The first scrimmage of the spring produced plenty of angst from Florida fans on Friday.
Feleipe Franks struggled (6-22, 133 yards) and for those who’ve already made up their minds about Franks, it was just more ammunition to say he can’t hack it. Kyle Trask played the best (11-19, 176 yards), but it’s obviously not ideal to have a lower-rated 3-star recruit (2123rd nationally) outplaying the blue-chip recruit (Franks was ranked 54th nationally). (scrimmage stats courtesy Jake Winderman)
This has obviously led many Gators fans to turn to other aspects of the offense that could potentially pick up the slack. The obvious choice is at running back.
Ignoring the fact that this is one scrimmage early in spring practice and the QBs will improve, it’s still worth looking at that assumption. With Lamical Perine starting and Malik Davis emerging last season, along with Jordan Scarlett returning from suspension, the expectation is that Florida will be carried by its backs.
But is that a reasonable assumption?
What is an elite running back performance?
The idea for this article originated from a Twitter discussion last week. I was responding to someone who called Jordan Scarlett at top-10 running back in 2016 and I made a flippant remark about 7 yards per rush being a cut-off for elite running backs.
122 RBs averaged more yards per carry in 2017 than Scarlett in 2016. Perine averaged almost a yard less (4.1) than Scarlett (5.0). Elite RBs are over 7. Davis was at 6.7. Lemons at 7.2 but only 19 carries.
— Will Miles (@WillMilesSEC) March 30, 2018
I was criticized for that relatively arbitrary cut-off. After all, Adrian Peterson (5.4), Le’Veon Bell (5.0) and LeSean McCoy (4.8) had averages well below that, and close to that of Scarlett (5.0).
But looking back with hindsight at guys we know were successful in the NFL is a bit unfair. It’s a lot like looking at Baker Mayfield (3-star recruit) blossom into an outstanding player and declaring that stars don’t matter.
But we know that they do in aggregate. In the same way, there has to be some statistical cut-offs that we use to classify running backs as elite, above average, average or below average. The issue is that with running backs there is the issue of efficiency and usage.
Take Alabama’s Derrick Henry for example. Henry won the Heisman in 2015, on the back of 395 (!) carries for 2219 yards and a 5.6 yard per carry average. But Leonard Fournette averaged 6.5 yards per rush on 300 carries, on a team with significantly worse QB play.
Who had the more impressive season, Henry or Fournette? Obviously Henry won the Heisman. But I think you could argue convincingly that what Fournette did was more impressive. Are they both good players? Undoubtedly. But if you had to give me one of them after that season, I would’ve taken Fournette.
That same season, Florida State running back Dalvin Cook averaged 7.4 yards per rush on 229 carries. Again, I think you could argue that Cook was a better back than either Henry or Fournette. Cook also had years where he averaged 5.9 and 6.1 yards per rush, so he wasn’t over 7 yards per carry for his career. But I do think there is something to be said for a player who can carry the ball more than 200 times and be that efficient.
This bears out if you look at the numbers in aggregate. What I’ve done here is compiled the statistics for the top 280 runners (running backs and quarterbacks) in 2015, 2016 and 2017. Those players averaged 5.27 yards per carry (black line) over the three years, so that’s what we should consider an “average” performance.
Running back efficiency (yards per attempt) versus usage (rushing attempts) for the top 280 backs from 2015-2017. (Will Miles/Read and Reaction)
The first thing to note is that there is a definite penalty for getting more carries from an efficiency standpoint. This should be fairly intuitive, as it is easier to average 8 yards per carry on 100 carries than it is on 300.
The second thing to note is that there are two ways to be an outlier. A back who has over 300 carries is clearly doing something that is rare (13 times in the 3 seasons).The majority of those backs were better than the average, but there aren’t any significantly over 6 yards per rush (D’Onta Foreman in 2016 at 6.3 is the highest). The three green circles are the performances of Henry (393 carries), Fournette (300 carries) and Cook (229 carries).
