Buffalo Bills QB Josh Allen

Josh Allen and the NFL's Roughing the Passer Problem

Week 5 began with a roughing the passer penalty on the fifth play from scrimmage of the Jets-Falcons Kipper Express, well before most of America even had time to pour our second cup of coffee. Week 5 ended in the wee hours of Monday morning with a roughing the passer penalty that erased a Chiefs interception and extended the final Bills touchdown drive, long after most folks gave up and went to bed.

Both fouls adhered directly to the strict letter of the written roughing the passer rule but only tangentially to its spirit. Neither hit was particularly brutal or placed Matt Ryan or Josh Allen in any peril. Both penalties had a minor-but-not-insignificant impact on the final outcome as they turned drive-ending defensive stops into positive offensive plays that led to scores. The penalties fostered quarterback safety in the broad sense that playing checkers instead of football would also foster quarterback safety.

A total of 48 roughing the passer penalties have been called through five weeks, putting the NFL on pace for 163.2 such fouls. That's the highest total by far in the NFL Penalties database, which dates back to 2009. The new 17th game is a factor, of course, but the league would still be on pace for 153.6 roughing fouls through 16 games. We can assume that's an all-time record: roughing the passer was rare through most of pro football history, because it required a broadsword.

Roughing the passer penalties have been on the increase for over a decade. Here are the annual totals since 2009:

Year Roughing the
Passer
Penalties
2009 69
2010 79
2011 100
2012 90
2013 90
2014 99
2015 102
2016 89
2017 107
2018 118
2019 136
2020 127
2021 163*

Linear regression suggests that league-wide roughing the passer rates are increasing at about 5.4 penalties per season. That's just enough of an increase to be noticeable across the years, because while the fouls are still relatively uncommon, they are often game-changing, controversial, memorable events.

The league changed the roughing language before the 2018 season, adding a prohibition against "landing on top of [the quarterback] with all or most of the defender's weight." A small spike in fouls in 2018 sparked one of those brief national debates on officiating that pop up roughly this time each year. The brouhaha abated, but roughing penalties steadily increased.

The likely cause of the steady increase before and after 2018 is the knock-on effect of other safety-oriented rule changes, such as helmet-to-helmet targeting. Simple drift in rule interpretation is also a likely contributor to the increase. The dip in 2020 might just be fluctuation, though it might also have been caused by the absence of crowds, who often act as Foley artists, making big wallops seem even more dramatic/severe/punishable.

The problems with the current interpretation of roughing the passer are obvious. It's nearly impossible to land anywhere without "all or part of your body weight," because that's what "landing" means. Defenders must now aim for a tiny strike zone between the quarterback's torso and thighs, even as the quarterback ducks or spins to escape. An errant swat that strikes the helmet or a dive that catches the quarterback's shin can earn a 15-yard foul; it's important to protect brains and knees, which is all the more reason to give defenders a fair chance when they aim for the midsection. This year, it appears as though a defender cannot even help his cause by slowing before impact or turning to cushion the blow.

Further, roughing the passer is not uniformly enforced and is subject to lots of potential biases. The Jets have been flagged for 19 roughing the passer penalties but have benefited from just two since 2020. It looks as though bad teams don't get any benefit of the doubt. Land Clark's officiating crew called the foul 15 times in 2020 and five times so far this year. Bill Vinovich's crew called it twice in 2020 and zero times so far this year. That's unlikely to be a purely random split.

Also, the Texans, Ravens and Cardinals combined to benefit from just one roughing the passer penalty in 2020. The Bills benefitted from a league-high 10. The potential implications are left to the class as an exercise.

All we can do about the increase in roughing-the-passer penalties is grin, bear it, hope that the rules are enforced more uniformly, hope even harder they don't impact the outcomes of games too severely, and hope hardest of all that the NFL someday tweaks the interpretation of the foul with some language like the defender must apply additional downward force after the quarterback has fallen for a foul to occur.

Roughing the passer was initially designed to prevent dirty plays. It was rewritten to prevent dangerous plays. Now, it's turning into a "Gotcha." That's bad for the game. It's also dangerous, because some defender who thinks that he'll get flagged for anything he does might end up doing whatever he wants.

