NFL Analysis: The Data Is Getting Worse for Proponents of the Running Game

The New Is Getting Bad for NFL RBs

The last eleven years have seen an onslaught of data-based analysis arguing that NFL teams should not draft a running back in the first round nor give them a second contract. At least in part because the advice was so contrarian, it often appeared with sensationalized headlines: ‘Running Backs Don’t Matter’ was a common refrain. The data was correct. The news just got worse for the ground game adherents. New analysis demonstrates that having elite run-stopping interior defensive linemen is directly correlated to losing in the NFL. 

Do Running Backs Matter?

Of course, running backs matter. The statement is hyperbolic by design. The underlying thesis is accurate. The argument rests on asset allocation and the rule changes. If we assume that teams are rational actors, then we can assume they are concerned with maximizing their expected points added (EPA) per play and their play success rate. If so, the path is well lit, draft running backs outside of the first round, and pass on early downs, often.

NFL Asset Allocation

NFL teams operate with artificial constraints to encourage parity. For example, the draft as the primary means to enter into the league, the allocation of draft picks, the draft order, and, of course, the salary cap. Given the contribution towards winning, reliance on other positions for success (offensive line and quarterback), longevity, and the ready availability of adequate replacements, the value of a running back falls outside of the first round. 

Since these arguments have been well-made for a decade, it is unnecessary to engage in detailed proof. Instead, I will give a few select examples to support the point.   

As NFL teams began expanding their analytics departments over the last decade, most teams accepted this sound advice. The following chart indicates the number of running backs drafted by round in the first three rounds of the NFL draft from 1990 to the present. The data is segmented into five-year increments and the most recent three drafts between 2019-2021.

The NFL is arriving at a far better understanding of the actual value of a running back. Still, some teams are stubborn, delusional, or both. When Dallas drafted Zeke in the top five, a colossal mistake, Jerry Jones explained that he understood the value argument but that Zeke was different, special. Of course, the things we just have to have are always the exception to the rule, we tell ourselves. When Gettleman came under fire for drafting Barkley in the top five, he initially became irrationally defensive and slandered analytics. Later he explained he was only joking but that Barkley was different, special.  

Since we have the most data on Zeke, the following chart illustrates the cost of drafting the “different, special” back in the top five of the draft.

Player DraftedRoundRush Avg.TD/Per RushRec Avg.TD/Per Rec.Rookie Contract
Zeke ElliotFirst4.530.728.124.104 yrs, 24.9M
Derrick HenrySecond521.499.125.334 yrs, 5.4 Million
Kenyan DrakeThird425.747.428.174 yrs, 3.4 Million

If you are a Cowboy fan, you probably have a little vomit in the back of your throat. I get it. And, it gets worse. Your team just gave Zeke a six-year, 90 million dollar extension in 2019. Just sit with that number for a minute. Honestly, I am speechless. And the very worst part, I have not even mentioned yet—the opportunity cost. Dallas passed on Jalen Ramsey. They could have drafted Ramsey and Henry or Ramsey and Drake. 

In anticipation of a counterargument, I am not cherry-picking players. Henry and Drake were the following two backs selected. This example is not an outlier. The next year Jacksonville drafted Fournette fourth overall, passing on Mahomes and Watson. In that same draft, Cook and Mixon went in the second round, Kamara, Hunt, and O’Connor went in the third. 

The examples continue until perhaps the most shocking, Josh Jacobs, a first-round running back in 2019. This viewpoint is not the result of positional hindsight. Every analytics person had been shouting this from the rooftops before, during, and after these selections. It is not that running backs do not matter. It is that they are plentiful and the skill differentials between the many good ones are small. Basic supply and demand economics indicates you can get one in later rounds, pay them less, run them into the ground, and then replace them. That’s cold, but cold wins.

Rule Changes Made Passing More Efficient

Rule changes instituted by the NFL over the previous two decades have altered the optimal strategy. Passing is easier now and more efficient. The numbers are unambiguous. Running on first down is a losing proposition. Running on first and second down is purely a masochistic act. As a result, running backs are not as valuable. Here are success rates and EPA for play type by down. 

Play TypeFirst downSecond DownThird Down
Pass Play49% +.073 EPA44.45% +.055 EPA38.57% -.08 EPA
Run Play31.4% + -.09 EPS37.1% – .07 EPA52.55% +.08 EPA

There is an inherent logic to the math. Teams that run on first and second down face longer yardage on third down. Third-down is the most difficult down to throw the ball because everyone is expecting it. Teams that throw on first down face a shorter distance to go on second and third down. Teams running on third down are typically facing short-yardage situations, which increases the success rate and EPA. 

The numbers get far more complex when we account for three-play sequencing. The most optimal play sequencing from a combined EPA and a play success rate are pass-run-run, followed closely by pass-pass-run. The least optimal play-calling sequence is the traditional run-run-pass. The Seattle Seahawks place in the top part of the league in this notorious sequence every year. Yet, the team is generally successful while running the least efficient offensive scheme. This is why you hear people calling for Pete Carroll’s head despite his overall career record of success. We want to see what Russell Wilson can do with an efficient offensive scheme. 

Do Elite, Interior Run Stopping Linemen Matter

Now for the even more terrible news for proponents of the run game, although perhaps it was all ultimately foreseeable. Last year, Brian Burke of ESPN introduced a metric called the “run stop win rate.” Essentially it measures the effectiveness of run stoppers. Then, Josh Hermsmeyer of Nate Silver’s 538 took the metric further. He adjusted the metric for the volume of runs faced, by defensive technique (for example, zero-technique, etc) of players on the interior defensive line. They call the new metric “run stop wins over expected” (RSWOE). 

The results were enlightening. Hermsmeyer concluded that there is no correlation between teams that have elite run stoppers and the points those teams allow in games, or vice-versa. So whether you are excellent or terrible at stopping the run, it does not impact scoring. A slight positive correlation does exist between teams with elite interior run stoppers and their EPA and success rate per play. Of course, that is a bad thing, as defenses do not want positive EPA. They want negative ones. 

Then came the kicker. Opposing coaches run less and pass more against teams with elite interior run stoppers. It seems so simplistically obvious. Since we know from the charts above that it is more efficient to pass, this creates a decidedly negative result for the defense. Thus, having elite interior run stoppers makes it less likely that your team will prevail, by forcing the opposing team to lean on the pass. This logic, if taken to the extreme, would naturally break down. A team with zero run stoppers is not the answer. The analysis demonstrates that the results likely sit on a bell curve. While you do not want a terrible RSWOE, you also do not want an elite RSWOE. 

It is the Goldilocks Theory of interior defensive linemen. You want to be bad enough to seduce the opposing team to run and good enough not to get gouged. You want to be neither too hot nor too cold. You want to be just right. If the analysis continues to produce these insights, we will likely see a steep drop in first-round picks that are primarily interior run stoppers. We will also see a reduction in salary allocation to those players. 

I grew up with John Madden (and pretty much every famous coach) in my ear, telling me successful football required you to run the ball, and stop the run. He might have been right at the time. It turns out now, you do not need to do either.

+ posts