Tuesday, August 21, 2012

A brief rant about poll philosophy

Lane Kiffin and Larry Scott and various other people have pointed out in the past week that the coaches' poll is fundamentally flawed because of its biases and coaches' lack of knowledge about other teams and so on and so forth. All those people are right, especially considering that the coaches' poll is one of the two (along with the Harris Poll) that actually matters in terms of deciding stuff at the end of the year.

The AP poll doesn't have that latter problem but provided the basis for this rant just by virtue of the public nature of its ballots, which are presumably assembled in largely the same way (from a philosophical standoint) as the coaches' ballots.

As you may or may not remember, Oklahoma had one first-place vote in the AP poll; it came from Kyle Meinke of AnnArbor.com. Sooner Nation called the guy up for an interview to find out exactly why somebody put a team other than LSU/Alabama/USC at the top, and while most of the answers were of the "polls are meaningless yadda yadda yadda" variety, he said two things that cumulatively are producing FIERY FIRE in my eyeballs.

The first:
I love how the schedule shakes out for Oklahoma. Kansas State is home, Notre Dame is home, Baylor is home. Among its first nine games, the most difficult might be playing Texas in Dallas, and OU will be favored in that one as well. It's easy to see the Sooners going 9-0 before that season-ending stretch against West Virginia, Oklahoma State and TCU. And by then, I think any issues they have up front offensively will have been addressed. 
Here's the thing: Schedule ease should not be a consideration in a preseason poll, which (IMO) is supposed to be a ranking of the best teams rather than a projection of the order in which they're going to finish. I realize that this is a somewhat subjective issue; if somebody wants to vote the latter way and is willing to totally throw out all preconceived notions once there are some on-field data points, I'd be willing to listen to his/her reasoning and consume the in-season ballots accordingly.

But here's the second quote:
If they struggle against UTEP or Florida A&M, and USC wins at Stanford, I'd have to give the Trojans a long look. Likewise, if Alabama beats Michigan in its opener and then Arkansas in Week 3 -- a pair of top-10 teams -- I'd give them a long look. But again, I think Oklahoma is the best team and I likely would keep them No. 1 if they take care of business as they should.  

You can't say "I think Team X will finish highly partly because the schedule is so easy" and then say "I won't drop Team X as long as they win all those games that I just got done saying are so easy." Doing the second makes the first a self-fulfilling prophecy: Of course a team will finish ranked highly if all/most of its games are easy and you refuse to drop them on your ballot as long as they win said easy games. That skews the poll inexorably in the direction of whatever preconceptions are used to assemble the preseason ballot (even though those are supposed to gradually be removed once enough data is in place to do so).

I'm not picking on Meinke because I think he's stupid but because I'm confident that his reasoning is widespread -- just check out any rankings explanation anywhere for verification -- and thus skews the poll results on a macro scale. I don't like it. Stop it, plz.


Post a Comment

Powered by Blogger.