Is the scramble the best play in football? How running QBs have revolutionized the NFL

How running QBs have changed the NFL: Dual-threat history, valueplay

Why Stephen A. was disappointed in Lamar despite numbers (1:52)

Stephen A. Smith explains why even though Lamar Jackson put up great numbers, he’s still disappointed in how the quarterback played in the big moments. (1:52)

Bill Barnwell, ESPN Staff WriterSep 12, 2024, 06:30 AM ET

Close

Bill Barnwell is a senior NFL writer for ESPN.com. He analyzes football on and off the field like no one else on the planet, writing about in-season X’s and O’s, offseason transactions and so much more.

He is the host of the Bill Barnwell Show podcast, with episodes released once a week. Barnwell joined ESPN in 2011 as a staff writer at Grantland. Follow him on Twitter here: @billbarnwell.

Let’s say you were a lapsed fan who hadn’t watched professional football in, oh, about 40 years. If you tuned in to an NFL game for the first time in decades in Week 1, what would have stood out as different? I’m guessing some of the technological advancements would have been the first thing you noticed. The first-down marker is helpful. The game looks better in high definition, sure. The absence of neck rolls might have seemed strange. Let’s look past the visual side of things, though.

What about the game itself? What would have seemed most different from the past? There’s something unique that has become commonplace in the current NFL, and it’s a development that has fundamentally changed the sport. The running quarterback — seen alternately as a fad, a problem to be corrected or simply not seen at all for long stretches of time — has become a firmly entrenched entity.

Subscribe: ‘The Bill Barnwell Show’

Now, I know this isn’t exactly new news. I’m not arguing quarterbacks such as Lamar Jackson and Josh Allen have just arrived, because they’ve been in the league since 2018. What’s fascinating and meaningful about this shift right now is how widespread and seemingly permanent it has become. For most of the past half-century of football, the running quarterback was limited to one player per generation, an exception to the rule of pocket passers. They were almost regarded as gimmicks. Many players who might have qualified to play under center were moved to other positions before they could ever prove they were capable or incapable of serving as an NFL quarterback, with Jackson as a recent example of a prospect whom some observers wanted to move to another position.

That’s no longer the case for top prospects who have excelled at the college level. The number of quarterbacks who are essential components in their team’s running game has grown and continues to rise. And the impact of those players is changing the game.

Jump to a section:

The staggering shift, and why 2018 is notable

The impact of designed quarterback runs

Is the scramble the best play in football?

The proliferation of running quarterbacksLet’s put this into context, because the numbers are jarring. There are several ways to define a running quarterback, but let’s use round numbers. One simple definition of a QB who both throws the football and runs more often than the occasional scramble, sneak or kneel-down would be a passer with 150 passing attempts and 100 rushing attempts in a season. That’s easier to do in a 17-game season than it was in the 14-game era, but we’re looking at a nominal passing workload and just over seven carries per game across a full year in the 14-game seasons to qualify as a runner by this list.

From 1950 to 1988, there was exactly one instance of a player throwing the ball 150 times and running 100 times in the same season. It was Bears quarterback Bobby Douglass in 1972. He spent about 2½ seasons as the starter in Chicago and came within 38 yards of becoming the first QB in league history to rack up 1,000 rushing yards in a single season. The league saw Lou Holtz attempt to implement a version of the option during his brief stint as Jets coach a few years later, but otherwise, the NFL simply did not employ quarterbacks who ran the ball regularly. They were more likely to punt than become part of the run game, even in an era in which offenses were far more heavily weighted toward the run.

Then Randall Cunningham came around and pulled off 150/100 in back-to-back seasons with the 1989 and 1990 Eagles. This opened up the door for individual exceptions to the league’s de-facto rules, although the only player who had a qualifying season over the next decade was Steve McNair in 1997. Daunte Culpepper did it with Minnesota in 2002, but the next quarterback to pull it off regularly was Michael Vick, who accomplished it four times for Atlanta between 2002 and 2006 and then once more with Philadelphia in 2010.

You could argue the tide shifted in 2011, when Tim Tebow and Cam Newton were 150/100 guys, or the following year, when Robert Griffin and Colin Kaepernick made playoff runs in offenses incorporating zone-read concepts from the college level. (Kaepernick didn’t qualify that season, although he made it in 2014.) I’m not sure they stuck the landing, though; Tebow was quickly out of the league, Griffin wasn’t able to adjust after tearing his ACL, and Kaepernick wasn’t as effective once Jim Harbaugh and Greg Roman were out of the picture. Newton was the one quarterback who was heavily involved in his team’s running game who made it through multiple successful seasons on his second contract, although the hits he took might have shortened his career.

