View Single Post
11-03-2012, 10:22 PM
Now 11.5% more Zen
Replacement's Avatar
Join Date: Apr 2005
Location: Hockey Hell
Country: Canada
Posts: 40,042
vCash: 500
Originally Posted by guymez View Post

Please explain how the Corsi is meaningless and while your at it please explain a better way to rate a QB.


Myself and many others here and elsewhere have debunked Corsi countless times. Even those that tabulate the Corsi stat were forced to realize its shortcomings and could no longer defend it thus the several revisions of Corsi called relative Corsi etc.
A couple years ago I also shot down the Team Corsi stat and I can direct you to some of my thoughts on that if you like.

Corsi is not at all a scientific or even statistically valid measure. Its a weighted derivative with arbitrary ratings involved.

I like stats, but mostly raw data, stats/min etc. Stats in otherwords that have not been manipulated by somebodies weightings and ideas about which factors are most important in a player.

No wonder we're having disagreement on this issue if you value Corsi as a statistical measure.

Next, QB efficiency ratings are specifically designed for US game, and not the CFL with one less down. As such many of the weightings used end up being less helpful in evaluating CFL QB's. The measure wasn't designed to be evaluative here.

I'll list some of the things that compromise QB ratings here.

This is of course a larger field and with more of the yds gained being inconsequential when compared to the NFL. Not all yds are created equal, much more the case here. Also that the nature of the CFL game is massive ebb and flow up the field. Whereas the NFL is much more of a field position battle of attrition. In the CFL yds attained are somewhat less important and more easily gained. So the weighting that passing yds would get in the NFL is not necessarily warranted here.

Next Interceptions. In the NFL game for similar field position reasons an interception is huge. In the CFL its essentially not the game changing event it is in the NFL. Unless you're Henry Burris and throw up 4 in a game.

TD ratio. CFL has endzones that are TWICE as large. A proper coefficient that is applicable to the CFL game would require QB's to have much better redzone performance, and TD passing numbers, to get a highscore in this game. But redzone performance isn't even an evaluation criteria. Ask yourself why not.

Next, yds/ pass average needs to be higher in this league (obviously) for a QB to get realistic weighting in this variable. Instead completion percentage(much more of a valid coefficient when applied to judging performance in 4 down football) can be less meaningful here. What good here is a series of 5yd completions. In the NFL a gain like that leads to probably moving of the sticks. Here it leads to punting as often as not. So passing % here is overrated compared to moving the chains and say average gain per throw which is arguably more important.

But if you want the readers digest version consider looking at Calvillo vs RR this year. RR has the higher QB rating but Calvillo has massacred Ray on any meaningful stat. He's tossed a dozen more TD's, over a thousand more yds, Calvillo has a superior TD throws to interceptions ratio.
But the one thing Ray has on him? Completion percentage. Thats it. A very overrated variable when applied to the CFL.

Last edited by Replacement: 11-03-2012 at 10:32 PM.
Replacement is online now