However, there is another way to be an outlier, which is to combine efficiency and usage. In this case, if you set cut-offs of 200 carries and 7 yards per carry, that was even rarer (6 times in the 3 seasons). The most impressive player to show this combination was Stanford’s Bryce Love in 2017. Love averaged 8.1 yards per rush on 263 carries (blue circle). Love also averaged 7.1 yards per rush on 111 carries in 2016, so he actually increased his efficiency with more opportunities.
The data in the chart above also allows us to rate running backs compared to their peers. The table below outlines the yards and yards per attempt from the top 280 running backs from 2015, 2016 and 2017 divided into the 90th, 75th and 50th percentile.
NCAA running back performance from 2015-2017 divided into percentiles. (Will Miles/Read and Reaction)
So I was wrong in my tweet. If we define elite running back performance as the top 10 percent of the top 280 backs in the country, they average 6.8 yards per carry not 7.
Obviously, elite performance in this particular metric has to decrease with increased usage. Peterson had 339 carries for 1925 yards in 2004. That’s a monster year. But had he only tallied 200 carries at the same average, he would have accumulated 1136 yards, or slightly worse than Cal running back Patrick Laird’s 2017 season.
Interestingly, there is very little variation in the percentiles from year-to-year. Additionally, the same analysis of SEC running backs from 2012-2017 shows very similar averages, meaning that playing in a tougher conference doesn’t change how we should rate a running back.
Application to Florida’s running backs
Why is this important? Well, Florida’s running backs don’t have much of a track record when it comes to efficiency or usage.
As mentioned in the tweet above, Jordan Scarlett would have ranked 122nd in efficiency (yards per carry) had he put up numbers identical to his 2016 performance last season. He also would have ranked only 70th in total yards despite his 179 carries, which would have ranked 66th most.
Performance of Florida running backs in their most recent season. (Will Miles/Read and Reaction)
Scarlett’s numbers aren’t poor, but they’re right in line with what I would classify as an average performance. His 889 yards is between the 50th and 75th percentiles, but his 5.0 yards per attempt is below the 50th percentile. This is the profile of a good player. But it isn’t the profile of a game breaking back.
Each of Florida’s other running backs have red flags as well.
Lamical Perine is a tough back, and has had some runs that are really impressive. But in aggregate, his 136 carries and 4.1 yards per carry average is well below average. Perine was a little better in 2016, but even then only averaged 4.6 yards per carry on 91 carries. The reality is that based on these numbers, Perine has performed below the average of his peers.
Malik Davis burst onto the scene last year as the most effective Florida back, and it’s was easy to see even for casual fans. His 6.7 yards per rush is really good, even with only 79 carries. If he could maintain that efficiency with a higher usage, Florida would really have something.
The good sign is that as his usage increased significantly early in the season (4 carries vs. Tennessee, 21 the next week vs. Kentucky), his efficiency didn’t fall off much at all. The bad news is that he is recovering from what is thought to be a significant knee injury. There aren’t any guarantees that he will be the same player as he makes his way back, especially early in the season.
As a bit of a cautionary tale, Georgia’s Nick Chubb averaged 7.1 yards per rush on 219 carries in 2014 and 8.1 yards per rush in 2015 before his knee injury. But he only averaged 5.0 and 6.0 yards per rush on 200-plus carries in 2016 and 2017, respectively. There was clearly an element of explosion missing from Chubb’s game in 2016 that came back in 2017. It’s possible the same thing could apply to Davis.
Adarius Lemons averaged 7.2 yards per rush, so he was quite efficient. But he only had 19 carries, with 11 of those carries and 89 of those yards coming against an overmatched UAB squad. He has shown explosiveness elsewhere, particularly on a couple of plays in last week’s scrimmage, but the jury is still out on whether he will be explosive on a play-by-play basis in the SEC.