To put all that in context, over a 65-year span between 1950 and 2015, just 11 quarterbacks recorded a 150/100 season. Just four — Newton, Vick, Cunningham and Russell Wilson — did it more than once.

Over the past five years, seven different quarterbacks have racked up 150/100 seasons. They’ve been aided by a 17th game, but even upping the threshold to 150 passes and 125 rush attempts, five have reached this feat.

And right now, as we begin the 2024 season, a full quarter of the league’s starting passers are involved in their team’s run game. In the AFC, there is Josh Allen, Lamar Jackson, Justin Fields and Anthony Richardson. The NFC has Jalen Hurts, Jayden Daniels, Daniel Jones and Kyler Murray. There are others, such as Patrick Mahomes and Dak Prescott, who have success as scramblers, but the eight just mentioned are involved in their team’s designed run game.

More than half the league is going to take at least some snaps on designed quarterback runs. Go back to 2009, and there were 225 designed runs called for the league’s 32 quarterbacks, according to ESPN’s tracking data. Those runs don’t include scrambles or sneaks, although a few quasi-sneaks inevitably slip through into the research. By last season, that number had more than doubled to 590 designed runs. There were 37 in Week 1, which would put the league on pace for a record-high 629 this season.

As it turns out, the draft that really shifted things was 2018, when five passers were selected in the first round. Sam Darnold and Josh Rosen, the two most traditional, pro-ready quarterbacks from that class, haven’t been successful pros. Baker Mayfield, the No. 1 pick, has mixed solid and frustrating campaigns and settled in as a low-end starter.

The two most successful quarterbacks from that class are the guys who have derived significant value from their impact on the ground. Jackson, a two-time MVP, became the successor to Kaepernick and Tyrod Taylor in Greg Roman’s offense — Roman was in San Francisco and Buffalo before he came to Baltimore — and took the league by storm as a dual threat. He was the league’s best player last season. And while Allen can make any throw, the element that sustained him during his early-career struggles — and the factor the Buffalo offense repeatedly falls back on when they need a play — is his ability to move the chains with his legs.

Since then, Murray, Jones, Richardson, Daniels and Trey Lance have been drafted in the top 10 as players who were all expected to make significant impacts with their legs. They’ve stepped into offenses and immediately been used as runners in the same way they were in college without a second thought. That simply wasn’t happening on a widespread basis without being questioned or dismissed as a fad before the breakthroughs of Jackson and Allen and their roles in their respective teams’ run games.

The impact of designed quarterback runsThere isn’t advanced metrics or play-by-play data stretching back deep into the past, which limits the comparisons to make here. The earliest expected points added (EPA) data ESPN has dates to the 2007 season, which is fine for our purposes, since there weren’t many quarterbacks featured in the run game. Vince Young led the NFL that season with 20 designed runs, and the players behind him were Jay Cutler and David Garrard, whose styles wouldn’t be aligned with that of running quarterbacks in the modern NFL.

Rushing EPA per play hasn’t changed much over that stretch. Even as the league has moved heavily toward the pass after the influence of the 2007 Patriots shifted teams’ openness toward the pass and spread offenses, rushing games averaged minus-0.03 EPA per attempt in 2007 and minus-0.04 EPA in 2023. Those numbers don’t include sneaks, scrambles or kneel-downs.

Even as their role in rushing attacks has grown, though, quarterbacks have been markedly more efficient than full-time running backs on these carries. Running backs generated minus-0.03 EPA per carry in 2007 and minus-0.05 EPA in 2023. Quarterbacks, even with their usage rate doubling, generated 0.16 EPA per rush last season. Their success rate on those runs was nearly 55%, well above the 43% success rate for halfbacks.

The place in which they’ve made an obvious impact is near the goal line, where the quarterback run game becomes more and more ubiquitous with each passing season. Quarterbacks were responsible for just over 10% of the league’s rushing touchdowns in 2007. That mark was up over 24% last season.

Quarterbacks took 10% of the total carries inside the 5-yard line in 2007, many of which were on sneaks and scrambles. That figure rose to 20% last season. Despite their usage rate doubling, quarterbacks scored on an identical rate of their rush attempts, ending their runs with celebrations in the end zone 47.7% of the time in 2007 and 47.5% in 2023.