Former Jaguars to Watch

Let's do something a little different and use this week's Five to Watch segment to check in on the 2017 Jaguars defensive diaspora!

Calais Campbell, Baltimore Ravens: Campbell is sackless this year, but he recorded a third-down tackle for a loss on Monday night, followed by a blocked field goal, making the Ravens' overtime comeback over the Colts possible. The Ravens need all defensive hands on deck when they face the Chargers on Sunday.

Malik Jackson, Cleveland Browns: Jackson is having a positive impact for the Browns as an interior pass-rusher. He will spend Sunday helping Myles Garrett and others chase down Kyler Murray.

Yannick Ngakoue, Las Vegas Raiders: Ngakoue recorded a pair of sacks against the Bears last week. He'll try to keep the Raiders from falling into a pit of despair when they face the Broncos on Sunday.

Jalen Ramsey, Los Angeles Rams: Having a great year, of course. Ramsey will help take away the Giants' deep passing game on Sunday, leaving them with nothing.

A.J. Bouye, Carolina Panthers: Bouye has rebounded from some subpar years and is playing well as a slot cornerback. The Panthers and Vikings face off on Sunday to determine who will linger at the bottom of the playoff race and end up with the final NFC wild-card spot.

Tashaun Gipson, Chicago Bears: Gipson recorded a sack against the Raiders and recovered a fumble against the Bengals. He has been a major factor when healthy, and he gives the Bears a very slight chance to upset the Green Bay Packers. Receiver Allen Robinson, another Jaguars escapee, is still atoning for the sins of a past life, and will also help the Bears try to accomplish the unlikely on Sunday.

Dante Fowler, Atlanta Falcons: He's on bye this week, but Fowler is playing rather well, so we did not want to leave him out.

Gosh, that's a lot of defensive talent. Could you imagine what a defense with all of these players could accomplish in 2021? It would probably be a top-10 unit in DVOA, given an ordinary supplemental cast.

The 2021 Jaguars defense ranks 31st in DVOA.

Leaderboard of the Week

Every Thursday, Walkthrough examines a random (and usually obscure) leaderboard from Football Outsiders, Sports Info Solutions, or elsewhere on the analytics Interwebs in search of deep truths and wisdom.

This week, we'll be looking at the empty backfield passing leaderboards, otherwise known as the "Matthew Stafford and a bunch of other guys" leaderboard.

Here are the numbers for all of the quarterbacks with more than 30 attempts from empty sets this season. I included NFL passer rating because it's a non-terrible quick-and-dirty metric for data like this.

Empty Backfield
Player Team Att Comp Yards TD Int Sacks Rating
Mathew Stafford LAR 62 40 549 4 1 0 107.5
Joe Burrow CIN 53 38 386 4 2 4 101.6
Kirk Cousins MIN 43 31 397 2 1 1 106.4
Derek Carr LV 41 28 277 2 2 1 83.1
Aaron Rodgers GB 31 22 419 1 0 2 124.1
Patrick Mahomes KC 31 26 347 7 0 0 152.9
Tom Brady TB 31 21 192 1 0 0 95.1

Football Outsiders' Derrik Klassen explained on our podcast a few weeks ago that Stafford's ability to do more from an empty set than toss quick passes to his interior receivers is one of the big reasons he's such an upgrade over Jared Goff. Stafford averages a healthy 8.9 yards per attempt, while Cooper Kupp averages a whopping 17.2 yards per catch on 18 receptions over 27 targets.

Stafford's rate stats are impressive enough, but the fact that he's operating out of an empty backfield over a dozen times per game, twice as often as quarterbacks such as Aaron Rodgers or Patrick Mahomes or Tom Brady, illustrates just how unique the Stafford-Sean McVay Rams offense has become this year.

Four sacks and two interceptions ruin otherwise impressive empty-backfield numbers for Joe Burrow. Zac Taylor still doesn't have the offensive line he needs to deploy empty backfields as often as he does, but Taylor will be the last person to realize it. Burrow will never be a Stafford-like rifleman from the empty set, but he can grow into an almost Brady-like ball distributor if he survives.

