Social media lit up on Sunday night in the moments after the Baltimore Ravens converted a fourth-down run to seal an exhilarating 36-35 victory over the Kansas City Chiefs. Amid the flurry of "WOWs," shock emojis and celebratory GIFs was a tweet from Michael Lopez, the NFL's director of football data and analytics. After watching Ravens coach John Harbaugh make a difficult decision from his team's own 43-yard line, sending quarterback Lamar Jackson up the middle for two yards, Lopez posted the headshot of a man that only his followers and a handful of other devotees to football analytics would recognize.
Pictured was Daniel Stern, the Ravens' football research coach and the man responsible for giving Harbaugh the information he needs to make the kind of choices that are taking over NFL game strategy.
"It was basically a win for analysts everywhere," Lopez said this week. "As a member of the league office, I'm a fan not of one team but all 32. Whether or not they converted makes a difference in the optics, but just the fact that there was enthusiasm about going for it in that situation, that was the biggest thing from my perspective."
Harbaugh's decision provided a visual capstone for a trend that has been building over the past decade. To the great thrill of analytics advocates everywhere, much of the NFL has shifted squarely into four-down thinking this season, when viewed either in raw numbers or through a "non-obvious" filter developed by ESPN sports data scientist Brian Burke called forfeited win probability (more on that later).
There were 88 offensive plays on fourth down through the first 32 games, more than any in league history. Forty-eight of them were passing plays, 32 were rushes and eight were sacks. The distribution among teams is not yet equal -- 10 teams have one or zero such plays -- but the idea has nestled itself securely in mainstream thought. Through two weeks, teams have gone for it on fourth down 20.8% of the time, a high since at least 2000, and offenses are converting 55.1% of those attempts.
"We've been going for it for quite a while around here," Harbaugh said. "We have learned a lot about the numbers and things like that, so we know more than we used to. ... But really, in the end, it boils down to if you think you can get it compared to your chances of not getting it. That's kind of the coaching decision there."
"Hell yeah Coach, let's go for it."
— Baltimore Ravens (@Ravens) September 20, 2021
Back in action Sunday at 1 p.m. on CBS! pic.twitter.com/TSv2TrQ4YL
Indeed, the truth is that two largely independent catalysts have collided to create this season's conditions.
The first is the idea, espoused as early as 1970 by former NFL quarterback Virgil Carter and codified in a 2005 academic paper by economist David Romer, that coaches would benefit from more aggressive offensive playcalling. The development of win probability models in recent years, especially those that account for the time and score in a game, have confirmed that teams often have a better chance of winning by attempting to convert a reasonable fourth down than by punting the ball away.
The second is an intentional and decade-long shift of advantage toward NFL offenses, especially with the passing game. Rules that limit coverage and pass rush have made offenses far more efficient, shifting the priority from field position to possession. Even if Harbaugh had no understanding of the numbers Sunday night, for example, he knew the Ravens had a significant advantage over the Chiefs' defense at the key moment. Conversely, his team would be at an obvious disadvantage if Chiefs quarterback Patrick Mahomes got the ball back -- no matter where on the field he took over.
"It's really changed the underlying fabric of the sport," Burke said. "The stronger offenses get, and the more of an upper hand they have, really what that does is give you a better chance of converting. So maybe some of this shift we've seen doesn't have anything to do with the fourth down analytics itself. It's just an understanding that the sport has changed and offenses have the upper hand anyway. Going for it on fourth down is always lucrative, but the shift toward offense has made it easier to do."
Burke, in fact, has theorized that the evolution of the game -- and not necessarily a commitment to data -- is what drove New England Patriots coach Bill Belichick's seminal 2009 decision to go for it on fourth-and-2 from his own 28-yard line during a 2009 game at the Indianapolis Colts. The play failed, and the Colts needed only four plays after gaining possession to score the go-ahead points in a 35-34 victory. Belichick, who famously insists that he does not pay attention to analytics, absorbed heavy public criticism for the move. But it effectively broke the ice for less secure members of his profession.
"Even though it didn't work," Burke said, "I think it made coaches feel safer. It started that process. If they go for it on fourth down and fail, they're not going to get fired on Monday morning."
