Ahhh! It’s that time of year again. March Madness, the three week odyssey in men’s college basketball that is sure to have more drama than an entire season of The Bachelor. It’s a time where office pools are the norm and everyone from a complete novice to a basketball “expert” tries to make sense of the pure craziness of the NCAA tournament.
The strategies for choosing a winner vary. Some people pick winners based on school or conference allegiances, mascots, school colors, fondness for favorites or upsets, or just gut feeling. But for one University of Minnesota biostatistics professor, the choices for his picks are based on statistical analysis, with a bit of game theory thrown in.
How do you apply what you teach in your School of Public Health biostatistics courses to make your NCAA tournament picks?
I’m able to use a lot of what I teach in my courses during March Madness. The outcome of a college game is normally distributed (i.e., according to a bell curve) with center equal to the point spread. That means half of the time the favorite in the game covers the spread, and half of the time it doesn’t. Using data from past games, we know how wide the bell curve is, so we can compute the probability of the favorite winning any particular game in the tourney. For games later than the first round, we can guess what the point spread will be using differences in team computer ratings, like those Jeff Sagarin prepares for USA Today.
In your opinion and experience, how much more accurate are your picks compared to other methods?
To give you the best odds of winning your pool, you want as many people who pick winners using unscientific methods (like team colors) in your pool as possible! Using statistical information my picks are generally more accurate and give me a better chance to do well in my bracket.
Does this give you an advantage over others?
In years past, I was almost the only person that had access to computer programs that used statistical analysis to pick winners, giving me an advantage over other players. As the years have gone on, these programs are now available online (see for example the Poologic site), so my advantage has greatly decreased!
How should you pick your upsets?
Picking some upsets is a must during the NCAA tournament, especially if you’re playing in a larger pool, since you want your bracket to look somewhat different from that of most of your competitors. You might want to pick a couple 2, 3, or even 4 seeds in the Final Four. In earlier rounds, 12 seeds are popular with some players since they tend to defeat 5 seeds at a higher rate than would be expected (one or two 12’s seem to get through each year). Picking slightly “against the grain” can give you a slim advantage by separating you from your competition.
How do you use the human element against your competition?
During the NCAA tournament especially, many people base their picks, at least in part, on their personal biases and emotions, ignoring what the numbers tell them to do. I actually do not watch many of the games before the tournament, so that the numbers will speak for themselves and help prevent me from getting too wrapped up in the emotional side. Still, you may want to use the emotion of your opponents against them; that is, you may want to pick a few teams that are generally disliked by the masses (i.e. Duke, North Carolina, Kentucky, etc.) to do well in the tournament. Picking against Big 10 teams may also be wise if you play with a lot of Big 10 fans.
Who do you think is going to win it all?
I wouldn’t hazard a guess, but all indications are that this year will be particularly challenging; all the top teams have several losses and there is no dominant team. For instance, the latest Sagarin ratings show that the nation’s #1 team (Louisville) is only 10 points better on a neutral court than the nation’s #30 team (the Gophers). Let the Madness begin!
For more on Carlin’s process for picking his winners, watch this video. Good luck on your picks and enjoy the NCAA tournament!