Skip to main content
Why Analysts Can’t Predict the Super Bowl

Predicting the outcome of a sporting event is difficult science. In the U.S., January brings Super Bowl predictions. Sports Illustrated's Peter King predicted the Broncos versus the Packers in the Super Bowl (after both teams lost this past weekend, we now know he was very wrong).

And analytics folk hero Nate Silver cast his prediction into the mix in an interview on ESPN’s “First Take” on January 10. Coming off his remarkably accurate U.S. Presidential election prediction, he cited data from a football analytics site, “Football Outsiders,” which aims to bring objective analysis to football. Silver predicted that the Patriots would meet the Seahawks in the Super Bowl. Since, the Seahawks lost to the Falcons last Sunday he is now half wrong approaching the Super Bowl on February 2.

So, if predictive analytics can be used to accurately predict the outcome of the presidential election, why can’t it work in predicting the outcome of a sporting event? Because there are too many variables creating too much noise in the system.

There are 53 players on the roster and they can be affected by injury (reported or unreported), emotions (first time in a Playoff game) and errors (dropped balls, missed cues). And not only do teams change during the off season, they change during the season. A team that plays in September is often different than the one that makes it to the post season.

The weather can also be a factor -- some teams thrive in rain, cold and snow, others shrivel. Wind affects the physics of throws and kicks. Creative offensive or defensive coordinators introduce a diverse amount of plays and schemes meant to make the opponent off balance. Penalties can change the outcome of a game.

And this is compounded by the fact that there are only 16 games in season, and at most each team will only have 13 opponents. Some playoff teams have played each other, but most have not. And there aren’t even a lot of common opponents between them.

On the bright side, Silver was more accurate than Peter King at Sports Illustrated. Of course there is still another weekend of games…

related posts