simple: I took the average NFL running back’s rushing attempts per game and team rushing share from 2000 to 2018 and then compared that to the NFL production of players who saw very little and a lot of work in college.

I bucketed players who were in the 25th percentile or lower (fewer than 9.0 rushes per game, or fewer than 20% team rushing share) and in the 75th percentile or higher (more than 15.0 rushes per game, or more than 33% team rushing share) and averaged those buckets’ eventual NFL production. The results ended up pretty striking.

The tables below show collegiate rushes per game and collegiate team rushing share and how those translate to the NFL on average, respectively.

2000-2018 NFL Rushes
Per Game
NFL Team Rushing
Market Share
Fewer Than 9.0 College Rushes Per Game 2.8 6.9%
NFL Average 5.4 14.9%
More Than 15.0 College Rushes Per Game 7.9 22.4%

This is the data based on market shares.

2000-2018 NFL Rushes
Per Game
NFL Team Rushing
Market Share
Fewer Than 20% College Team Rushing Market Share 3.0 8.1%
NFL Average 5.9 16.7%
More Than 33% College Team Rushing Market Share 8.0 23.4%

In my previous piece, I found that there was about a 0.40 R-value to the correlation between college receiving rates and NFL rates -- a noteworthy strength for predictive football statistics. As for the relationship between college and NFL rushing rates? It’s nearly identical in strength to its aerial counterpart.

As we can see from the above groupings of data, the low-end college rushers ended up seeing about half the volume in the NFL that the average NFL running back did over the last 20 years. The high-end college rushers saw about 50% more work on the ground than the league average. Also, it seems that rushes per game in college is actually the more predictive metric, producing slightly more distinct results between the ends of the workload spectrum.

It seems that NFL coaches may be more willing to ride with a player they have seen handle heavy workloads in college than they are to trust players who didn’t earn big volume collegiately. We also might assume that as the NFL has gotten more pass-happy and experimental, these correlations would become weaker. The offensive coordinators of today’s pro game are, for the most part, less bound by tradition and may give more rushing chances to players such as Antonio Gibson -- who basically never ran the ball in college -- right?

2011-2018 NFL Rushes
Per Game
NFL Team Rushing
Market Share
Fewer Than 9.0 College Rushes Per Game 3.2 8.4%
NFL Average 4.6 12.8%
More Than 15.0 College Rush Per Game 8.4 23.5%

While the R-value for recent years (2011-18) between college and NFL rushes per game drops, it goes to only 0.38; that’s still plenty good. The difference between the low-end and high-end buckets gets a little muddied over the last decade, but there’s still a clear preference by NFL teams to use running backs who have rushed a lot before.

All of this is backed up by Sam’s aforementioned piece, where he concludes that the ideal top-tier rusher for fantasy has probably passed the 500-carry threshold in their college years.

Division I college teams have averaged 472 rushing attempts per season since the new millennium. That would make the per-year average for those 500-rush backs 125 to 167 attempts per year (if they play four or three seasons in college, respectively) with an average team rushing share of 26.5% to 35.3% -- right in line with our study’s threshold. As for the average rushes per game -- 10.4 to 13.9, which is just about in line with our threshold there, as well.

Extrapolating the Curve

Like calculus is to physics, data in a vacuum is all interesting and well, but it’s not worth much if I can’t apply it to something tangible. It won’t matter to you, playing fantasy in 2021, that Najeh Davenport's college rushing profile indicates he won’t be used as an early-down workhorse. No, no, no -- we need to look at the two most recent draft classes to see which still-young NFL backs have hidden rushing potential and which are potential busts.

First, the table below shows running backs who missed both low-end college thresholds in carries per game and team rushing market share.

Low Rushing, 2019-20 Draftees NFL Rushes
Per Game
NFL Team
Rushing Share
CFB Rushes
Per Game
CFB Team
Rushing Share
Antonio Gibson 12.1 42.5% 1.7 2.8%
Alec Ingold 0.4 3.0% 2.9 4.4%
Tony Pollard 6.0 21.3% 3.5 9.0%
Raymond Calais 0.0 0.0% 4.8 10.9%
Javon Leake 0.0 0.0% 4.8 10.8%
Khari Blasingame 0.0 0.0% 5.5 13.4%
Tony Jones 3.0 0.6% 6.7 18.5%
Sewo Olonilua 0.0 0.0% 7.1 17.2%
Dexter Williams 1.0 0.8% 7.3 12.8%
Miles Sanders 12.3 40.0% 7.7 18.3%
Jamycal Hasty 4.9 8.9% 8.6 18.8%
D'Ernest Johnson 1.2 4.2% 8.6 19.4%
Damien Harris 11.8 14.9% 8.8 19.4%

The most fantasy-noteworthy of these players are Antonio Gibson, Tony Pollard, Miles Sanders, and Damien Harris.

Gibson and Pollard were actually used as slot receivers and scatbacks at Memphis -- one after the other -- so I don’t have much trepidation about their ability to sustain a strong NFL rushing workload. Context for these two actually satisfies me that they were getting much rushing work at all while still being solid offensive weapons.

Sanders has found his way into the bottom grouping in this category as well as the receiving one from my receptions piece, which does give some cause for concern. That said, in college he was behind an extremely gifted running back in Saquon Barkley, who gobbled up a ton of the touches in both phases of the game. Some assuaging of fear can be done here, though it shouldn’t surprise anyone that the Philadelphia Eagles continue to try to upgrade their depth at that position.

Finally, Damien Harris fits into a similar situation to Sanders' from his time at Alabama, sharing a backfield at various points with Derrick Henry, Kenyan Drake, Josh Jacobs, Bo Scarbrough, and even 2021 rookie Najee Harris. He was right on the cusp of the thresholds, too, so I’d be willing to extend the benefit of the doubt to him if the New England Patriots commit to him.

The next table shows the top backs from the last two draft classes who met both high-end thresholds.

High Rushing, 2019-20 DrafteesNFL Rushes
Per Game
NFL Team
Rushing Share
CFB Rushes
Per Game
CFB Team
Rushing Share
Alexander Mattison7.520.8%15.337.7%
Trayveon Williams1.23.3%15.839.3%
Cam Akers11.230.7%16.344.0%
David Montgomery15.862.1%16.946.3%
J.K. Dobbins8.924.1%17.339.9%
Justice Hill2.56.1%17.641.8%
Myles Gaskin10.522.9%18.246.3%
Devin Singletary11.035.0%18.842.0%
Joshua Kelley7.923.8%20.648.3%
Jonathan Taylor15.550.5%22.651.4%
AJ Dillon4.210.4%24.146.3%

There are a variety of fantasy-relevant (especially dynasty) names on this list -- some who have performed well already and some who have flopped. It might be worth asking around about guys like Trayveon Williams, Justice Hill, or Joshua Kelley in your dynasty leagues and taking a flier on them.

As we’ve seen, a big college rushing workload isn’t a guarantee of NFL usage, but it tends to matter. And that probability is worth something in the fantasy world.