Writers of Pro Football Prospectus 2008

Most Recent FO Features

ReplayIns14.jpg

» Scramble for the Ball: Getting it Right?

Instant replay review is one of the cornerstones of the modern NFL. The process and its myriad special rules have been internalized and constantly debated. Mike Kurtz wonders: is it worth it?

02 Aug 2010

Wisdom Of Crowds 2010: QB Part II

In the second week of our Wisdom of Crowds feature for 2010, we'll take a look at five more quarterbacks to see what the masses -- well, the people who follow my account on Twitter, at least -- think about the 2010 numbers of several very intriguing players.

For the uninitiated, every weekday, we ask folks on Twitter to predict three statistics for an NFL player; for quarterbacks, those three stats are passing yards, passing touchdowns, and interceptions. To keep the scale of the statistics uniform, we ask that the readers to assume that the players stay healthy for a full 16-game season.

You can also check out last week's projections here. Onward with this week's fab five!

Monday: Aaron Rodgers
Average: 4,440 passing yards, 32 TD, 12 INT
Maximum: 4,850 passing yards, 40 TD, 7 INT
Minimum 3,821 passing yards, 27 TD, 18 INT

Several factors come into play with projecting Rodgers's numbers. He played an extraordinarily easy schedule a year ago, which inflated his raw stats a fair amount. His schedule should be tougher this year, which alone should depress his raw totals from last year's 4434/30/7. Those seven interceptions came on 541 attempts, which yielded a league-low interception rate of 1.3 percent; previous research on interception rates (see Pro Football Prospectus 2008) suggests that such a low rate is almost sure to increase.

It's reasonable to suggest that Rodgers would benefit from having a better offensive line in front of him, as he did during the second half of last season, but he actually threw for more yards in the first half of the campaign (2,255), while he was getting sacked at a historically-high rate. His yards per attempt also declined despite the superior protection. Throw in the murky status of Donald Driver and, well, there are enough reasons to believe that Rodgers might struggle to reach that average projection.

Tuesday: Jason Campbell
Average: 3,247 passing yards, 19 TD, 15 INT
Maximum: 3,950 passing yards, 25 TD, 7 INT
Minimum: 2,600 passing yards, 13 TD, 21 INT

I find the 15-interception average for Campbell to be very interesting. He threw 15 a year ago, a career-high, but his career interception rate is at 2.3 percent; one flukily-low year at 1.2 percent, and then three years in-between 2.6 and 3.0 percent. Assuming that the 2.3 percent figure is the most accurate of the four, he would need to dropback about 650 times to get to 15 picks. That won't be happening. Even if we throw out the outlier year, his career rate would be 2.75 percent; that would take 546 passes to hit 15 INTs. I just don't think Campbell will be throwing the ball that frequently with Michael Bush and Darren McFadden around; a year ago, the assorted Raiders quarterbacks only threw the ball 484 times.

On the other hand, I could be underestimating the change in context here. Campbell's almost unquestionably playing with a worse group of receivers, although Zach Miller might be better than anyone the Redskins have in their lineup this year. That could drive an unexpected increase in his interception rate. The Raiders also might choose to throw the ball more frequently with a quarterback that's not terrible, and Campbell might be throwing deep more frequently than he did in Washington. I still think the interception figure is a tad high.

Wednesday: Jay Cutler
Average: 4,062 passing yards, 27 TD, 20 INT
Maximum: 4,800 passing yards, 37 TD, 14 INT
Minimum: 3,400 passing yards, 19 TD, 28 INT

Speaking of dramatic changes in context! This time last year, we were debating whether Cutler had some sort of third down yips that had led him to a league-high 11 picks in 2008. Well, he certainly regressed to the mean in 2009; unfortunately, he was beset by a rising tide of interceptions on all downs. So here we are again in 2010, suggesting that Cutler's going to regress to the mean with picks.

