Writers of Pro Football Prospectus 2008

Most Recent FO Features


» 2018 Free Agency Cost-Benefit Analysis

Is Kirk Cousins the best free-agent quarterback in recent memory? Should Trumaine Johnson or Malcolm Butler have gotten the larger contract? And what makes a free-agent contract good or bad, anyway?

16 Dec 2014

Any Given Sunday: Bills Over Packers

by Andrew Healy

At its best, sports analytics is like good science. New evidence comes in and forces everyone to rethink what they thought they knew. Nearly a hundred years ago, astronomers saw evidence of relativity and suddenly gravity wasn't so simple. The apple didn't just fall towards the center of the Earth. Curved space pulled it. Newton was out and Einstein was in.

A century later and equally significant, football analytics research would establish that offense is more important than defense. More specifically, overall team quality is four parts offense, three parts defense, and one part special teams. Sunday's game between the Packers and Bills seemed to illustrate that idea. The Packers came in with the NFL's top-ranked offense and No. 1 passing offense while the Bills entered the game with the second-ranked defense by DVOA, and No. 1 pass defense. Consistent with the supremacy of offense, the Packers came into Buffalo with a 10-3 record and as Vegas's favorites to win the Super Bowl. The Bills were 7-6 and on the fringes of the playoff picture.

But just like relativity left open questions at the atomic level, the offense-first football theory can't quite explain what happens in those rare cases when the best passing defense goes up against the best passing offense. That has happened 17 times now since 1991, 11 times in the regular season and six times in the playoffs. And what happened to Aaron Rodgers on Sunday actually isn't that out of line with what happened in those other 16 matchups. The top-ranked passing defense has consistently gotten the better of the top-ranked passing offense.

On Sunday, Aaron Rodgers came into the game as the top-ranked quarterback by DVOA. He was averaging 9.79 adjusted net yards per attempt (ANY/A), an amazing number even in this era of inflated quarterback stats. Against Buffalo, Rodgers threw for 1.98 ANY/A, his worst performance in his career as a starter in a game that he finished.

A similar pattern holds across the ten regular-season matchups from 1991-2013 that matched the top-ranked passing offense for the year against the top-ranked passing defense. On average, the passing offense dropped about 3 yards in ANY/A from how well it performed in other games to how well it performed against the best passing defense.

When the Irresistible Force Meets the Immovable Object, 1991-2013
Top-Ranked Offense vs. Top-Ranked Defense 53.8% 3.8% 3.3% 5.28 4.56
Top-Ranked Offense Vs. All Other Defenses 63.3% 6.1% 2.0% 7.42 7.71
All Other Offenses Vs. Top-Ranked Defense 53.2% 3.2% 4.2% 4.75 3.48

To put those numbers in perspective, the drop in ANY/A for top-ranked passing offenses when they face the best pass defense is about equivalent to going from Tom Brady this year to Geno Smith.

Note that three of the earlier cases involved teams in the same division, and thus two games in the same year: Washington and Philadelphia in 1991, San Francisco and New Orleans in 1992, and Indianapolis and Miami in 2000. These differences stay almost exactly the same if we don't include those years.

Thus, great passing offenses like the 2014 Packers often fall flat against top-ranked passing defenses like the 2014 Bills. Perhaps surprisingly, the top-ranked passing defenses have only performed a little bit worse against the top passing offenses than they have against the rest of the league. On average, the top pass defenses have given up just about one more ANY/A to the best passing offenses than to the rest of the league.

The Bills' dominance of the Packers' offense also fits with what happened in the six playoff matchups. In each case, the dominant offense averaged somewhere 7.81 and 8.87 ANY/A over the regular season. In only one of the six playoff games did the offense exceed 6.0 ANY/A, which is about what Mark Sanchez has posted this season. Two of those games had historically good passing offenses that were obliterated by top-ranked defenses. In 1999, the Buccaneers held the Greatest Show on Turf Rams to 11 points and five fewer ANY/A than their season average. In last year's Super Bowl, the Seahawks did about as well against the Broncos.

No. 1 Pass Offenses Vs. No. 1 Pass Defenses in the Playoffs
Regular Season ANY/A
Playoffs ANY/A vs. No. 1 Pass Defense
1999 NFC Championship Game (STL vs. TB) 8.31 3.33
2003 AFC Wild Card (TEN vs. BAL) 7.81 1.69
2006 Super Bowl (IND vs. CHI) 7.93 5.49
2008 AFC Divisional Playoff (SD vs. PIT) 8.04 7.44
2009 AFC Divisional Playoff (SD vs. NYJ) 8.30 5.07
2013 Super Bowl (DEN vs. SEA) 8.87 4.18

The Bills outdid even those two memorable performances on Sunday, holding Aaron Rodgers to eight ANY/A below his historic pace. And more than just the extent of Buffalo's dominance in pass defense, the way in which Buffalo accomplished that success was surprising. Buffalo leads the league both in sacks (with 49) and Adjusted Sack Rate, but they had only one sack of Rodgers (the strip sack on Green Bay's last offensive play) and three other quarterback hits on the day. The Bills mostly played coverage. According to my charting, the Bills blitzed the Packers just six times, in keeping with their strategy for most of the season.