The Vikings use so many constricted formations and 21-12-22 personnel that it's shocking to see Cousins third on this list. As you might expect, 25 of Cousins' targets went to Justin Jefferson or Adam Thielen, for 20 receptions and 303 yards. A team that likes to run lots of empty sets might want to invest in a No. 3 and No. 4 receiver, but that's not how the Vikings do business.

The Raiders use empty sets mostly for a quick underneath game, with middling results. Hunter Renfrow is 8-of-10 for 79 yards on empty-set targets, Darren Waller 7-of-12 targets for 53 yards. It's all very adequate, like most of the Raiders offense.

Mahomes has thrown four touchdown passes from empty backfields from inside the 20 and three touchdowns on just four attempts from inside the 10. Stafford leads the league with eight empty-backfield passes from inside the 10-yard line. With the addition of Josh Gordon and likely absence of Clyde Edwards-Helaire for a few weeks, Andy Reid might want to consider emptying the backfield more frequently.

Rodgers and Brady are excellent football players.

Most teams have deployed empty-backfield sets between 20 and 31 times this year, or about four to six times per game. That indicates that empty sets fill the same niche in most offensive ecosystems: a mini-wrinkle to spread the offense, define a mismatch, set up a quick-throw opportunity, etc. Again, that just makes the Rams offense look even more remarkable.

As for rushing from an empty backfield, Trevor Lawrence leads the league with five carries for 29 yards and one touchdown. Urban Meyer thinks Lawrence is Tebow and that he can keep "fooling" opponents with empty-set draw plays at the goal line. Trey Lance has five carries for 19 yards and a touchdown. Lots of Lamar Jackson/Josh Allen types have four carries, as does Taysom Hill. But Jonathan Taylor leads the league with 38 rushing yards for this run, which officially counts as a carry from an empty backfield.

Some other notes:

  • Dak Prescott has been sacked four times in just 24 empty-backfield dropbacks and averages just 7.3 yards per attempt. The Cowboys do so many other things well offensively that they might be better off scrapping their empty backfield package altogether.
     
  • Andy Dalton and Justin Fields have attempted just three plays each from empty-backfield formations. Each has completed one pass for a combined 15 yards. Fields has been sacked twice. The Bears are more likely to deploy a sixth offensive lineman than an empty backfield. When discussing Matt Nagy, the formations and personnel groupings themselves are unlikely to be the problem.
     
  • Lamar Jackson has thrown three interceptions from an empty backfield and has been pressured 16 times on 34 dropbacks. Such formations probably don't make best use of his diverse talents.
     
  • Ryan Tannehill has thrown 20 passes from an empty backfield. His catchable pass rate of 73.7% is the lowest among starting quarterbacks. The Titans have a thin, injury-riddled receiving corps, issues along the offensive line, and Derrick Henry. Mike Vrabel should threaten to bite coordinator Todd Downing the next time he empties the backfield.
     
  • Zach Wilson averages a league-low 42.9% completion rate and 3.5 yards per attempt from empty sets. As Leaderboard of the Week so often concludes: the Jets are the Jets are the Jets.

Walkthrough Thursday Night Sportsbook: Tampa Bay Buccaneers -6.5 at Philadelphia Eagles.

This line has remained under a touchdown because A) Tom Brady has a thumb injury he is being typically coy about; and B) the Eagles just beat the Panthers, and folks still think beating the Panthers is an accomplishment for some reason.

If Brady's thumb scares you away from just rolling with the Bucs, consider taking the Eagles with a +235 moneyline instead of the points. If Brady ends up throwing flutterballs like he did in 2020 against the Saints in Week 1 and the Bears, an Eagles upset becomes about as likely as a narrow Buccaneers win/Eagles cover, so you might as well go for the payout gusto.

Walkthrough isn't enough of a Philly homer to take that action. Instead, we're eyeing the Buccaneers (-4) to lead in the first half, because the Bucs have outscored opponents 79-57 in the first half, while the Eagles start every game with 20 consecutive screen passes and have therefore been outscored 69-44 in the first half.

