The NCAA basketball tournament has entered the national consciousness to such a degree it seems like everyone has filled out a bracket and has a favorite to win. While most of the predictions are based on gut feel and emotions, in the last couple of years data scientists have gotten into the act. Even my own employer.
While everyone is trying to figure out who will win the tournament, Brigham Young University statisticians have built a model which predicts which teams would create the largest TV ratings. With many popular teams – including Ohio State and Duke – already out of the tournament, the conventional wisdom is that TV ratings will be down. However, the BYU model predicts the highest Final Four ratings would come from Dayton versus Virginia (17.2 million viewers) and Michigan versus Arizona (16 million viewers). A Dayton vs. Michigan final would draw an estimated 22.2 million TV viewers.
The model suggests the impact of an underdog is significant. A Final Four game featuring a Cinderella team or a smaller market school would have 35% more TV viewers than a game featuring two national powerhouse schools. As one of the statisticians explained,
The Cinderella teams, with all the national media attention they get, become a national star. It’s not that these schools have an established national fan base, it’s that the NCAA tournament celebrates the Cinderella more so than other sports.
Of course, underdog status isn’t the only thing that influences ratings. Expectation of a close game, the star power of the coach, and the team’s overall appeal also matter. With Dayton and Virginia losing this weekend, that’s probably a relief to advertisers everywhere.
[30 March 5 PM update: All four teams from the BYU model are out of the tournament. It will be interesting to see if this year’s Final Four can surpass 2013 ratings which averaged 15.7M viewers.]