The most important play in Sunday's Broncos-Jets game didn't officially happen, but it might be the basis for an entire offensive scheme. The 46-yard pass interference penalty in the fourth quarter drawn by Santonio Holmes against Broncos safety Renaldo Hill on fourth-and-6 turned what looked to be a sure Jets loss into a miracle win. But in the NFL play-by-play, you'll find it listed as "No Play." Neither Holmes nor Mark Sanchez get credit for the yardage, even though it moved the ball just as effectively as a catch in the same situation would have.
When it comes to adding these yards into a player's total, we've come down on the agents' side. Those yards should belong to Holmes, who was prevented from making a catch by Hill, who grabbed his facemask. But after the pass interference penalty totally shifted the course of the game, both Football Outsiders' Tim Gerheim (via e-mail) and ESPN's Bill Simmons (via Twitter) wondered whether a team could build an offense around drawing pass interference penalties.
The answer depends on how frequently you can draw them. And our research shows such a strategy to be an ill-advised proposition.
We can categorize the result of each long pass with one of seven general outcomes: completion, incompletion, interception, defensive pass interference, completion resulting in a fumble recovered by the defense, other defensive penalty (like illegal contact or holding), and other offensive penalty (like offensive holding, offensive pass interference). To keep things simple, we won't worry about the final three options.
In 2009, teams threw the ball 25 yards or more on 1,261 plays. They completed 350 of those passes, for a completion percentage of 27.8 percent. On those completions, they picked up an average of 42 yards and scored a touchdown 117 times. The pass fell incomplete 744 times, or 59.0 percent of the time. Teams threw interceptions on the passes 9.7 percent of the time.
As for defensive pass interference? It just doesn't happen very frequently. Those 1,261 bombs yielded just 45 pass interference penalties. Only one -- the bomb to Calvin Johnson that resulted in a game-winning touchdown on the subsequent play for the Lions in Week 11 -- took place in the fourth quarter with two minutes or less left to go. Overall, only 3.6 percent of the passes resulted in pass interference. That's just not frequent enough for a team to rely on the pass interference call as a reliable offensive weapon. If you threw 30 times a game, you could expect to pick up about one pass interference call a game, but you would also be throwing three picks.
Let's say that you built your offense around acquiring behemoth receivers that could get downfield and turn every jump ball into a nightmare for the defense. You'd have just built the 2008 and '09 San Diego Chargers, a devastating downfield passing attack. In those two seasons, the Chargers threw the ball 25 yards or more 96 times, but they picked up just six pass interference calls. Even that elevated rate of 6.3 percent wouldn't be anywhere near enough to justify throwing downfield 25 or 30 times a game.
The other factor that comes into play is the Law of Unintended Consequences. Teams that know of your game plan would provide far more safety help, lowering the completion percentage for your downfield passes. Just like they do for Hail Mary passes, referees aware of their potential complicity in this DPI scheme would be less likely to actually call a pass interference penalty, rendering your strategy moot.
While Holmes' play drastically changed the complexion of the Jets-Broncos game, plays like that just don't happen all that often, especially in such a dramatically important context. And while Holmes deserves credit for his work, relying on defensive pass interference as a primary offensive weapon is more likely to produce losses than wins.
Three Good Performances
1. Matt Schaub, Houston
Once the Texans went down two touchdowns in the third quarter, Schaub took over. Before that point, he was 10-of-14, but it was for 91 yards and all of two first downs (along with a touchdown). Half of his yardage came on a throw to Andre Johnson against blown coverage. Afterwards? 15-of-19 for 204 yards, with 11 first downs and that game-winning toss to Johnson in the back of the end zone. As for the other debatable pass interference call of the week? It's been overblown. It's true that Brandon Flowers didn't deserve to be charged with defensive pass interference, but Andre Johnson caught the pass anyway. And Johnson didn't do enough to come close to justifying offensive pass interference. These two teams may very well meet again in the playoffs, so Flowers may have a chance to get his revenge.
2. Kevin Kolb, Philadelphia
Now, let's compare. In his two starts since Michael Vick went down, Kevin Kolb's completed 73.3 percent of his passes. Vick was at 58.4 percent in the two starts he finished. While Vick unquestionably has the stronger arm, he averaged 8.8 yards per attempt; Kolb's at 9.6. Vick threw five touchdowns without an interception, better than Kolb's four scores versus one pick, but Vick fumbled twice to Kolb's lone fumble. For all his elusiveness, Vick took a sack every nine dropbacks; Kolb's taken one every 13. On the other hand, Vick is clearly the better rushing quarterback, but the soft factors also favor Kolb: He started against pass defenses ranked 12th (Atlanta, 27th last year) and 23rd (San Francisco, seventh last year) this year, while Vick was against teams ranked 15th (Detroit, 32nd last year) and 30th (Jacksonville, 31st last year). Finally, Kolb spent most of Sunday without DeSean Jackson and didn't have left tackle Jason Peters, either.
