Writers of Pro Football Prospectus 2008

printable version

Football Outsiders

Innovative Statistics, Intelligent Analysis

Need Help with KUBIAK or Premium?

blog advertising is good for you


06 Oct 2010

Demystifying The Passing Efficiency Stat

David Fox of Rivals.com takes a look at the passing efficiency tool, its history and its usefulness. In the process, he interviews Arizona's Mike Stoops, California's Jeff Tedford, Syracuse's Doug Marrone ... and yours truly.

As a whole, my sentiment toward the passer efficiency rating is how Fox quoted me in the article. It is better than simply looking at yards, TD's, and INT's, and it is very useful to have in the box scores, but a) it takes a while to know whether you're looking at a good score or not (while the NCAA may have once wanted 100 to reflect "average," it very much does not anymore), and b) it does not necessarily reflect good quarterback play as a whole. It could stand to be refined.

Posted by: Bill Connelly on 06 Oct 2010

11 comments, Last at 07 Oct 2010, 4:18pm by tuluse

Comments

1
by The Ninjalectual :: Wed, 10/06/2010 - 6:01pm

Huh. I never even knew this existed.

2
by The Ninjalectual :: Wed, 10/06/2010 - 6:07pm

""If we tweak the efficiency rating, you're not going to be able to compare past years with current years"

No, you would just have to re-calculate the rating for previous years, and old players would have different ratings.

3
by Overrated (not verified) :: Wed, 10/06/2010 - 6:18pm

I think what's being argued is that changing the multipliers to reflect what is "average" now is no more appropriate than using the "average" from then.

4
by drobviousso :: Wed, 10/06/2010 - 6:39pm

Or you could just run a regression to see what is correlated with winning, normalize to 100, and you have your weights.

5
by Pass to Set Up ... :: Wed, 10/06/2010 - 8:56pm

Without an understanding of why any individual metric contributes to winning, and in what context it contributes to winning, you'll probably just be overfitting your model to your past data, and it won't end up having as much predictive or explanatory power as you would think.

6
by drobviousso :: Wed, 10/06/2010 - 10:04pm

Couldn't have less than "Lets just get it around 100"

7
by Theo :: Thu, 10/07/2010 - 4:12am

I'd be more interested between a correlation between pass efficiency and NFL success.

11
by tuluse :: Thu, 10/07/2010 - 4:18pm

You might, but I imagine most people care more about their performance in college than pro potential.

8
by andrew :: Thu, 10/07/2010 - 7:25am

The article quotes Tim Clifford's 10-14, 345 yards, 5 TD performance as having a rating of 403.4, highest all time since college doesn't cap any of the factors.

But if you plug those values into the calculator they included, you get 396.29.

9
by Revenge of the NURBS (not verified) :: Thu, 10/07/2010 - 12:34pm

You forgot to add in the sales tax.

10
by verifiable (not verified) :: Thu, 10/07/2010 - 2:52pm

From the two yard line,if a QB throws a pass one yard and the receiver runs in for a TD it resultsin a 440 rating but if the receiver falls down at tfe one yard line then the QB gets a 110. This stat may be weighing TD's a bit heavily.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.

More information about formatting options