clock menu more-arrow no yes mobile

Filed under:

The Drive: In defense of running all the damn time

NFL: SEP 17 Seahawks at Bears Photo by Robin Alam/Icon Sportswire via Getty Images

If Chuck had his wish, it’d be three yards and a cloud of dust and he’d win, 7-6.” —Steve Largent on Chuck Knox

Bill Walsh did it in 1981. Or maybe Don Coryell did it in 1974. Could be that Mel Blount did it indirectly in 1977. Or maybe it was Steve Largent who did it in 1978. Great change has many progenitors. Whoever did it, and surely each had a part, for longer than I’ve lived the NFL has been a passing league. I have never known an NFL in which pass-centric strategies are not praised and rush-centric strategies are not scorned. Yet for a very long time this popular opinion made little impression on NFL coaches.

But, like me, fewer NFL coaches than ever can remember in any meaningful way a league that wasn’t ruled by passing the ball. Nickel is now a kissing-cousin of so-called base formations. Ezekiel Elliott and Melvin Gordon have followed in the footsteps of Le’Veon Bell and opted to go on strike rather than accept their reduced value and commensurate reduced pay. That change which was seen as overdue ten or twenty years ago is now orthodoxy, and that orthodoxy is often said to be not changed enough.

In the words of writer Josh Hershmeyer:

That’s where the NFL is currently living. The NFL is a passing league that somehow doesn’t pass enough. NFL teams know the medicine works yet stubbornly refuse to take a clinically effective dose.

Very much apart from that orthodoxy is Seahawks offensive coordinator Brian Schottenheimer. Seattle was the only team in the NFL to rush more often than it passed. It worked, insomuch that Seattle’s offense was above average in its efficiency, and the Seahawks made the playoffs. And it didn’t work, insomuch that despite having the sixth most efficient passing offense and sixth most efficient running offense, Seattle finished with only the ninth most efficient overall offense, because, one could argue, it ran too damn much.

The Seahawks finished in a dead-heat with the Falcons for overall offensive efficiency. Both finished the season averaging 8.8% DVOA per play. Atlanta though performed much more poorly rushing the ball, -10.7% (22nd) compared to Seattle’s 4.3%, and also a bit less efficiently passing the ball, 26.7% (7th) to Seattle’s 27.2%. Yet, by whatever fractional margin, the Falcons were the more efficient overall offense. Three of the eight teams which finished ahead of Seattle had poorer combined run and pass efficiencies, but Atlanta, Green Bay and Pittsburgh all made the (one might call common sense) decision to play-call more passes. Thus they did not best the Seahawks through better execution but through better strategy.

DVOA can be a bit abstruse. What that percentage means and how that percentage was arrived at is often not very clear. A relatively simpler way to think of this is, NFL teams on average added 78.1 points through passing the ball and lost 3.0 points rushing the ball in 2018. Only 12 teams added any value running in 2018, and those teams added only from 1.30 to 65.94 points. All but six teams added value by passing and half of the league added more value passing the ball than the best rushing team (the LA Rams) added through rushing. The very best passing team, Kansas City, added 266.71 points through passing the ball. Schottenheimer, it would seem, shot his feet all to bits in 2018.

Even the assumption that highly valuable play-action passes are contingent on successfully running the ball has been, in the words of Steven Ruiz “debunked by stat nerds.” And in the arms race of ideas, no one wants to be the metaphorical equivalent of the Polish cavalry charging German tanks. But, as it turns out, that metaphor is founded on a myth.

The value of passing the ball in the modern NFL is not a myth. Nor do I intend to patch together a few factoids in the face of overwhelming evidence. I do, however, wish to offer a counterargument, however strained it may seem to be. Not because I wish to defend Schottenheimer or Pete Carroll. And not because I necessarily believe that the NFL is not a passing league. But because, fundamentally, I do not believe much and prefer to remain curious and undecided in the process of discovery. It is simply much more fun to attempt disagreement, and, I think, potentially much more valuable to bet against prevailing trends. After all, no one ever got rich buying Google stock at $1,200.

Argument 1: Passing endangers the health of Russell Wilson

Per an old and perhaps outdated post on Slate:

“It turns out that the only gameplay variable that explains injuries with any statistical significance is sacks. On average, a 1 percent increase in sack share—the percentage of plays called for the QB that end in a sack—is associated with a 2.6 percent rise in starts missed due to injury (0.7 percent standard error). This link holds when we use the career-wise dataset and when we use sacks per start instead of sack share.”