More than the pass rush, the seven men in coverage for Buffalo controlled the game. While Rodgers was out of sync with his receivers early and those receivers were credited with seven drops (the most for any team in one game since 2008), the Buffalo linebackers and secondary gave little space to Packers' receivers. OK, yes, there was this almost unbelievable drop from Jordy Nelson that would have helped Rodgers have a different stat line.

Overall, however, the number of drops probably doesn't give the Bills' secondary enough credit. Rodgers' second interception also got labeled a drop when it was caused by excellent coverage from Ron Brooks. Brooks did a great job undercutting an attempted pick from Davante Adams and then blanketing Jarrett Boykin.

The Bills had that kind of close coverage on Packers' receivers all day. As a team, the Bills were credited with ten passes defensed. Safety Bacarri Rambo (two interceptions), linebacker Nigel Bardhan, and Stephon Gilmore each had two. Although he dropped a potential pick-six, Gilmore had a very good day. I had Rodgers going 1-of-9 for 6 yards when throwing in Gilmore's direction.

For Green Bay going forward, Sunday's struggles against the Bills might not mean that much. Top-ranked passing offenses like the 1991 Washingtonians, the 1995 Cowboys, the 1998 Broncos, the 1999 Rams, and the 2006 Colts all won the Super Bowl despite struggling to varying degrees in games against the top-ranked pass defense. While great pass offenses lose a surprising number of battles to top-ranked pass defenses, those great passing teams are still the ones most likely to win the war.

If They Only Had a Quarterback

After two weeks of largely shutting down the best two passing games in football, the Bills have surged ahead of the Lions into first place in overall defensive DVOA. It doesn't take analytics or a football Einstein to tell us why the Bills are still just 8-6. Kyle Orton is not very good. The Bills overcame more mediocrity from Orton on Sunday, but their ineffective offense is the reason that they remain on the outside of the playoff picture.

BUF -34.2% -50.1% 33.6% 49.5%
GB -10.5% -21.1% -22.8% -12.2%

BUF -32.9% -29.3% 33.6% 30.0%
GB -28.1% -31.4% -22.8% -19.5%

It's pretty incredible that the Bills won with how little they got from their offense, and in particular from Orton. Boobie Dixon (have to love anybody with that first name on a football field) had a few impressive runs. But Orton's successes were mostly limited to a long completion on a defensive breakdown to Bryce Brown and a late-game third-down conversion to Scott Chandler.

For the day, Orton matched Rodgers's 17.2 QBR. For the year, Orton's 39.3 QBR puts him in 27th place out of 30 qualifying quarterbacks. He does a little better by DVOA because of positive opponent adjustments, but he's still 27th (out of 40). Overall, the Bills offense is very good at setting up punts. Only five teams (Oakland, Jacksonville, Cleveland, Arizona, and Tennessee) have punted more often than the Bills. In offensive drive success rate, which measures the percentage of down series that result in a first down or touchdown, the Bills rank 29th, ahead of only Oakland, Jacksonville, and Tennessee. By that same statistic, the Bills' defense ranks first even though Buffalo has faced the third-toughest slate of opposing offenses in the NFL.

All of this is to say that what the Bills' defense has accomplished in recent weeks against great offenses is as impressive as what Seattle's has done against mediocre-to-bad ones. The Bills won't get the same amount of attention because the win-loss record will always drive the narrative. And that record has a relatively low ceiling as long as the defense and special teams have to carry the Kyle Orton albatross to wins where the offense averages 4 yards per play.

Posted by: Andrew Healy on 16 Dec 2014

26 comments, Last at 18 Dec 2014, 1:38pm by Andrew Healy


by Mike B. In Va :: Tue, 12/16/2014 - 3:13pm

For the day, Orton matched Rodgers's 17.2 QBR. For the year, Orton's 39.3 QBR puts him in 27th place out of 30 qualifying quarterbacks. He does a little better by DVOA because more quarterbacks qualify, but he's still 27th (out of 40). Overall, the Bills offense is very good at setting up punts. Only five teams (Oakland, Jacksonville, Cleveland, Arizona, and Tennessee) have punted more often than the Bills. In offensive drive success rate, which measures the percentage of down series that result in a first down or touchdown, the Bills rank 29th, ahead of only Oakland, Jacksonville, and Tennessee. By that same statistic, the Bills' defense ranks first even though Buffalo has faced the third-toughest slate of opposing offenses in the NFL.