4-star recruits Dameon Pierce (201st nationally) and Iverson Clement (329th nationally) fill out the Gators collection of running backs. Pierce lit up the circuit in high school, with a 9.4 yards per carry average on 227 carries. This was significantly up from his averages of 6.7 and 6.9 yards per carry in 2015 and 2016 and he has carried a heavy workload. Clement didn’t have quite as explosive a senior year (7.1 yards per rush in 199 carries), but has been pretty explosive overall in his career, averaging 6.0 and 7.4 yards per rush in 2015 and 2016.
Both Pierce (7.2) and Clement (7.0) had very good rushing averages in their high school careers. But for reference, Scarlett averaged 8.1 yards per rush in high school. Whether these stats transfer is not something I’ve looked into in much detail. But Scarlett’s transition to what is essentially an average back after putting up those numbers in high school concerns me.
None of this means that one of these players won’t put in an elite-level performance in 2018. You don’t have to look far to see what a coaching, a QB upgrade and better health can do. Sony Michel and Chubb averaged 5.5 and 5.0 yards per carry in 2016, respectively. They averaged 7.4 and 6.0 in 2017 as the overall play of the offense improved. There is quite a bit of year-to-year variability in running back efficiency.
Comparison of Florida’s RBs to the rest of the SEC
Even with the shortcomings of each back, Florida does have enviable depth compared to many teams. The problem is that the Gators depth at running back doesn’t really differentiate them from its chief competition at the top of the conference: Georgia and Alabama.
While Georgia did just lose Chubb and Sony Michel to graduation, there is still quite a bit of production coming back to Athens. D’Andre Swift had 81 carries (similar to Malik Davis), but he amassed a per-carry average of 7.6. Elijah Holyfield had 50 carries with a 5.9 yard average and Brian Herrien had 61 carries with an average of 4.3.
And Kirby Smart has reinforcements coming with two commits from the 2018 class. Zamir White is the number one running back prospect in the country (9th overall) and James Cook is ranked the third best all-purpose back (41st overall). White is the true elite 5-star talent who should be able to step in the first day and Cook isn’t far behind.
At Alabama, the return of Damien Harris combined with the performance of Najee Harris in the National Championship Game should have the Tide feeling pretty comfortable about the position. Not only do those players have a 5-star pedigree, but they averaged 7.4 (Damien) and 6.1 (Najee) yards per rush in 2017. Combine that with Josh Jacobs (6.2) and Brian Robinson (6.9), plus 4-star recruit Jerome Ford and Alabama is going to be really strong at the running back position yet again.
Depending on how you look at the statistics of the returning players of each team gives a good look at where Florida’s running backs sit in the SEC pecking order. The chart below shows the aggregated 2017 performance of each team’s returning running backs sorted by yards per attempt.
Stats of returning running backs for SEC teams in 2018. (Will Miles/Read and Reaction)
It’s no surprise that Alabama and Georgia are at the top of this list given what I wrote above. Interestingly though, Missouri has a relatively formidable running game coming back to Columbia this season and Auburn still has quite a bit returning, albeit with a lot less experience with the NFL departure of Kerryon Johnson.
Florida ranks sixth on this list. The Gators 5.1 yards per carry average is just slightly below the SEC average from 2012-2017 (5.35). If Scarlett is excluded from the Gators stats, the Gators average rises to 5.23, moving Florida just above Arkansas.
Even when looking at usage, Florida falls in the same range. The Gators do have the most aggregate yardage gained on this chart, but that is with Scarlett and some combination of Perine/Davis being the lead back in two different years. If Scarlett’s stats are removed, Florida ends up sixth in total yards rushing.
Takeaway
It probably sounds like I’m really down on Florida’s running backs. That isn’t the case at all. I think they’re very good players and are going to be an effective unit. I just don’t think that the current statistical profile suggests that they can carry a team that again may be hamstrung by poor QB play.
There are definitely extenuating circumstances to consider as well. You can’t really blame the running backs for their performance against Michigan when they were getting blasted in the backfield by four Michigan defenders on every play. And you can’t blame them for a lack of performance on second-and-long running plays that everyone knew was coming (aka “the Nuss”).