Walkthrough also loves Longest Completion props, but both Brady and Jalen Hurts are way up at 38.5. Brady's thumb makes a long-completion wager a bad risk, while the house has figured out that Jalen Hurts completes at least one bomb per game after Nick Sirianni has established the screen. So instead, we're hitting some Thursday Night Lenny: Leonard Fournette Over 88.5 Rushing + Receiving Yards at -115.

If Brady is hurt, he'll force-feed Fournette both handoffs and screens, and not even Fournette can drop all of them.

Comments

38 comments, Last at 15 Oct 2021, 12:20am

1 AJ Bouye may not be tested…

AJ Bouye may not be tested much by a slot receiver on Sunday. He may get opportunities to tackle Cook and Conklin, but the Vikings mostly run and pass attempts to WRs are usually for Justin Jefferson.

2 fournette

the quick turnaround might take touches away from fournette.

look at rojo yards (16.5), and fournette longest carry (every RB that plays philly since graham went down has a field day)

3 "The dip in 2020 might just…

"The dip in 2020 might just be fluctuation, though it might also have been caused by the absence of crowds,"

What makes you think that 2020 is a dip, and not that 2019 is a spike? 2020 fits the prior curve way better than 2019. 

As to calls, you can't just call the dirty/dangerous ones - because the goal is to prevent those, not to punish them. The 15 yards doesn't matter if we've got a QB with a concussion or blown knee. You have to punish the behavior that leads to those dirty hits. 

7 2018 rule change

The 2018 rule change that all or most weight be called a foul is the best reason to think that 2020 is a dip, along with the extrapolation this year leading to even more calls. I wasn't awake to see the hit on Josh Allen, but was working through my coffee when the hit on Matt Ryan rescued the Falcons opening drive. I am no Jets fan, but obviously enough of a fan of football that I was watching the putrid Jets/Falcons, and that call made me want to spit out my coffee. It was basically a fundamental tackle. The defender hit Ryan perfectly midtorso with both arms wrapped around him in the prototypical tackle formation as the ball was coming out and momentum carried them both to the ground. This was only called because the rule change in 2018 made a fundamental tackle on a QB a potential penalty.  The roughing penalty call is on the referee, of course, and my guess is some referees can't and won't bring themselves to throw a flag on that kind of play, whereas others do. Which makes it eve harder for defenders, since some games they won't be called for that kind of play, and others they will. To that end the league may be trying to encourage more consistency, which by the letter of the law is going to lead to the kind of drift Tanier mentioned into more enforcement rather than less.

26 "along with the…

In reply to by NYChem

"along with the extrapolation this year leading to even more calls. "

My understanding is that these sort of calls are always weighted towards the beginning of the year. So I don't think the extrapolation is useful data - I'd give it +/- 50% error bars.

22 Which one??

So which behavior are you talking about, Sleeves? Gravity, or tackle football? If you've got a third candidate, as John Lennon phrased, we'd all love to see the plan.

32 His whole interpretation…

His whole interpretation seems to be missing the data.
2009 and 2010 look like they are part of a linear trend.
Then 2011 through 2017 is basically 95 +/- 6. That's stable with noise.
2018 - 2021 looks different but knowing something happened in 2018 helps explain all of that.

OK 2017 is a bit outside that, but that could be for any number of reasons. An emphasis, an actual trend, one player who couldn't get the rules through his head, I don't know, but it would be interest to look and not just write it off because some linear fit completely misses there is an obvious step/inflection point.

To me it looks like something happened to the rules in 2011, then again in 2018. The 2018 rule was discussed. The smaller increase from the 2011 - 2017 plateau vs what we see in 2019 and 2020 could simple be because crews were still getting used to calling the new rule change. 2021 may not be increasing as much as it seems, as was pointed out I'm pretty sure these calls happen more frequently early season (something testable). If it is then we should fix the 2021 project to see if it fits with 2019 and 2020. If it does you have an explanation. 2018 rule change that took a bit of time to settle in for how it's called. If not we have something interesting.

So the question is, what happened in 2011 that jumped from the 09/10 numbers then stayed basically stable till the 2018 rule change?

37 So 2011 is when the…

So 2011 is when the launching/spearing rule changes went into effect. 