From 1990-2008, there were 31 instances of a quarterback throwing 20 or more interceptions in a season and getting to throw 400 or more attempts in the subsequent season. (Cutler threw 26 a year ago, and barring injury, he'll hit 400 attempts this year sometime around Week 12.) In the subsequent season, those quarterbacks averaged just under 15 picks. Cutler might be throwing more attempts than those players, of course, but KUBIAK has a somewhat controversial opinion on the run/pass splits of Chicago's offense. Much like Campbell, I think the picks figure is a little high, and while 27 TD may also be a little on the optimistic side, his passing yardage seems pretty reasonable.

Thursday: Mark Sanchez
Average: 3,138 passing yards, 18 TD, 16 INT
Maximum: 3,740 passing yards, 28 TD, 8 INT
Minimum: 2,600 passing yards, 11 TD, 22 INT

We can apply the same interception logic to Sanchez, who threw 20 picks as a rookie in nearly 200 fewer dropbacks than Cutler. Regardless of Sanchez's inexperience, it seems close to impossible that he'll have 5.5 percent of his passes picked off this year. With that in mind, Sanchez's attempt total is also going to rise. He only threw 24 attempts per game last year, the eighth-lowest total of any quarterback with 15 starts or more since 1990. (That list is mostly pretty middling quarterbacks, guys who ran the ball a lot, and Steve Young's 1997 season; Young threw 356 passes in 15 starts, but completed nearly 68 percent of those throws and averaged 8.5 yards per attempt.) Throw in the expected decline in his running game and the expected improvement coming from a year of NFL experience, and well, there's a lot to digest.

If he averages 6.7 yards per attempt again, Sanchez would need 468 attempts to hit that average passing yard total. I don't think he's going to throw that many passes. I also think the Jets are going to continue to rely heavily on the run inside the red zone. With all that in mind, I think each aspect of his average prediction is probably a little high.

Friday: Tony Romo
Average: 4,287 passing yards, 30 TD, 13 INT
Maximum: 4,727 passing yards, 39 TD, 9 INT
Minimum: 3,760 passing yards, 24 TD, 20 INT

Romo's prediction is another one that KUBIAK's taken some heat for on our message boards. His projected line is for 3,800 passing yards, 24 TD, and 17 INT on 517 attempts. That's just about the minimum projection for Romo from the public.

In this case, I'm probably going to subjectively say that KUBIAK is struggling to predict Romo's line. Last year, we had Romo at 3666/25/12 on 492 passes; in all fairness, KUBIAK didn't think Miles Austin was going to become Miles Austin, but it's a pretty big miss for a guy who ended up at 4483/26/9 on 550 attempts. If you scale Romo's figures to 16 games, his passing yardage and attempt totals have been pretty consistent. His scaled attempt totals go 520, 553, and 550; his yardage has gone 4211, 4244, and 4483.

I think the issue here is that KUBIAK struggles to project the Cowboys' offense in the context of having three running backs that get significant playing time. It's seeing an expected decline in passing attempts that I don't think is likely to happen; if we scale Romo's projection to 550 attempts, he gets to 4042 yards, which is still a bit of a decline from his previous numbers, but a less unreasonable one.

This is a good example of how to combine KUBIAK and your own opinions on players. Subjectively, I think Tony Romo will exceed his KUBIAK projection and put up a passing yardage total similar to his average projection, although with fewer touchdowns. I also suspect his interception rate -- which has gone from 3.9 percent in 2006 to 3.7 percent, 3.1 percent, and then 1.6 percent a year ago -- is going to bounce back up a bit. KUBIAK gives me something to think about in Romo's case, but I'm going to discount some of its expected regression here.

This week, we'll be wrapping up the quarterback section of Wisdom of Crowds with a look at five more signal-callers. If you're interested in participating, march over to the Twitter box on the left side of the page and follow me, or any of the fine members of the FO staff.

