There are no perfect quarterbacks. Even Peyton Manning throws a pass to the other team from time to time. It happened Sunday, but as his Denver Broncos' 51-48 victory over the Dallas Cowboys demonstrated, there are interceptions and then there are interceptions from hell. If every turnover were equal, the emotional damage would be easier to endure. It's the killer pick that leaves the deepest scars -- deep enough, sometimes, to make the long-criticized game manager seem attractive by comparison.
Of course, a cautious game manager wasn't going to suffice in the type of shootout Tony Romo and the Cowboys lost so painfully to Manning's Broncos. The late interception Romo threw hurt so much precisely because he had been so dynamic to that point. The play will only harden Romo's reputation as prone to the critical mistake, but he was not alone Sunday. Killer interceptions stood out for losing quarterbacks Matt Schaub, Eli Manning, Cam Newton and Blaine Gabbert.
Romo and Schaub have become known for breaking hearts. Eli Manning has had his moments as well. Categorizing interceptions by corresponding changes in team win probability gives us a context for just how much each turnover hurts. For example, the pick Romo threw against the Broncos reduced the Cowboys' chances by 25.9 percentage points, an unusually large drop for a single play. It was the ninth time since 2008 that a Romo pick reduced Dallas' chances by at least 20 percentage points. But if you forget Romo's reputation for a second and look at the numbers you'll find that Schaub is an even worse offender. That fact begs the question: Can Schaub survive his early season mistakes in Houston? Or, more to the point, should he?
A little context is in order before we get to Romo, Schaub and the other NFL heartbreakers. Ten years of play-by-play data have set expectations for winning based on specific game circumstances including the score, down, distance, quarter, time remaining, venue and field surface. This data showed the Cowboys with a 55.9 percent win probability when they had first-and-10 from their own 20-yard line with 2:39 remaining in a tie game Sunday. The sack Romo took on first down reduced those chances by 7.7 percentage points. Then came the fateful pick, and just like that, the Broncos had the ball at the Dallas 24. Denver's win probability stood at 77.6 percent and the Cowboys were essentially finished.
Through the late, late game Sunday night, NFL quarterbacks had thrown 2,610 interceptions since 2008. An interception thrown on a 50-yard Hail Mary to end the first half doesn't matter much. A pick thrown during the fourth quarter of a tie game matters a great deal. For the sake of this discussion, I've divided those 2,610 picks into three categories:
Bad: 60 percent of those picks (1,573 of the 2,610) reduced their teams' win probability by 10 or fewer percentage points. For example, the interception Andy Dalton threw Sunday on the final play of the first quarter of a scoreless game against New England produced a 10.8-point drop. This was a costly play, not a fatal one.
Worse: 30 percent (778) of all picks since 2008 produced reductions in win probability between 10 and 19.9 percentage points. The fourth-quarter interception Tennessee's Ryan Fitzpatrick threw against the Kansas City Chiefs fell into this category. The Titans trailed by three points with 6:23 remaining. The pick reduced their win probability from 38.9 percent to 22.4 percent, a drop of 16.5 points.
Worst: 10 percent of interceptions (259) since 2008 reduced their teams' chances by at least 20 percentage points. Romo's pick Sunday fell into this category. So did the interception Eli Manning threw against Philadelphia while trailing 22-21 with 10:49 remaining. That one produced a 24.1-point drop.
Quarterbacks aren't to blame for every interception. Bad luck, bad route running, bad pass protection and other factors come into play. Also, quarterbacks playing well enough to keep their teams competitive will have their teams in position to win -- and therefore in position to suffer significant drops in win probability on a single play. It's impossible to suffer a 20-point drop in win probability if your team has a 10 percent chance of winning. But losing is losing, and this has become a familiar refrain for Schaub.
Worst of the worst
The pick-six Schaub threw on the third play of the Houston Texans' 34-3 defeat at San Francisco gave him one in four successive games, a record. It was also his least damaging of the four, by far, because so much time remained in the game. His previous three came deeper into one-score games, affecting win probability by between 24 and 39 percent on each. Those three killer picks give Schaub a league-high 14 in the "worst" category reserved for interceptions reducing win probability by at least 20 percentage points. That's five more than anyone else since 2008.