So that with the 2018 changes seems to explain the two increases resulting in the steps. Though as mentioned the 2018 changes may not have settled out yet and the 2017 data is enough off of the 95 +/- 6 that was 2011 - 2016 to be more than noise.

Not everything is linear, and this data definitely looks like it can be explained by the two rule changes and higher rates of roughing calls earlier in the season, though that is still reasonably uncertain that something else besides the 2018 rule change/point of emphasis is going on.

4 From the front page blurb:…

From the front page blurb:

The NFL is on pace to call a record number of roughing the penalty penalties.

I almost don't want it to be fixed because I want the image of Chase Young or TJ Watt doing an elbow slam on a thrown yellow flag.

5 "If Brady ends up throwing…

"If Brady ends up throwing flutterballs like he did against the Saints in Week 1" 

Brady threw non-flutterballs against Dallas in Week 1.

12 "If Brady ends up throwing…

In reply to by MikeTanier

"If Brady ends up throwing flutterballs like he did against the Saints in Week 1 and against the Bears in 2020"

That sentence is confusing the way you meant it. I think the most reasonable interpretation of "Saints in Week 1" is the current season, and that your "Bears in 2020" is the example from a previous year.

13 I wish he threw flutterballs…

In reply to by MikeTanier

I wish he threw flutterballs against Dallas in week 1.

But yeah, you need to rewrite that sentence: "If Brady ends up throwing flutterballs like he did against the Saints in Week 1 and against the Bears in 2020..." You just need to delete "in week 1" and the second "against."

6 really?

the hit on allen in the chiefs/bills game is exactly what you described as a problematic behavior, not a problematic call. frank clark clearly hits allen, then leaves his feet to drive his entire weight onto allen's collarbone. the bills offensive lineman has a hand on clark's back but doesn't push or put any weight on him. clark had allen dropping and then left his feet - he could easily have swam over him or rolled to prevent throwing his entire body weight on him. this is a textbook example of a (modern) RTP call, similar to how ed oliver's penalty that canceled out a critical third down for the chiefs was textbook in that oliver flopped onto the side of mahomes' knees instead of rolling to avoid it.

both players have histories of bad RTP penalties that they could have easily avoided. this is not rocket surgery here, these are easy examples of well-called penalties that prevent QB injuries. nobody wants mahomes to blow an ACL or allen to break his collarbone and these penalties help prevent the behavior that causes those things to happen more often.

there have been many terrible RTP calls this year. these were not two of them. calling out especially clark's is baffling. complain about chase young's a few weeks ago all you want, but these were good calls.

21 It's the leverage, not the call itself

In reply to by prophetikmusic

The Clark penalty on Allen wasn't an egregiously bad call, as has already been mentioned it fits the definition of what isn't allowed. I think the problem most have with it is the leverage of the situation. But that's incidental. It does point to the real problem though: no one wants a hard fought football game to be decided by someone's judgment, where someone else's judgment would have led them to not intervene. Which is completely understandable. Parsing individual calls though probably isn't the answer, because there'll always be grey areas or disagreement. What's more important to the credibility of the league is for a certain very high percentage of these calls, say 90% or more of them, to both pass the 'letter of the law' test AND an eyeball test that accounts for the speed of the game as it's actually played by defenders. Now I'm not saying I know how to do that, but if I were the league I'd be more concerned with consistency and credibility than with trying to cover every eventuality, all the time. And I agree with Mike that if there's going to be a flag anyway someone will decide to make it count. The Bills/Chiefs game is an example, the Bills chose to take holding and PI flags in order to get physical in place of what probably would've been yards gained in the absence of such contact. 

29 ". It does point to the real…

". It does point to the real problem though: no one wants a hard fought football game to be decided by someone's judgment, where someone else's judgment would have led them to not intervene. Which is completely understandable. "

I don't disagree - but that's never going to be a realistic thing. Holding, PI, catches, etc, are all going to always require judgement calls. 

10 The Allen roughing call had…

The Allen roughing call had more than a minor impact on that game. If the play stands, the Chiefs have the ball in good field position down 11 with nearly the entire fourth quarter left to play. Instead, their weak, exhausted defense let the Bills march down the field again, and by the time the Chiefs saw the ball again, they were down 18 points with six minutes to play. 

