November 25, 2010

Adding Rushing to Quarterback Rating: In the New York Times, Luis Deloureiro is working on a new quarterback stat that includes a way to measure mobile threats like Michael Vick. The formula is (passing yards + rushing yards - sack yards)/(pass attempts + sacks). Vick's 8.8 almost doubles second place finished Ben Roethlisberger at 4.6. Philip Rivers (4.4), Vince Young (4.2) and Aaron Rodgers (3.7) round out the top 5.

posted by rcade to football at 07:10 PM - 16 comments

I'm confused on the argument..I thought passer rating was primarily used as an indicator to show how well a QB is playing.

Now I understand the desire to introduce RUSHING into that scenario, but based on his approach is he saying a Vince Young(4) and Josh Freeman(6) are playing at a higher level at the QB position than Brady (8)?

If so, I think he needs to go back to the drawing board.

posted by bdaddy at 09:24 PM on November 25, 2010

Useless

posted by sgtcookzane at 09:35 PM on November 25, 2010

It also doesn't take TDs and/or INTs into account. That seems like a problem, though I'm sure if the current QB rating does either.

posted by jmd82 at 12:02 AM on November 26, 2010

So a rush doesn't count as an attempt? So a 30 yard scramble gets put into the total but not in the denominator?

OK...

As a modification for one of the four sections of passer rating, sure, go ahead and put this calculation in (remove the foolish thresholds and TD pass percentage crap while you're at it)... it'd improve it. But by itself it's even more flawed than most stats. Many of the names above Manning by his measure amuse me.

posted by Bernreuther at 12:14 AM on November 26, 2010

Why not just add rushing yards to the QB passing stat. 80 passer +20 yards rushing = 100.

posted by Jackjeckyl at 03:08 AM on November 26, 2010

Passer rating is a pretty meaningless statistic as it is. Compared to many stats in baseball, there does not appear to be any rhyme or reason on how things are combined and the ratio in which they are combined. The modifications he's suggesting are even more stupid. Just because you can combine two statistics mathematically, doesn't mean you should, nor does it mean that it will have any actual value.

Heck, using this formula, any running back who makes one pass attempt is going to have a stupid high rating.

While I appreciate the effort, the results are less than meaningful, because it doesn't include carries in the denominator.

posted by AaronGNP at 08:03 AM on November 26, 2010

Finally, a stat to keep FIFA's world rankings company.

posted by Mr Bismarck at 08:24 AM on November 26, 2010

I think the NFL's passer rating is a nice gauge of how efficient the QB is playing. I believe Tom Brady had a perfect passer rating against the Lions. I don't understand how, he was like 21 of 26 but still got a perfect rating. While he had an awesome game, it wasn't perfect as his rating would indicate.

posted by sgtcookzane at 08:47 AM on November 26, 2010

It's probably one of those things like errors in baseball: you don't call an error on a ball that a fielder never had a chance for. Likewise, I'm guessing that incompletions somehow don't count if the quarterback really never had a chance to connect (although who judges that, I don't know), but got rid of the ball rather than get sacked for a loss.

posted by lil_brown_bat at 09:16 AM on November 26, 2010

Likewise, I'm guessing that incompletions somehow don't count if the quarterback really never had a chance to connect (although who judges that, I don't know), but got rid of the ball rather than get sacked for a loss.

Or when Brandon Tate lets a perfectly thrown deep ball bounce off his hands.

posted by Ying Yang Mafia at 09:29 AM on November 26, 2010

Passer rating doesn't take into account drops, it just calls it an incompletion.

Interceptions are what kill passer rating, to the point where literally throwing every pass into the turf at your own feet will give you a better rating than if you complete 10 passes but throw two interceptions.

Going 10/20 for 150 yards 0 TDs and 2 Ints = 35.4 Rating. Going 0/20 for 0 yards 0 TDs and 0 Ints = 39.5 Rating.

Passer rating is a decent indicator of who's good and who isn't over the course of a season or career, but in individual games if people do very well or very poorly it gets a bit squirrely.

posted by Mr Bismarck at 10:40 AM on November 26, 2010

I don't understand how, he was like 21 of 26 but still got a perfect rating. While he had an awesome game, it wasn't perfect as his rating would indicate.

From Wikipedia: A perfect rating requires at least a 77.5% completion rate, at least 12.5 yards per attempt, a touchdown on at least 11.875% of attempts, and no interceptions.

Basically, any values HIGHER than those bolded are not going to add to the final score.

This is why I have a problem with the QB Passer Rating system as a way of judging quarterbacks. However they do the calculations, they need to allow for more spread. If two quarterbacks have identical numbers in a game, except QB Dave threw for an extra 20 yards more than QB Bob, then I believe QB Dave should have a higher rating.

Come up with some other calculation, and then set up a similar comparison system like OPS+ and ERA+, where an "average" game is listed with a score of 100, while great games could have scores of 150 or 200 (while terrible ones could be 50 or 25).

posted by grum@work at 12:00 PM on November 26, 2010

Football Outsiders already do something similar with DYAR - a rating that is adjusted for the quality of the defence faced and then measured against what an theoretical replacement would have achieved and DVOA which is similar, except measured on a "per-play" basis against the average.

The FO stats also include sacks and aborted snaps and count fumbles against the QB even if his own team recover the ball.

PlayerRatePlayerDYARPlayerDVOA
M Vick108.7T Brady1,224T Brady+43.8%
T Brady105.8P Manning1,117B Roethlisberger+40.2%
P Rivers105K Orton1,063M Vick+35.2%
B Roethlisberger101.9P Rivers1,052P Rivers+33.2%
V Young98.6A Rodgers1,007A Rodgers+32.3%
D Garrard98.5M Ryan1,003K Orton+28.5%
M Cassel96.2D Brees967M Ryan+28.3%
A Rodgers95.7M Schaub704P Manning+26.1%
T Romo94.9J Flacco695V Young+24.5%
K Orton94.5C Henne566T Romo+23.7%
So having played the defences he has faced so far this year, Matt Ryan is about a thousand yards ahead of where Joe Backup would be and each individual play is likely to be 28% more successful with him under Center.

Conspicuous in their absence from the QB Rating top ten are Manning, Brees, Ryan and Schaub. Garrard is sixth for rating, but 23rd for DYAR and 19th for DVOA.

posted by Mr Bismarck at 02:14 PM on November 26, 2010

Also conspicuous (though not surprising) by their absence from all three of these rankings, MrB, are all the 49ers and Raiders QBs, plus Manning the younger. [Homerism alert]

posted by billsaysthis at 02:34 PM on November 26, 2010

Interceptions are what kill passer rating, to the point where literally throwing every pass into the turf at your own feet will give you a better rating than if you complete 10 passes but throw two interceptions.

Interceptions can also be a misleading stat if a perfectly thrown ball bounces off a receiver's hands and ends up being intercepted.

And how about 4th quarter comebacks? Do they factor in at all?

posted by cjets at 10:54 AM on November 27, 2010

To slightly counterbalance, the QB Rating also doesn't count when a QB hits a defensive back in the hands and they drop the ball. I'm looking at you, Dallas secondary.

posted by Mr Bismarck at 03:43 PM on November 27, 2010

You're not logged in. Please log in or register.