*Sigh* Yep, welcome to being a Bills fan. Even competent offensive performances make this team a playoff contender (for the first time in 10 years.) Instead, well, our defense is really excellent for a .500 team...

by JoeyHarringtonsPiano :: Wed, 12/17/2014 - 5:22pm

Do you think they give up on Manuel? It's not totally unreasonable that they don't (what is totally unreasonable is the rumblings that St. Louis isn't giving up on Bradford yet). I'm not sure Manuel has enough attempts under his belt yet to consider him a lost cause, but I could be wrong. I still think he was overdrafted, but that's sunk cost at this point.

by Julio :: Tue, 12/16/2014 - 3:17pm

This explains why GB lost to NO. Actually, it doesn't. What explains
GB losing is the fact that they and Rodgers don't play well on the road
this year.


by RickD :: Tue, 12/16/2014 - 5:21pm

Rodgers was injured midway through the Saints game. That's when the Packers fell apart. I think a healthy Rodgers would have won that game.

by DisplacedPackerFan :: Wed, 12/17/2014 - 11:26am

Agreed. They were also missing Sam Shields and Morgan Burnett. Having half of your starting secondary out, doesn't help. Rodgers doesn't get hurt, and Shields and Burnett play and I think it would have still been a shoot out like the first half, but one where the Packers scored more and the Saints scored less.

Also, while the team doesn't play as well on the road, they have been healthier at home, and have faced a much easier slate of defenses at home. They have HFA, but I think it's constantly overstated this year.

by BobbyDazzler :: Wed, 12/17/2014 - 4:02pm

Nonsense. Rodgers was barely injured and while it did contribute to the loss the bigger factor was the Packers defense. They couldn't stop the Saints offense all game and Rodgers pretty much sucks when playing from behind as referenced by his 4th quarter comeback stats.

The Packers lost that game cos their defense couldn't stop a nosebleed, not because of a marginal injury that caused Rodgers to miss no playing time in any other game. He was only pulled late in the game when they were 3 scores down, so if he was so badly injured why did he play nearly the whole game?

by Perfundle :: Tue, 12/16/2014 - 3:30pm

One interesting thing this year is that so many of the best defenses are thoroughly mediocre in the red zone. Not counting this week's games, #2 Buffalo is 19th in points allowed per red zone trip, #10 New England is 18th, #5 San Francisco is 16th, #7 Philadelphia is 14th, #3 Denver is 25th and #4 Seattle is a shocking 32nd.

by Andrew Healy :: Wed, 12/17/2014 - 12:44pm

Red zone performance is pretty random from year to year. Even accounting for that, it's strange to see Seattle dead last.

by pdxfan :: Tue, 12/16/2014 - 4:03pm

How is the second photograph above not pass interference? The ball isn't at the receiver yet and he is being held/blanketed...

by Perfundle :: Tue, 12/16/2014 - 4:11pm

How can you possibly tell from that screenshot whether he's being held? It could be that the two made contact just as that picture was taken, in which case the ball arriving and the contact are near instantaneous. Also, blanketing someone is not pass interference.

by RickD :: Tue, 12/16/2014 - 5:25pm

I can't tell if the receiver's being held, but it certainly looks like the defender could use a saddle there. And he's clearly riding the receiver well before the ball has arrived.

by tuluse :: Wed, 12/17/2014 - 1:04am

Look at the blue line, he's within 5 yards of the line of scrimmage so he could have been jamming at the point the ball was released.

I'm not sure exactly how this rule works but it seems like it would be very exploitable if the offense could just throw immediately to any jammed receiver and get a PI call.

by pdxfan :: Wed, 12/17/2014 - 2:22am

subjective for sure.


"It is pass interference by either team when any act by a player more than one yard beyond the line of scrimmage significantly hinders an eligible player’s opportunity to catch the ball. Pass interference can only occur when a forward pass is thrown from behind the line of scrimmage, regardless of whether the pass is legal or illegal, or whether it crosses the line.

"Defensive pass interference rules apply from the time the ball is thrown until the ball is touched."

"If one guy is not playing the ball and one is, and you have that kind of contact, the one not playing the ball will always be called. You can have plenty of contact on a play where a defensive player tries to undercut a route, and he gets shoulder to shoulder, and it's not a foul because he's entitled to that ball."