When scrambles and sneaks are removed? Again, usage rate for quarterbacks has about doubled. Running backs and quarterbacks have grown more efficient at converting their carries inside the 5-yard line into scores, but without sneaks and scrambles, backs converted 39.6% of their touches into scores last season. Quarterbacks were at 60%.

It shouldn’t be a surprise that running backs have grown more efficient over that time frame, and it’s in part because of the rise of running quarterbacks. The evidence we’ve seen suggests playing next to a running quarterback is likely to increase a running back’s efficiency and effectiveness. Anecdotally, we’ve seen Mark Ingram and James Conner improve after joining teams with mobile quarterbacks, while Miles Sanders had his efficiency crater after leaving the Eagles and Hurts in 2023.

There aren’t enough data points on a player-by-player basis to draw meaningful conclusions, but the on-field evidence suggests playing next to a quarterback who is a threat in the run game should make a running back’s life easier. The threat of a quarterback keeping the football on zone-read or designed quarterback run plays can cause defenders to slow down and second-guess whether they want to attack the running back.

While defenders aren’t flummoxed by the concept of the zone-read the way they were in 2011 — in part because every defender has seen those concepts in high school and college football before making it to the league — the uncertainty about the quarterback run game is still enough to create unorthodox problems for opposing rush defenses.

Adding the quarterback to the run game changes the numbers for teams on the defensive side. In the past, when they were exclusively handing the ball off and staying out of the run game altogether, defenses could dictate the numbers game. With two wide receivers and two cornerbacks aside, defensive coordinators could put seven players in the box (four defensive linemen, two linebackers and a third linebacker or nickel defensive back) against seven offensive players (a running back, a tight end and five offensive linemen). If the defense wanted to stop the run, it could move one of its two deep safeties into the box and play eight against seven.

If the quarterback is in the picture, though, that changes. Now, if defenses play a team with a running quarterback, they’re outnumbered if they have two deep safeties and need to have a safety in the box if they want to be even with an opposing offense. And as you might have heard, given how quarterbacks such as Allen and Hurts are throwing the football, defenses don’t want to live in a single-high world.

As a result, that problem of how to stay in a two-high shell while keeping the numbers even in the box has led to a few solutions, the most notable of which has been the proliferation of Vic Fangio-style defenses, although that might have peaked in 2023. One of the key components of Fangio’s fronts has been asking defenders to play what would be considered “gap and a half” responsibilities up front, where they attack one gap and then remain responsible for a secondary gap nearby. The goal is to try to defend as many gaps as possible without adding extra bodies to the box, creating something like an illusion of a full box long enough for the safeties to fly in and help make plays while also being able to stay deep in case of pass plays.

Naturally, as the league has also realized the advantages of play-action, the threat of the running quarterback has only made the threat of a play fake even more lethal. As an example, just about every team runs the boot and flood concepts that see a signal-caller fake a handoff to one side and then run to the other side, where receivers at short, intermediate and deep levels follow him to stretch the defense horizontally and vertically.

The boot concept almost always calls for the weakside defensive end to be left unblocked, which means a pass rusher will be free to attack the quarterback as he turns back around and runs his way. Less mobile passers can survive on these concepts by gaining the appropriate depth and reading the field quickly, but they’re always more likely to take sacks and be forced into throwaways by their lack of agility and quickness.

Mobile quarterbacks can simply outrun that free defender, and that creates more opportunities. They can extend those plays further, which allows more time for the deep concepts to open. They can scramble off those concepts into open space, which creates easy yards and/or forces the shallow defender to honor the threat of the scramble, opening space behind for receivers. The threat of all that in the passing game has led teams to send that defensive end directly at the quarterback and totally abandon the threat of the running back getting the football in recent years, which in itself allows backs to be more efficient, since there’s less of a threat they’ll be brought down when they cut back.

Theoretically, the threat of mobile quarterbacks should also have an impact on what coverages end up being employed in pass situations. If teams believe they need to match the quarterback in the run game by getting an eighth defender in the box, they’re prevented from playing split-field coverages and have to play single-high looks like Cover 1 and Cover 3. Modern NFL defenses try to disguise coverages and mess with box counts by moving safeties and changing looks on quarterbacks after the snap, but if they’re getting gashed on the ground, they’re going to get more bodies near the line of scrimmage.

