The "establish the run" mentality is finally dead.
Eagles owner Jeffrey Lurie debunked the NFL's most long-established, outdated and illogical myth during a wide-ranging interview with Sheil Kapadia of The Athletic. When asked for his opinion on the "establish the run" wisdom that often comes up in broadcasts, Lurie straight-up rolled his eyes, per Kapadia:
"'What's the right way to say this?' [Lurie] asks himself out loud. "It's just not a truthful way of reporting based on all the information we now have. OK? That's sort of a nice way to say it.'"
Lurie finally said what analytics experts and statisticians have been screaming for decades.
Bob Carroll, John Thorn and Pete Palmer decried "establish the run" logic and its corollaries (The Lions need to give Barry Sanders 25 carries, because they are 15-1 when he gets 25 carries) in their groundbreaking book The Hidden Game of Football way back in 1988. My longtime friend Aaron Schatz founded Football Outsiders and began the modern football analytics movement 15 years ago by debunking run-to-win myths.
More recently, Ben Baldwin of The Athletic studied the topic extensively in a series of articles for Football Outsiders. He found no quantifiable relationship between running and play-action passing; nor between running the ball and controlling time of possession; nor between the amount of rest a defense gets (because the run was "established," keeping it off the field) and defensive performance.
Every scrap of evidence shows that NFL teams run too often and overemphasize the value of the running game. But NFL coaches still cling to "establish the run" rhetoric, even as the leaguewide run-pass rate keeps climbing toward 60 percent passing (it was 58.8 percent last year, with sacks counted as pass attempts but quarterback scrambles still considered runs).
Coaches have passed down their philosophies through the generations, from Amos Alonzo Stagg to today, largely unchanged. Establishing the run made sense in 1955, when players drove ice trucks in the offseason and might run out of steam in the third quarter. It still made sense in 1977, when quarterbacks threw 1.4 interceptions per game (they now throw just 0.8 interceptions per game) and the completion rate was just 51.3 percent (it's now 64.9 percent). It still makes sense in high school, where your teenagers might be bigger and have played less Overwatch during the week than next-town-over's teenagers or where your quarterback might not be able to throw the ball 30 yards without the wind at his back.
But the NFL has been a passing league for decades. And while coaches have also handed down a drill-sergeant patois about trench warfare, plus a tough-guy suspicion of bean-counters with calculators, Lurie is an owner with a Super Bowl ring and one of the NFL's more progressive thinkers. That may be why he felt comfortable admitting that he's at least dabbled with the dreaded four-letter m-word: math.
So that settles things, right? Running backs don't matter, coaches who emphasize the run are stupid and anyone who believes that the physicality of the running game still has value is an anti-intellectual troglodyte. The analytics community has been right all along.
Real research isn't so simple.
True, the research shows that pounding the ball up the middle early in a game just for the sake of doing so—think of last season's Seahawks loss to the Cowboys in the Wild Card Game here—is counterproductive in every measurable way. It has also shown for decades that saying Team X has a great record when they run the ball 30 times is a mixup of cause and effect; teams usually run 30 times because they have a commanding lead and are munching on the clock; the 30 rushes don't usually get them that lead.
We've learned other things about the value of running backs (the margin between an All-Pro and the average rookie is surprisingly small) and usage patterns (too many carries are bad), and those discoveries, among others in every facet of the game, have been shaping draft, contract and strategic decisions around the NFL for years.
But there is no research that proves that the running game is somehow useless or that running backs are as replaceable as the pens we all steal from bank lobbies. Those are talking points often associated with "analytics" when they are really broad caricatures of analytical concepts that are always being reappraised and refined.
What would happen if teams just threw the ball 90 percent of the time? No one knows, because all the analytical models we have are based on an NFL in which teams pass 50 to 60 percent of the time. Mathematical models are unreliable at the extremes, and everyone who understands modeling or has created an overpowered video game character and "broken" a sophisticated simulation or listened to Jeff Goldblum's dialogue in Jurassic Park knows that.
Here's a simple example that won't make your eyes glaze over. Teams averaged 4.4 yards per rush last season—the highest per-carry rate in NFL history, in the most pass-happy season in NFL history! Last year's offensive explosion had almost as much to do with the creative, high-efficiency rushing attacks of teams like the Rams and Chiefs as with the passing game.
The running game may actually grow more marginally valuable as passing rates increase because of the element of surprise, the proliferation of dime defenses full of 195-pounders (see last season's divisional-round matchup between the Patriots and Chargers), quarterback runs, the elimination of ineffective two-yard plunges into stacked defenses and other imaginative options. And yes, letting the right tackle fire off the line and clobber his opponent now and then probably has some value that doesn't show up on a spreadsheet.
Or maybe not. We need more data and more research. One goal of analytics should be determining whether some optimal play-calling balance exists, finding it and exploiting it (or disrupting it on the defensive side). Coaches have been doing that by trial and error for decades, despite their establish-the-run mantra. The best analytical minds are exploring that topic, not taking victory laps on Twitter.
Decision-makers like Lurie, meanwhile, are becoming increasingly savvy to the data and more skeptical of the slobberknocker conventional wisdom. Not coincidentally, NFL scoring and excitement are at an all-time high, and the game on the field grows more interesting and innovative every year.
So let's get an R.I.P. in the comment thread for establishing the run. And here's some old-timey coaching advice for the analytics community: You celebrated the win yesterday. Tomorrow, it's back to work.
Mike Tanier covers the NFL for Bleacher Report and has been a co-author of the Pro Football Prospectus and Football Outsiders Almanac series of annuals since 2005. Follow him on Twitter: @MikeTanier.