Why would I do such a thing? Because every year, there are some people who like to gripe about the silliness of preseason college football polls, the premise being that no one really knows who will be good or not before the season starts. For example, last year in the preseason AP poll, Florida State was ranked #3, Michigan #11, Louisville #16, and Florida #17, and none of those teams finished ranked in the final poll. Meanwhile, in the final poll, Central Florida finished #6, Texas Christian #9, Notre Dame #11, and Michigan State #15, and none of those teams were ranked in the preseason poll.
Now, I have no problem with preseason polls myself. They're just for fun, and I enjoy looking at them. Yes, teams that are ranked highly in the preseason AP poll may have an "unfair" advantage in later polls, but that's only because many sportswriters aren't very good at rating teams. That's not the fault of the preseason poll. And while some of the preseason rankings will turn out to be wrong, for the most part preseason polls tend to be pretty close: last year, 17 teams from the preseason AP poll finished ranked, and that is a lot more than you're likely to get by simply picking teams out of a hat (on average, that will get you about 5 correct out of the 25).
Still, a few years ago I got to thinking: what if we did simply pick teams out of a hat, and then go from there each week? And we could treat all teams with a blank slate, ignoring their histories, performances from last year, returning starters, etc. We would only focus on what each team does this year, with a completely random ranking to start the season. Now, is this what I think people should be doing when they rank teams early in the season? Heck no, this is just for fun.