14 "exhausted" is a stretch

The Bills only ran about 57 plays to the Chief's 75+ because they were scoring so quickly. They also had the whole rain delay and forced the Bills to punt three times in a row after 4 or fewer plays. The KC D is bad, but exhaustion is no excuse.

31 Clear makeup call ...

... for the nonexistent holding on the prior play that got the Bills up to their own 26. The Bills never should have been in that spot (3rd and long in the shadow of their own end zone) in the first place, full stop.

11 "Week 5 ended..."

NOT in the wee hours of Monday morning, but near midnight on Monday, as the Ravens completed their comeback against the Indy Abominations.  

A comeback aided, in large part, by a rare roughing the passer call in favor of Lamar Jackson.  Jackson even commented after the game that the call on roughing him was so rare, he was able to thank the official by name.

Land Clark.  As to why roughing is called so rarely on behalf of Lamar, as a good writer I know put it, "make of it what you will."

16 an Eagles upset becomes…

an Eagles upset becomes about as likely as a narrow Buccaneers win/Eagles cover, so you might as well go for the payout gusto.

I would love for an Eagles upset to be remotely possible, but even if Brady decides to rip his arm tendons out or something, I can't see the Eagles being able to capitalize. The Bucs defense is just built to destroy offenses like theirs.

23 I think it's more like:…

I think it's more like:

Sirianni: OK, Jalen, our plan is to throw screens, slants, wheels - quick passes to guys that are open. Maybe a few runs to get Miles warmed up too. Then when the linebackers close in, and safeties come down to help, we sucker-punch 'em with a bomb to Watkins or DeVonta. Fundamental football.
Hurts: Solid plan, Coach!

(game starts)
(screen tackled for a loss by White)
(run for a gain of 1)
(slant tackled for a gain of 3 by a linebacker)

Jordan Whitehead, lounging 15 yards deep: Dude, this game is easy.

18 The Allen roughing call was…

The Allen roughing call was clearly a make-up for a ridiculous ticky-tack Holding call on Mitch Morse the play prior.

24 Landing on the QB

In the 2nd paragraph after the table the quote is "all or most" of the weight, while two paragraphs down it's "all or part".  I hope "most" is the actual rule.

25 To me, it looked like a…

To me, it looked like a legitimate call. Allen was tackled, and then Clark added to it, in an attempt to injure. There was an additional motion in which Clark attempted to drive Allen into the ground.
It might look like a close call to some, but when you are being tackled, you know the difference between someone tackling you and someone trying to take you out. That was definitely on the side of taking Allen out.  
The announcers were all over the Bills, but KC got what, 5 first downs from penalties?

33 The hit on Allen looked to…

The hit on Allen looked to me like 49% of Clark's weight, so bad call (humor). Honestly it seemed like Clark could have tried to roll off a bit but it was not a drive him into the turf hit (IMOH). The roughing calls that  bother me more are the glancing hits to the QB helmet that are flagged. The rule says hits to head need to be made "forcibly" to be a penalty but it seems the referees are emphasizing the note to the rule " When in doubt about a roughness call or potentially dangerous tactic against the quarterback, the Referee should always call roughing the passer"

35 RTP is a tough one

Lost in this is that the Chiefs previous TD drive would have ended, but an RTP was called on 4th down after a stop. Was the hit low? Sure, when a defender has been knocked to the ground that's where he grabs the QB.

Both the Chiefs' and Bills' last TDs came on drives that were extended by RTP calls. In both cases I felt (as a neutral observer) that the defenders would have been unable to avoid contact. 

I didn't feel like either of these hits warranted a penalty, but this is where the league is now.

38 Looking at it again, the…

Looking at it again, the ball was gone and the natural motion of Frank should have been to keep going in the direction he was driving. Instead, he changes the direction of his body to drive Allen into the ground. 
The ball was gone. He could have hit Allen without drawing a penalty. But the tackle into the ground was roughing the passer. Frank knew it. Look at his reaction as soon as he finishes the tackle.