Joseph: The Problem with Polls (Oct. 10)
Allan Joseph | Monday, October 10, 2011
It’s time to face two hard facts about college football. First, the polls don’t work. Second, there’s no good way to fix them.
It’s pretty clear to anyone who watches college football that the rankings are essentially arbitrary. The preseason polls come out each year based on subjective interpretations of last year’s results ⎯ with some spring practice observations taken into account if the writers and coaches are feeling particularly motivated when they fill out their rankings.
From there, the season progresses while coaches and writers move teams up and down based on just the score of the game and possibly the highlights on SportsCenter. That leads to obviously false conclusions fairly regularly. For example, Notre Dame soundly defeated Michigan State in week three. Yet, in the week five polls, Michigan State still ranked above Notre Dame in both the AP and coaches’ polls. The Spartans fell from No. 15 to No. 20, but it’s absurd to argue that they are better than a team who beat them soundly just two weeks ago.
When it comes down to it, the polls rarely, if ever, actually reflect an idea of who the top 25 teams in the country are. No matter what, a loss drops a team in the polls. It doesn’t matter whether that was a last-second loss to the top-ranked team in the country – the losing team will sink in the polls. There’s no reason for that.
The rankings are supposed to suggest relative rankings. Teams are supposed to lose to higher-ranked teams ⎯ it should be no surprise when the No. 1 team wins. The No. 1 team beating the No. 10 team doesn’t mean the No. 10 team should drop in the polls, unless the game was a blowout. If the No. 10 team gave the No. 1 squad a close game, in fact, it should rise in the polls. But it doesn’t.
The polls never actually give a true representation of who the best 25 teams in the country are, and in what order. That’s the entire point of having the polls. The system is broken.
What’s worse, human polls make up two-thirds of the Bowl Championship Series formula that determines who plays in the national championship. This stuff matters.
That’s why it’s time to figure out a way to fix the polls. Obviously, a computer-only system is never going to be acceptable. Fans, players and coaches alike want to feel that there are humans in control of the process.
But the problem is that there’s no good way to make the polls better. It’s unreasonable to expect coaches (or really, the assistants who fill out the polls) and writers to do in-depth analysis of every game. That’s simply impossible.
So the answer is to make the polls matter less, while making football matter more.
That’s right, a playoff system.
It wouldn’t be perfect, certainly. A playoff would still depend on rankings to select its teams. Even the most expansive proposal, a 16-team playoff with automatic bids for every conference champion, relies on the polls to fill its at-large slots.
It might not be perfect, but expanding to a playoff system keeps the polls from doing as much harm as possible.
Sure, it’ll be difficult for the polls to decide between the No. 8 and No. 9 slot. But it’s a better decision to make than whether to leave an undefeated No. 3 Auburn team out of the national championship. Having the top four, eight or 16 teams play in a playoff would make the process fairer to everyone involved. If a team thinks it is good enough to win the national championship, it can simply play its way to that championship without having to impress voters.
So while those two facts about the polls are difficult, there’s one more fact to consider: a playoff would be a lot of fun.
Contact Allan Joseph at email@example.com
The views expressed in this column are those of the author and not necessarily those of the Observer