So maybe if the receiver had acted a bit and flopped instead of trying to make the catch through the ride...?

by The Powers That Be :: Tue, 12/16/2014 - 4:18pm

What is the sound of one hand clapping?
Where is the bachelor's wife?
Was that pass interference?

by Nevic :: Tue, 12/16/2014 - 4:39pm

By how the rule is written, it is PI. You can even tell from this screenshot. I remember complaining about it in real time during the game. Others have already noted that this whole game was called very loose in the defensive backfield (despite a ton of overall penalties). The only PI called was OPI when a Bills WR was blocking about 6 yards downfield on a WR screen. All homerism aside, I have trouble believing that any team that covered that closely did not commit any fouls. When the Packers DBs were good (back in 09) with Charles Woodson and Al Harris, there was always complaining of constant grabbing and holding; some games it got called, some it didn't.

by 12 Days :: Tue, 12/16/2014 - 9:13pm

There were zero PIs, defensive holding, or illegal contact penalties on either secondary the entire game. Very surprising given it's 2014 and we normally see multiple such penalties per game. This crew was definitely "letting them play" which, all things being equal, you'd have to assume was more of a benefit to the team with the really good passing defense instead of the really good passing offense.

by Andrew Healy :: Wed, 12/17/2014 - 12:56pm

Great point on the lack of penalties in the secondary (thanks Nevic, too). Watching the tape, I didn't see much in the way of uncalled PI. One exception was the second INT, which I think is debatable. I'd rather not see it called, but it often gets flagged now.

The All-22 doesn't post until after I write, though, so it's hard to tell and my impression could be wrong.

by Nevic :: Wed, 12/17/2014 - 6:51pm

Just from my (homerish) memory there was a blatant hersey tug on Nelson on an incomplete pass on 3rd down in or near the red zone. Then later there was an obvious play where the defender was all over Quarless' arms down the sideline, not looking for the ball, but definitely intefered, but it was not called. The Packers got away with one as well, but I can't remember the situation very well. It was not like the refs missed a ton of calls, just a few, but it was in stark contrast to how they officiated EVERY other aspect of that game (19 total accepted penalties, plus 2 for the offsetting penalties).

by tuluse :: Wed, 12/17/2014 - 1:07am

"It doesn't take analytics or a football Einstein to tell us why the Bills are still just 8-6. Kyle Orton is not very good."

Of course Orton didn't play 4 of those games.

by The Ancient Mariner :: Wed, 12/17/2014 - 10:50am

If Orton were very good, he would have.

by tuluse :: Wed, 12/17/2014 - 1:34pm

Perhaps. I remember one time a fellow named Kurt Warner sitting on the bench while a highly drafted young QB was given a chance.

Not saying Orton is as good as Warner, but his DVOA is 20% higher than Manual's and he seems to be doing fine in an Alex Smith "don't make any mistakes" role. If he had been the starter through training camp and from the beginning the of year he might have been doing even better.

by stevied561 :: Wed, 12/17/2014 - 1:01pm

"...linebacker Nigel Bardhan..."

Not even close.

Really though, great article. It's awesome to finally see the Bills defense in the national spotlight.

by BobbyDazzler :: Wed, 12/17/2014 - 4:03pm

Double post. D'oh!

by big10freak :: Wed, 12/17/2014 - 6:32pm

With all credit to the Bills pass defense if McCarthy accepts that running the ball is still an option maybe he finally forces the Bills defense to adjust to THAT change in approach.

Instead, like so many times before, he keeps having his qb try what is not working.

Those two do share one common trait, stubbornness. Or it's arrogance. Either way it translates in refusing to compensate to what a really good opponent is doing in the way of defense.

by Andrew Healy :: Thu, 12/18/2014 - 1:38pm

Great point. Almost never would I want to take the ball out of Rodgers's hands, but Sunday might have been the exception. Packers averaged 6.2 YPC on 21 designed runs, compared to 4.1 YPA on 42 passes (and two INTs). Second-biggest gap of Rodgers's career in favor of the run in starts that he finished. Biggest was this game in 2010 (http://www.pro-football-reference.com/boxscores/201010100was.htm), but the YPC driven by one 71-yard run. Lacy's longest run Sunday was 22 yards.

And, wow, bad job by me on spelling. Nigel Bradham. He was really good, too, so deserved better. Great coverage on Quarless down sideline in fourth quarter, even if close to PI.

by Uncle Paul :: Wed, 12/17/2014 - 9:42pm

Wow, Any Given Sunday solves the paradox- the immovable object beats the irresistible force, as long as the IOs get some special teams luck.