That’s a huge advantage for mobile quarterbacks! From 2018 to 2023, the three who faced 2-man coverage (man-to-man coverage with two deep safeties) least often are three mobile players: Jackson, Newton and Murray. Jackson has faced Cover 2 over that run at the lowest rate in the league. Jackson, Newton and Murray faced Cover 3 and their single-high looks, meanwhile, at three of the five highest rates over that stretch. If a quarterback’s style of play can help dictate what coverages he sees, he’s reducing the complexity of what he’s facing on a week-to-week basis and can make more informed decisions in the passing game.

There’s another more widespread way quarterbacks make an impact with their feet, and that has grown into a hugely valuable element of modern NFL offenses.

Is the scramble the best play in football?It might be! The average NFL scramble last season generated 0.31 EPA. That’s significantly better than the typical non-scramble run play (minus-0.07 EPA) and even better than the average pass (0.13). In terms of yardage, the average scramble even generated slightly more yards (7.2) than pass plays (7.0). Should teams toss their run and pass plays out of the playbook and build the entire offense out of scrambles?

Well, not really. There’s serious selection bias built into that data. Scrambles by definition are going to gain positive yardage because they all pass the line of scrimmage. A run that loses yardage is still a rush. A scramble that doesn’t make it back to the line of scrimmage is marked as a sack.

It’s possible to work with the data to suss out some of those failed scrambles. What if failed scrambles were treated as sacks, where the quarterback made it out of the pocket and the sack took place after four seconds? This isn’t a perfect definition, as some scrambles will produce sacks while the quarterback steps up inside the pocket and some could take place on designed boot concepts where a quarterback isn’t trying to scramble, but this is a reasonable proxy for failed scrambles.

When folding that in and including all the other sacks that aren’t considered as failed scrambles as part of the passing performance, scrambles still rate as the best play in football, albeit nowhere near as drastically as they do by our first look. I’d still argue there’s selectiveness built into those numbers. Consider that a quarterback such as Mahomes often will scramble only if he feels confident he can get a first down.

I’m not sure the scramble is the single best play in football, but I’d certainly argue it’s a valuable play and a hidden source of value for quarterbacks who create first downs with their legs. In 2023, quarterbacks ran for nearly 7,400 yards and 398 first downs on scrambles. That’s about double what they produced in 2007, when they ran for 3,614 yards and 195 first downs. Efficiency on those scrambles isn’t up dramatically — they averaged 6.8 yards per scramble and a 36.6% first down rate in 2007, going up to 7.2 yards with a 38.7% rate last season — but the influx of rushing signal-callers has led to them scrambling far more often.

Traditional passing metrics aren’t incorporating the impact of those scrambles, which are nominally pass plays that end up turning into runs. ESPN’s Total QBR includes them, along with the impact on designed runs, fumbles and sacks taken by each quarterback.

To put that in context, consider a simple stat we’ve used to judge quarterbacks for a long time: yards per attempt (Y/PA). It’s simply passing yards divided by pass attempts, with no inclusions for scramble yards or sacks. What about incorporating yards gained on scrambles, yards lost on (all) sacks and changing the denominator from attempts to dropbacks to account for the added plays? This is similar to what net yards per attempt accomplishes, but we’re adding scrambles to the mix as well.

When ranking last season’s quarterbacks by yards per attempt and then comparing them to our metric with scrambles and sacks included, most of the changes align with the general consensus about these various quarterbacks relative to the rest of the league:

Mahomes, who averaged 8.1 yards per attempt in his career before dropping to 7.0 yards last season, jumps from 17th in strict yards per attempt to 10th with scrambles and sacks included. That shouldn’t be a surprise; his secret superpower is avoiding sacks, and he scrambled for 406 yards last season, second in the NFL behind Jackson.

Allen jumps from ninth to fifth. His ability to extend plays without taking sacks is spectacular; he posted the league’s best sack rate and averaged nearly eight yards per scramble, with his 370 scramble yards ranking third. Murray also improved from 22nd to 18th.

On the flip side, a metric like this doesn’t favor all rushing quarterbacks. Justin Fields, who was traded by Chicago to Pittsburgh in the offseason, ranked fourth in scramble yards and was among the league leaders in averaging 8.4 yards per scramble attempt, but taking sacks more than 10% of the time drops him from 18th in traditional yards per attempt to 22nd in our modified metric.