Wilson has averaged a sack taken on 8.4% of all pass attempts. Last year, that percentage ballooned to a career worst 10.7%. We might wish to believe that Wilson is special and therefore not prone to larger statistical trends, and in some ways he might be, but when it comes to taking sacks, I doubt he is any less injury prone than most quarterbacks. 2016, what could have been a prime season for contention by the Seahawks, was more or less undone in week one when Ndamukong Suh stepped on Wilson’s heel badly turning his ankle.

Runs do not endanger Wilson’s health. Even runs by the quarterback—perhaps because the quarterback can anticipate and protect himself from violent contact—have not be shown to greatly endanger the quarterback.

Argument 2: Scrambling quarterbacks seem disproportionately negatively affected by catastrophic injury

I am defining “catastrophic injury” as an injury that requires major surgery and lands the quarterback on injured reserve. I am using ESPN’s Total Quarterback Rating because it factors in the value of rushing and sacks taken. What defines a “scrambling quarterback” is debatable. Feel free to disagree with any of my selections.

My group of pocket quarterbacks who suffered catastrophic injury: Tom Brady, Joe Flacco, Peyton Manning and Matthew Stafford

Scrambling quarterbacks: Robert Griffin III, Colin Kaepernick, Donovan McNabb, Deshaun Watson

There’s no way to do this perfectly, but for the sake of argument.

Tom Brady QBR in last full season before injury: 88.5

Average of next three seasons following injury: 75.6

Average number of games played in those next three seasons: 16

Flacco pre: 68.2

Post: 53.7

Games: 13.7

Manning pre: 78.4

Post: 77.5

Games: 16

Stafford pre: 50.5

Post: 58.0

Games: 16

Griffin pre: 69.4

Post: 38.8

Games: 9

Kaepernick pre: 66.1

Post: 46.5

Games: 10.5


Pre: 65.5

Post: 55.4

Games: 14.7


Pre: 83.6

Post: 63.0

Games: 16 (in one season)

I am not going to try and sell you on the irrefutable truth of this method or its findings. Gridiron football is full of things which do not make for good subjects for statistical analysis. But pocket quarterbacks depend most on their decision making. Perhaps this can be injured through concussion. Perhaps previously good decisions can become bad as talent is diminished. But, overall, pocket quarterbacks, the great like Brady and Manning, and the so-so like Flacco and Stafford, do not depend on great athleticism because they do not have great athleticism.

Scrambling quarterbacks do, and when that athleticism is severely diminished due to injury or imperfect recovery, they do not reinvent their game to compensate—or do not effectively. Russell Wilson played the 2016 season injured, and it was his least effective season. His total quarterback rating was 56.8, the worst of his career and the only time he has finished in the 50s. And a year after finishing with the most efficient offense in the NFL, Wilson and the Seahawks finished 16th.

Wilson is not clay to be shaped. He is a person with incredible talent but limitations, and so long as he builds his game around incredible displays of athleticism, keeping him healthy and thus retaining that athleticism is paramount. Running the ball forwards that goal.

Argument 3: Running has rebounded in value

A cumulative value of -3.0 is hardly inspiring but it is a marked improvement from seasons past. Starting with 2017 and counting backward, the average value per team of all runs were: -26.6, -20.1, -30.5, -26.1, -23.0, -19.9, -19.5 and -27.8 rounds out the decade. The NFL is adjusting to its previously mistaken estimate of the value of running the ball, and as pass rushers and pass blocking are emphasized, and passing formations and nickel and dime defensive backs are played more often, running is becoming more valuable. That’s not a reversal, but it might be a first data point in a trend, and should somehow the league flip or at least approach something closer to equality, teams which invested in running the ball will be way out in front.

Argument 4: Rushing is more weather-proof and thus more valuable in the playoffs

Only three teams in the NFL play in fixed dome stadiums. Which means there’s a pretty good chance that weather will be a factor in a majority of playoff games. Running the ball, as you can imagine, is much less negatively affected by wind, cold and precipitation. A good run game is especially important when the outcome of a game is especially important.

