Tuesday, August 19, 2008

Ranking NFL Quarterbacks- A More Accurate Passer Rating

Before I delve into adjusting passer rating, one issue needs to be cleared up. The QB rating system, as it stands now, is not that bad. True, there can be some adjustments, two of which I will make in this post, but overall, the QB rating fairly accurately measures a quarterback's performance.

I hear pundits and commentators laugh and sneer all the time about the rating, because it requires some convoluted formula that few people understand. Well, if they were to look at the actual criteria that make up the quarterback rating instead of loking for something to complain about, maybe they wouldn't always be so quick to complain. The QB rating in the NFL is based on four different categories:

1. Completion percentage
2. Touchdown rate
3. Interception rate
4. Yards per attempt

Are any of those four categories unreasonable? Are there better categories for getting to the core of what a quarterback is supposed to do? I dunno, these four categories seem pretty sensible to me. Plus, the reason that the formula is so complicated is because the four categories produce very different numbers that cannot simply be added up. Sophisticated weighted scales and formulas must be constructed in order to give all categories equal weight.

A final defense of QB rating before I make my moderate adjustments to it: Look at the top ten QB's in passer rating in 2007:

1. Tom Brady- 117.2
2. Ben Roethlisberger- 104.1
3. David Garrard- 102.2
4. Peyton Manning- 98.0
5. Tony Romo- 97.4
6. Brett Favre- 95.7
7. Jeff Garcia- 94.6
8. Matt Hasselbeck- 91.4
9. Donovan McNabb- 89.9
10. Kurt Warner- 89.8


Now, you may have some slight qualms with a name on the list here and there, but one would be hard-pressed to argue that this list was not a pretty good reflection of QB performance in 2007.

That said, there are a couple of adjustments that can be made to make the list a little more representative of QB performance. There are three factors I see that the QB does not control that get factored into QB rating:

1. Dropped passes by the receivers lower a quarterback's completion percentage.
2. Yards after catch by a receiver give the quarterback extra yards per attempt.
3. Sacks affect the quarterback in numerous ways.

As for sacks, some of them are his fault, while others are the offensive line's fault. I will try to figure out how to distinguish between these better, but as of now, I do not have a good strategy. Therefore, I will not address sacks now.

I will address YAC and Dropped Passes. I will factor out the YAC from Yards per Attempt and factor out Dropped Passes from Completion Percentage to come to what I called QB Rating Plus. Below are the 2007 rankings for QB Rating Plus:
Rating Rating Plus
BUF Edwards 70.4 69.4
MIA Lemon 71.0 70.6
NYJ Clemens 60.9 59.4
NE Brady 117.2 116.9
IND Manning 98.0 99.7
TEN Young 71.1 70.8
HOU Schaub 87.2 86.1
JAC Garrard 102.2 102.7
CLE Anderson 82.5 83.7
PIT Roethlisberger 104.1 104.0
BAL Boller 75.2 75.6
CIN Palmer 86.7 86.1
KC Huard 76.8 77.5
SD Rivers 82.4 81.1
OAK McCown 69.4 69.7
DEN Cutler 88.1 87.5
NYG Manning 73.9 76.8
DAL Romo 97.4 97.0
PHI McNabb 89.9 89.5
WAS Campbell 77.6 78.6
TB Garcia 94.6 96.1
NO Brees 89.4 90.1
ATL Harrington 77.2 78.2
CAR Testaverde 65.8 67.7
DET Kitna 80.9 81.3
CHI Griese 75.6 78.4
GB Favre 95.7 93.0
MIN Jackson 70.8 71.8
SEA Hasselbeck 91.4 90.4
SF Dilfer 55.1 58.6
ARI Warner 89.8 90.2
STL Bulger 70.3 71.8


As you can see, the differences are not tremendous, but for several quarterbacks, the changes alter the rating enough to changes their ranking in passer rating by several spots. The biggest winners under the new system are: Eli Manning, Peyton Manning, Jeff Garcia, Brian Griese, Mark Bulger, Vinny Testaverde

The biggest losers under the new system are: Brett Favre, Kellen Clemens, Philip Rivers.

During the 2008 season, I will be studying sacks. I believe that this will provide the best improvement yet in measuring quarterback performance.

2 comments:

Unknown said...

New to your blog... its a blast to read..

Perhaps you could use sacks whether or not the QB was in or out of the pocket. While not perfect its likely the OL's fault if he gets schlacked in the pocket. If he leaves the pocket and gets steamrolled hes on his own more or less.

Kudos and keep these coming.

Mike

The Football Know-it-all said...

Thanks for the comment. I'm actually working on a stat, and it bears some resemblance to what you're saying. I'm noticing that sacks that happen within three seconds lose fewer yards than sacks that occur after three seconds.

Sacks that happen more quickly are more the line's fault, whereas sacks that occurs after several seconds are mostly the qb's fault, for not getting rid of the ball.

Therefore, I'm toying with the idea of including yards lost per sack as a metric to measure qb's, with more yards lost per sack indicating a tendency on the wb's part to get sacked.

More research is forthcoming...