Dak Prescott drops from fifth to ninth. The Cowboys star took sacks at an average rate, but he hasn’t scrambled much since returning from his 2020 ankle injury. His 28 scrambles generated 193 scramble yards, which ranked 14th. Desmond Ridder, Sam Howell, Baker Mayfield and Zach Wilson all fell by at least three spots.

Expanding the data to the past five full seasons and quarterbacks who have at least 1,000 dropbacks over that stretch, including scrambles and sacks aligns with our perceptions of quarterbacks. Mahomes jumps from eighth in raw yards per attempt to first in the adjusted metric. Allen goes from 19th to 10th, which still underrates him because of his impact on designed runs. Tom Brady, who avoided sacks at an elite rate into his 40s, jumps from 24th to 17th. And on the flip side, top-10 quarterbacks by yards per attempt, including Deshaun Watson, Ryan Tannehill and Russell Wilson, all drop by nine spots or more.

This metric isn’t anything extraordinary. Stats such as Total QBR use a better framework (EPA) and do a better job of contextualizing the impact of each of these plays on performance. Yards per attempt wasn’t a perfect metric from the jump. What I’m hoping sinks in here, though, is there’s meaningful value being generated by quarterbacks with scrambles (and by avoiding sacks) that isn’t being captured through traditional statistics.

Scrambles only amount to about 3% of all offensive plays, up from 1.7% in 2007, but the combination of quarterbacks in the designed rushing attack and working on scrambles is making a bigger impact on the broader game.

In Week 1, with those eight running quarterbacks in starting lineups around the league, signal-callers ran for 787 yards. That’s the third-highest total for any week since at least 2007, which is as far back as ESPN’s data goes. Given the paucity of run-friendly quarterbacks before that time, it’s likely the third-most rushing yards produced by quarterbacks in any given week in NFL history. That came in a week in which no passer had an extraordinary game on the ground by their standards. (The current single-week record, 801 yards, came during Week 9 in 2022, when Fields led the way by running for 178.) It would hardly be a surprise if this record were reset at least once this season.

The rushing volume might already be shifting the game. Just two quarterbacks threw for 300 in Week 1. Teams were running at a higher rate than they have in recent years, a product of several factors, including defenses continuing to shift to lighter boxes and the wider spread of run-friendly passers. The runs also produced shorter games; while the average Week 1 contest between 2019 and 2023 saw each team run 64.2 offensive snaps per contest, teams in Week 1 ran 60.2 snaps. That might not feel like a lot, but over the course of a full season, a four-play difference would be the equivalent of missing a full game’s worth of snaps.

As I wrote in November in my look at running back contracts and why backs were struggling to get paid, the shift toward running quarterbacks has impacted the value of players at other positions. Since quarterbacks are scoring a higher percentage of rushing touchdowns than they did in decades past, it’s difficult for backs to rack up the gaudy touchdown totals peak players would have, limiting their cumulative numbers and their chances of winning awards.

Teams with running quarterbacks might also feel more comfortable building the offense around that signal-caller and trusting that his gravity might allow their running game to remain efficient with a wider range of potential running backs, which would make those backs less valuable and more replaceable. At the same time, we’ve just seen the Eagles (Saquon Barkley) and Ravens (Derrick Henry) invest in veteran backs in free agency.

What feels true, at least in the near future, is it doesn’t feel like the league is about to head back in the other direction. Vick and Cunningham were anomalies in their respective generations. Griffin and Kaepernick were regarded, at least by some portion of the football universe, as gimmicks with a short shelf life.

Now, most high-level high school and college games on a weekly basis feature quarterbacks who are mobile and play meaningful roles as scramblers. Many offenses have some semblance of a quarterback run game built in. Forty-seven FBS quarterbacks carried the ball on at least 50 designed runs (without sneaks or scrambles) a year ago. Fewer elite athletes are being moved off the position at lower levels. And pro teams are more open to installing concepts from college playbooks than they were decades ago, meaning quarterbacks who run those plays at lower levels can do so at the highest one as well.

At this point, it seems more likely the NFL will shift toward having more quarterbacks who feature meaningfully in their team’s run game as opposed to heading back toward the zero-run game that filled up most of the league’s first few decades. That’s going to impact the league and how we think about the sport. Will true pocket passers be marginalized for not producing enough with their legs? Can a quarterback taking hits on designed runs every week last for 20 years? Will defenses start returning toward heavier boxes and more traditional lineups? If that lapsed fan tuned in after 40 years, he would be looking at an evolved version of football that is going to keep accelerating forward in the decades to come.