September is here, which means crisp fall air, pumpkin spice everything — and countless hours of football.
But what if I told you that the more college games you watch this year, the less you'll actually know about who the best teams are?
As it happens, when it comes to predicting success in bowl games, preseason polls are actually a better predictor than end-of-season polls. That means right now, before any games have even been played, the collective wisdom of our sportswriters and coaches is actually better than it will be come November.
Look at this data from Ed Feng, who runs the sports analytics site The Power Rank. His figures show that forecasts of team success before the season kicks off are more accurate than when it winds down.
It suggests that voters tend to get it right before the first kickoff, relying on instinct and a loose sense of talent, rather than actual game results. It's almost as if actually watching real games fools people into seeing trends that don't really exist. Poll voters are developing an overinflated sense of a team's talent, and expecting a level of performance that might not be sustainable.
(Just for those who don't know: Pre-bowl polls are at the end of the regular season, right before the bowl games start. Closing market spread refers to the betting lines set at gambling houses right before the games kick off.)