At the same time, the expectation of elite playmakers is that they need to make someone miss from time-to-time. Scarlett’s longest run in 2016 was 46 yards against Iowa. Perine’s longest in 2017 was 29 yards. Lemon’s longest in was 22. Malik Davis had the 74 yard run against Tennessee that resulted in the fumble out of the end zone, but that was an explosive play. I don’t think it’s a coincidence that he was the back who looked the best before his injury.
Maybe I’m wrong. Maybe Florida’s running backs will be an elite group that will lead the Gators to victory. But I just don’t think the statistical record suggests that is the case. Instead, I think it suggests that it is a solid group that is going to be middle-of-the-pack. That’s actually really good considering that nothing about the Gators offense has been middle-of-the-pack for the past three seasons.
But if your expectations are that the running backs are going to carry the offense while Mullen figures the QB situation out, I just don’t think that’s a realistic expectation. Mullen is going to have to generate improvement from the QB position – and significantly so – to see any appreciable offensive improvement.
Otherwise, the Gators offense is going to be stuck in neutral like it has ever since Tim Tebow left Gainesville.
Carl W
A thorough but mildly depressing read. What seems to be the main problem keeping our backs from elite status is breakaway speed. Chubb, Michel, the Bama backs that have dominated so long, Dalvin Cook, they were all threats to score from anywhere on the field, and could outrun DBs. Other than Davis, I don’t see that from our backs, and he lacks the size and power of those mentioned above. Maybe Lemons or Pierce will bring that, and of course Davis might still reach elite status, at least after he recovers from the knee. But until a Gator RB hits a hole 60 yards from the goal line and takes it home, I’ll still be looking for that speed.
The one thing I wonder about: Given that we do have a number of good, maybe very good, RBs, what is the chance Mullen dials up some creative multi-back sets that create mismatches for defenses? Say, Scarlett at TB, Perine at FB with Lemons, Davis or Pierce split out as a receiver and/or motioned through the backfield? With Jones or another running QB behind center, the options seem unlimited.
Joe
Seriously this “it’s not ideal to have a 3* QB “is insane , 4 out of the last 6 heisman winners were guess what 3* QBs (they actually started in highschool ,trask garnered a 3 star ranking while sitting the bench imagine what he would have been rated had he played) and even if you count RG3 and make it the last 7 years that’s 5 QB’s who were either 3 stars or in RG3’s case a very low rated 5 star national ranking :293 so for the better part of a decade it’s been 3 star QBs playing for national titles winning heisman’s and setting records he’ll baker mayfield was a walk-On in fact this decade I believe Jamies Winston and cam newton are the only 5 star heisman winning QBs so of the 10 last winners there’s more 3* QBs on there than any other position/ ranking
Will Miles
It’s actually not insane at all. My colleague Bill Sikes’ research shows that since 2004, the average national championship QB has averaged a 4.2 star rating (0.9526) from 247Sports composite. If you exclude Bama, that increases to 4.5 (0.9707). SEC Champion QBs in the same timeframe have averaged 4.2 stars (0.9411). Playoff QBs since 2014 have averaged 3.7 stars (0.9142) with an average national ranking of 356. Trask was ranked 2123. He may end up a very good QB. But the odds are very much against it, regardless of the success of some of the 3-star QBs you cite. Yes, sometimes 3-star players become very good. The probability of that happening is very small, which is why we make such a big deal out of it when it does happen.
Carl W.
Excellent response. Certainly there’s nothing that says Trask CAN’T turn into an excellent or even championship QB, and I hope he does, but the odds are against it. Every rule has its exception, but the data over time proves out. Mullen might work a miracle, and again, I’m hoping he does, but it will be easier with higher-ranked recruits. In this case the numbers are pretty convincing.
Terry
I believe all running back statistical numbers should be used in context of their offensive line and QB play. Any way you can see the correlation between average yards per rush, how their line grades out, and QB rating or other passing related stat?