Schaub needs a high pain threshold. He has thrown 64 picks overall since 2008, and 21.9 percent of those fell into the "worst" category. That's twice the league average and the highest rate for any current NFL starter.
The chart to the right shows the 10 quarterbacks with at least eight interceptions in the "worst" category since 2008. The final two columns show these plays in relation to total pass attempts. Schaub not only has the highest raw total, he also has one of the highest rates per pass attempt. While Drew Brees is tied for second with nine "worst" interceptions, his percentage is about half of Schaub's.
INTs Reducing Win Probability by 20+ Pct.
QB INTs Att. Pct.
Matt Schaub 14 2,585 0.54
Mark Sanchez 9 1,867 0.48
Carson Palmer 9 2,254 0.40
Tony Romo 9 2,571 0.35
Philip Rivers 9 2,805 0.32
Drew Brees 9 3,335 0.27
Chad Henne 8 1,468 0.54
Kyle Orton 8 1,766 0.32
Jay Cutler 8 2,532 0.32
Eli Manning 8 2,855 0.28
Stats cover 2008 to present
The pick Arizona's Daryl Washington collected off Carolina's Cam Newton in the fourth quarter of a one-score game Sunday fell into this category, changing win probability by 20.9 points. It was only the third time Newton has thrown one of these.
Some of the quarterbacks most closely associated with these costly interceptions also rank among the leaders in plays producing the largest positive swings. Fifteen of Eli Manning's touchdown passes since 2008 produced gains in win probability of at least 20 points, by far the most in the NFL over that span. Matthew Stafford (11), Romo (9), Tom Brady (9), Brees (8), Rodgers (8), Ryan (7) and Cutler (7) are next on the list.
The total for Schaub? Two, tied with a group featuring Troy Smith, Stephen McGee, Shaun Hill, Matt Moore, Curtis Painter, Christian Ponder, Alex Smith and a few players whose careers remain in their formative stages.
This is not an air-tight measure. Interceptions aren't the only way a quarterback can significantly hurt his team's chances, just as touchdowns aren't the only way he can affect a positive swing. But if Schaub or any quarterback is going to give away games with so many killer interceptions, we might reasonably expect greater positive return on the other side.
Schaub remains near the bottom when we expand the criteria from interceptions and touchdowns to include all QB action plays featuring win probability changes of at least 20 points. He has nine plus-20-percent plays and 16 minus-20-percent plays, producing a minus-seven differential that ranks second-worst in the league since 2008 among current starters. Philip Rivers is at minus-13, including minus-3 this season, but he's done enough right to trail only Peyton Manning in Total QBR this season. Schaub ranks 27th in QBR, one spot below Geno Smith and one ahead of the winless Eli Manning.
Schaub also has less equity built up. Since 2008, Eli Manning is plus-13 in QB action plays featuring win probability swings of at least 20 points -- and he's won two Super Bowls overall. Next comes Stafford (plus-10), Ryan (plus-6) and Rodgers (plus-5), followed by an odd mix of quarterbacks in the next tier. Schaub, meanwhile, is not only second from the bottom but he's also at minus-4 this season, meaning he's trending in the wrong direction.
Coach Gary Kubiak keeps saying Schaub will remain his starter, and Schaub has a good opportunity to stop his slide. A home game against the St. Louis Rams' pass defense should help in Week 6. Even though the Rams collected a pick-six off Gabbert on Sunday, producing a 26.7-point change in win probability, St. Louis still ranks among the bottom five in passer rating allowed (106.7) and Total QBR (69.1) allowed.
If Schaub can't beat the Rams, then what? A trip to Kansas City follows in Week 7. The Chiefs lead the NFL in passer rating allowed (62.3) and Total QBR allowed (25.1). They're tied for the league lead with two interceptions returned for scores.
Week 6 could prove pivotal for Schaub. His current pace is simply unsustainable.
why matt schaub has to go:
http://insider.espn.go.com/blog/mike-sando/post?id=1220