So why should Andy Reid stick with Michael Vick? Because just as benching Kolb after one half of poor play was shortsighted and didn't give the team a proper chance to evaluate Kolb, benching Vick after an injury doesn't give the team a proper chance to evaluate him. Going back and forth just erodes each player's confidence. What if Vick comes back and plays poorly? When does Reid go to Kolb? Philadelphia loves its backup quarterbacks more than any other city in America. No matter who ends up getting the job, he may not get it for very long.
3. Chris Ivory, New Orleans
The Saints' nominal starter isn't exactly a versatile back. He can't pass block; at least, he can't pass block well enough for Sean Payton to trust him with Drew Brees's life. He had 15 catches for 51 yards in three years at Washington State. He's already fumbled three times on 45 touches this year. What he can do, though, is run between the tackles and pick up consistent positive yardage, and with guards Jahri Evans and Carl Nicks in front of him, that's what the Saints need right now. Eleven of Ivory's 15 runs went for more than 5 yards, and eight of them went for first downs. He even picked up 17 yards and a first down on his first NFL catch. While a minor ankle injury in the fourth quarter forced him to the sideline, Ivory did enough work during the first three quarters to finish as the best back of the week.
Three Bad Performances
1. Hakeem Nicks, New York Giants
Even stars like Nicks have a bad game or two during their breakout seasons. Nicks spent most of his day against Broncos castoff Alphonso Smith, and although Nicks has about four inches and 30 pounds on Smith, the Lions cornerback was able to limit Nicks to three catches for 8 yards and a 10-yard defensive pass interference penalty. The Giants instead chose to attack the guy Smith was replacing at right cornerback, Jonathan Wade. It was Wade in coverage on Mario Manningham's long touchdown catch.
2. Jason Campbell, Oakland
Let's break down Campbell's day by quarter. After a ticky-tack 46-yard defensive pass interference penalty to start the game, Campbell went 5-of-7 to finish the first quarter. Not bad. In the second quarter, he went 0-for-4 with an interception. About as bad as you can get. In the third quarter, he got an incomplete grade because he didn't throw a single pass. The Raiders ran only three offensive plays, taking a 10-yard penalty in the process. Campbell did his part with a 7-yard scramble on third-and-long, ending the drive quickly. In the fourth quarter, he missed an open Louis Murphy deep by about a foot for what would have been a 98-yard touchdown, but then he completed a pass to Zach Miller that bounced off of a 49ers linebacker's hands and into Miller's. The incompletion was a much more impressive and positive throw than the completion. After that play, Campbell was 1-of-6 for 7 yards with another pick and a fumble on a sack. That's three potential turnovers (the Raiders recovered the fumble) on 24 dropbacks against a mediocre pass defense.
3. Ray Rice, Baltimore
To an extent, Ray Rice will always look worse by advanced numbers because the Ravens rely on him as a dump-off and safety valve in the passing game. While DYAR might see a 2-yard catch on third-and-10 as a negative play, it's better than a sack. On the other hand, all those meaningless catches and yards inflate Rice's raw numbers, too. Rice caught eight of the 10 passes thrown to him, but only two of those passes went for more than 4 yards, and only one was considered a successful play. The even bigger disappointment: Rice's 28 carries produced just two first downs. His long run of the day was only 8 yards. That's how the Ravens got 13 possessions against one of the league's worst pass defenses and scored just 20 points.
Three Stunning Performances
1. Dwayne Bowe, Kansas City
A punch line for his poor play as recently as the pregame shows on Sunday, Bowe's confidence came back with the opportunity to play one of the league's worst pass defenses. He was able to hold onto a 17-yard strike from Matt Cassel for a touchdown in the second quarter, and followed that by running through half the Texans' defense on a quick slant for a 42-yard score (with 35 yards after catch) in the third quarter. Five of his six catches came on third down, and the ones that weren't touchdowns all resulted in first downs. Things shouldn't get much more difficult next week; Jacksonville's pass defense is almost as bad as Houston's.
2. Deion Branch, New England
It was as if he had never left. Enigmatic and disappointing in Seattle, Branch hopped onto the Patriots' roster midweek and promptly put on a first-class showing. He outplayed FBO Top 25 prospect Lardarius Webb for the majority of the game, with half of his 12 targets resulting in a first down or a touchdown. The 43 DYAR he produced was more than he's had in a game since the 2007 season.
3. Miles Austin, Dallas
The minus-31 DYAR Austin put up doesn't represent his worst performance since he emerged as a franchise wideout against the Chiefs last season; he had a minus-43 DYAR day against the Packers in Week 10 of the 2009 campaign. It seems impossible to cover Austin at times, but the Vikings were able to hold him to two catches for 12 yards on five attempts. He actually had the ball in the end zone after a 68-yard catch-and-run, but he was called for offensive pass interference on the play. Cornerbacks Antoine Winfield and Asher Allen did a good job staying with Austin underneath, and the Vikings provided help over the top. Austin was able to take advantage of that as a decoy, occupying the safety long enough to create space for Dez Bryant's game-tying touchdown in the fourth quarter, but we're not at the point yet where we can assign an empirical value to that. It's also true that Austin could have drawn that safety and still made a bunch of big plays at other times in the game, and he simply did not.
Bill Barnwell is a writer for Football Outsiders.