Posted by: Bill Barnwell on 02 Aug 2010

12 comments, Last at 09 Aug 2010, 11:38pm by Braada

Comments

1
by Key19 :: Mon, 08/02/2010 - 2:07pm

Thank you.

2
by mrh :: Mon, 08/02/2010 - 2:18pm

I think the Cutler INT crowd number is about right, maybe a little low.

Here are the INT rates of the offenses coached by Mike Martz from 1999-2004 in STL (I left out partial year 2005), 2006-2007 in DET, and 2008 in SF:
2.8, 3.9, 4.0, 4.3, 3.8, 3.8, 3.7, 3.7, 3.7.
Mean = 3.8% on 575 att or 21.7 INT per year.

Those are remarkably consistent numbers across a variety of starting QBs, but primarily: Warner, Bulger, Kitna. Only 1999 sticks out as a significant deviation from the norm.

Assuming Martz' career averages of 575 team pass attempts with a 3.8% INT rate, with 550 attempts for Cutler in 16 starts and 25 for the other QBs in mop-up duty, Cutler throws 21 INTs.

3
by tuluse :: Mon, 08/02/2010 - 3:56pm

I expect Cutler to throw for both more touchdowns and more picks than Kubiak is predicting.

It's interesting that Martz's interception percentage is rather close to Cutler's career int percentage (3.5%).

I do see some signs for less picks than last year (it would be hard to throw for more). 1) Aromashadu playing more and possibly starting, 2) Cutler becoming more accustomed to the other receivers in Chicago, and 3) improvement on the offensive line even if it's just to mediocre.

9
by Big Johnson :: Wed, 08/04/2010 - 2:43am

"Aromashadu playing more and possibly starting"

That doesnt even make sense.

10
by tuluse :: Wed, 08/04/2010 - 3:48am

Why not?

He had the highest DVOA of any Bears receiver last year, and he is a large reciever who can jump ans use his body to shield defenders. The exact kind of receiver Cutler has shown a preference for his in career thus far.

11
by Eddo :: Wed, 08/04/2010 - 1:11pm

Two more (subjective) points in your favor, tuluse:

1) Cutler and Aromashodu seemed to be on the same page when the latter actually played. (See Doug Farrar's Cover-3 article from last December.) Cutler did not seem to have this rapport with any other receiver during the season.

2) Related to point (1), Cutler reportedly had been lobbying for Aromashodu to get playing time all season.

4
by Led :: Mon, 08/02/2010 - 4:09pm

If Sanchez manages to perform consistent with the average, the Jets will be a pretty good team. That would probably put him somewhere in the Chad Henne/Jason Campbell DVOA range and lift the Jets offensive DVOA from a -9.0 to a zero or slight positive. I would sign up for that.

5
by starzero :: Mon, 08/02/2010 - 4:33pm

i'd really like to see jay cutler throw 37 touchdowns and 28 interceptions.

--
hail damage

6
by Nathan :: Mon, 08/02/2010 - 11:43pm

Campbell's almost unquestionably playing with a worse group of receivers, although Zach Miller might be better than anyone the Redskins have in their lineup this year.

I came here to say "you know, I'm not so sure about that." Then I Googled the Raiders receivers and yeah, they're worse. But still, it's really hard to separate their production from Russel, no? Next year will be very interesting, to say the least.

7
by tuluse :: Tue, 08/03/2010 - 1:07am

Louis Murphy passed the eye test to me in a couple games, but his advanced stats are pretty horrific. I'm curious to see how he'll do this year.

8
by Jeff M. (not verified) :: Tue, 08/03/2010 - 12:23pm

Are we still boycotting news re: a certain well-travelled QB?

12
by Braada (not verified) :: Mon, 08/09/2010 - 11:38pm

Sorry I'm so late to the party, but can someone explain the "murky status" of Donald Driver? I know he's getting older, but after having his knees cleaned out this offseason, reports suggest he looks rejuvenated.