Argument 5: Rushing may indirectly help a team’s defense

I cannot find any data to support this, but this is a long held belief. It may be that building a good rushing offense helps to improve a defense through rest. Rushing is typically thought to be less effective at creating explosive plays but more effective at sustaining drives. I am not sure statistical analysis bears that out, but more on that in a second. It may be that something about fielding a good rushing offense improves a defense through practice. Ultimately, I do not know, but there is strong anecdotal evidence that running effectively somehow indirectly improves the performance of a team’s defense.

Argument 6: Things Change

If you’ve read older literature, you have perhaps encountered an idea that is called the Malthusian Catastrophe. In short, it is a belief that population growth will one day outstrip the earth’s ability to produce food. One estimate I remember reading is that the world could support no more than 1.5 billion people. Excuse my poor memory if that’s inaccurate, but my point is, Thomas Malthus could show how much food people could produce in his own era but he could not anticipate innovations in farming.

This is pretty typical of research. It is very often extremely accurate at reading the past but hopelessly inaccurate at anticipating the future. In fact, one thing which often greatly affects the future is keen awareness of the nature of the recent past. I am willing to concede that running has been inefficient, that play action passing has not recently depended on the effectiveness of the passing team’s run game, and that for a very long time coaches have stubbornly run too much. But much harder to know is whether running will stay inefficient as teams adjust, whether play action will remain uncoupled from the effectiveness of the run game, and how efficient passing and running will be as prevailing strategies change.

One change which is still in its infancy is NFL coach’s increased inclination to go for it on fourth down. When Bob Carroll, Pete Palmer and John Thorn wrote The Hidden Game of Football, they determined that for a run play to be successful, ie for a run play to improve a team’s chances of scoring, it had to achieve four yards on first and ten, 50% of the yards necessary to convert the first on second down, and 100% on third down. But, logically, a running back who could rush for 2.5 yards every down would be unstoppable. And so those numbers are not absolute. They represented the nature of the league at the time of the study.

As it becomes more common to go for it on fourth down, I imagine those numbers will need to be revised. And as teams think in terms of four rather than three downs, consistent incremental gains toward converting the first may become more valuable.

Gridiron football is wonderfully capable of reinventing itself and how it’s played. Most teams, I imagine, are aware that statistical analysis suggests they are overplaying the run against play action. What we do not know is what happens if they adjust. How fine a line can defenses walk? How much must change before opposing defenses have overcompensated?

Gridiron football is also wonderfully impervious to the scrutiny of statisticians. Its base measurement is yards, which is composed of three feet. Every data input taken from play-by-play is fuzzy, at best. If all things were measured in feet and not yards, would ten feet still be considered too little gained by a run on first and 30 feet? Perhaps rushing 2.6 yards, which may be recorded as rushing three yards, is falsely creating the perception that rushing for 3.4 yards is lacking in value. Sample sizes in football are minuscule, and smaller still when free of hasty generalizations. Twenty years ago, the brightest thinkers in baseball extolled to all who would hear that only winning drives attendance. The modern game, less indebted than ever to rules of aesthetics, is all but unwatchable, and attendance seems to be in terminal decline. No NFL team can ever achieve anything like Moneyball, and thank goodness for it.

Ultimately I have seen far too many passing-game gurus crash and burn, and too many old school thinkers excel, to believe statisticians are offering a truly holistic analysis of football. But we are surely in a time when statisticians have the ears of owners, management and to a lesser extent coaches. If, very possibly, they are dealing bunk, and some most definitely are, 2019 might be the exact moment to reject their ideas. It is a timeless pattern. New ideas fail. Old ideas, ameliorated by new ideas, become new again, effective again.

Jalen Rose said something very interesting in his interview with the New Yorker. The piece is a bit rambling and so I’ll paraphrase. Athletes in a sport have the intuitive knowledge of thousands of hours of direct experience. They cannot always explain what they know. They almost certainly cannot prove it statistically. But they intuit good methods and good strategies. Offensive linemen, they like to run the ball. So do quarterbacks and coaches. Defensive coaches, defenders, they prize their ability to stop the run. A lot of coaches may think that’s trivial—players’ intuition. But intuition is a magical thing capable of profound, often mysterious insight. Maybe the efficiency experts are playing chess and the rest of us, Pete Carroll, Brian Schottenheimer, Wilson, Fluker, Carson, we’re all playing football.