Why Polls Don't Matter

These Poles don't matter, either.

There's been a lot of bitching and moaning about the college basketball and football polls lately. Kentucky ranked ahead of Vandy at the end of last football season even though Vandy beat Kentucky (and a couple other pretty good teams). Kentucky still isn't ranked in the ESPN/USA Today Coaches Poll and two six-loss teams are, including Notre Dame, who has a 3-5 conference record. Duke is somehow number one in this week's poll by virtue of the fact that they are Duke. Pre-season college football polls have way too much influence over the in-season rankings. (Aside: I love how Mark May et al. at ESPN fumes for a ban on polls until mid-October in college football, but ESPN actually sponsors said early rankings and pimps out games for ranked teams. End aside.) Joe Lunardi has Team X at **fill in the blank** this week in his bracketology thingy. Bracketology, the made up science of who won last week!

While all of the above observations are true, and Tru at A Sea of Blue thinks that polls matter, I'm prepared to argue that they do not. At least not right now. The AP poll takes the opinions of sports writers all over the country. First, consider what a sports writer does for a living: writes and watches TV with the occasional post-game interview thrown in. The sports writer does not play sports, he does not coach sports, and he certainly cannot watch every game every night, as a lot of teams don't even play on TV. You know who plays on TV? "Top 25" teams. That's who the lazy writers see, and so that's who they rank. The teams they don't see are judged (inaccurately) by linescores. No way some writer is actually going to take the time to evaluate whether a team is statistically or even subjectively better than some other team before putting his poll ballot in every week. Will not happen.

Then there's the even dumber poll, the ESPN/USA Today Coaches Poll. Now, in theory, this poll might make better sense than the AP poll because its voters are comprised of people that actually know a little bit about basketball. Moreover, they know about other teams because they have to game plan for them all the time. But outside of their own schedule, these coaches probably do not know much about the 300 some-odd basketball teams in Division I. Billy Gillispie spends approximately 20 hours a day during the season with practices, games, film sessions, planning session, recruiting, and reconnecting with Rick Pitino's hot women-on-the-side. He does not have time to watch other teams' games unless they're the next opponnent (and, given the exclusion of the SEC from the polls this year, the opponent is not in the Top 25). Most other coaches are in this same mode of living during the season. So their poll votes consist of looking at who's in the Top 25, seeing if they won or lost last week, and moving the teams around accordingly. That's it.

The point is, people get all bent out of shape about the opinions of people who probably know about or watch less college basketball than they do. A people's poll might make better sense, then, but knowing Kentucky fans and their improbable love of the internet, Kentucky would rank in the Top 5 if they had lost every game for the past 3 seasons, so that wouldn't be accurate either.

The most accurate indicator of how good a college basketball team is in a given week is its performance against competition, not polls. Polls should be taken every week with a clean slate. Did Duke really play better than Team Z last week, or did their competition just really suck? Should Duke be ranked so high when they never actually play anyone of substance away from their home floor? Is Duke a closet farm for a fascist world takeover? (Definitely.)

End rant. Enjoy your new Lunardi bracketology bracket, which will not look anything like it does now in three weeks.