Since then, the trend line has moved up steadily. Burke developed and tracks a metric he calls "forfeited win probability," which essentially measures the extent to which a team decreases its chances of winning by making less-optimal decisions. Since the start of 2010, the league has cut its forfeiture rate roughly in half. In fact, during that same period, NFL teams raised their "go rate" on 4th-and-1 from one in every three opportunities to two out of every three opportunities, according to Lopez.
Forfeited win probability has dropped steadily in the NFL as more teams incorporate 4-down offense into their game strategy. Via @bburkeESPN: pic.twitter.com/q2eZWBqWJm
— Kevin Seifert (@SeifertESPN) September 23, 2021
"So within a [short] period," Lopez said, "you're going from a situation where the common thing is to punt to a situation where the common thing is to go for it. From our perspective, it has been a pretty massive change."
The surge has largely been powered by roughly one-third of the teams in the league, estimates Ryan Paganetti, who spent six seasons with the Philadelphia Eagles in a role similar to the one Stern has with the Ravens. In 2021, the list of teams with comprehensive and successful optimization policies include the Indianapolis Colts, Cleveland Browns and Green Bay Packers.
Paganetti was the primary advisor to former Eagles coach Doug Pederson when the "Philly Special" call on fourth down from the goal line helped power the franchise to its first Super Bowl victory after the 2017 season. Every team operates differently, but in the Eagles' case, the majority of the analytics advocacy and nudging coaches toward more optimal decision-making took place in the offseason. Paganetti convened meetings with coaches throughout the staff to explain his analysis and make adjustments where needed. Coaches appreciated knowing how much they could impact win probability -- one way or the other -- with various decisions, Paganetti said.
"For example," he said, "you may have data saying that 4th-and-5 or less from the 42-yard line would be a 'go for it' situation. But the magnitude of the advantage is nowhere near as large on 4th-and 4 and 4th-and-5, so I might communicate that 4th-and-3 or less would be a very clear go for it, while 4th-and-4 or -5 leans toward go for it but is closer to a toss-up."
Those parameters were always mapped out before a game, eliminating debate during the precious seconds available to make a decision. Paganetti, connected to the coaching staff via headset, tried to preview upcoming fourth-down decisions during the course of a possession in time to impact playcalling on third downs as well.
The impact of such acceptance on the game has been almost universally positive, Burke noted, in contrast to the changes imposed by analytics in baseball and, to some extent, basketball. Not only has it provided moments such as Sunday night's game-sealer, but it has also pushed coaches to maximize advantages in the passing game that thrill a majority of fans who prefer elevated scoring and more aerial highlights.
"It's really the two biggest things analytics has done for football," Burke said. "We've shown that passing is much more efficient than running, and then we have this fourth down thing. As opposed to other sports, these analytics really improved the product. It's more fun. In baseball, you could say that analytics have crippled the game. You see strikeouts, walks and home runs. In basketball, it's a three-point shooting contest now. Some people like that, but on net, I don't know that it's a positive."
It's fair to wonder, however, if the search for football optimization could go too far. Could win probability prompt teams to emphasize strategies that are less interesting, even if more efficient? Neither Burke nor Lopez could conceive of an analytical approach that would generate such an outcome, but both said the NFL will need to monitor the impact of its constantly changing rules. Burke pointed to the Big 12 conference at the college level, which currently has six teams averaging at least 39 points per game, as a cautionary tale.
"You see a game there, and if they kick a field goal, it's like they know they are going to lose," Burke said. "You almost need a touchdown on every possession there. If the NFL allows offenses to get too much of an upper hand, then you're looking at arena football, and it takes away from what makes the NFL version of football special."
Too much passing would elevate "a lot of other outcomes that aren't all good for the game," Lopez said.
"You don't have defensive pass interference on run plays," he added. "So your risk of a really big penalty is higher the more you pass. Your risk of stoppages is higher, because every time there is an incomplete pass, the clock stops. You could see longer game times that way. Those types of trends are worth thinking about, but that's more about rules and officiating than analytics."
For the foreseeable future, at least, the NFL seems poised to move into the golden era of analytically-driven game strategy. In a copycat league, four-down offense is one idea worth emulating.
ESPN's Jamison Hensley contributed to this story.