Professor Tim Chartier works at the college library with students (from left) Miles Abbett ’14, Sarah Klett ’15, Jordan Wright ’14, (on floor) Annie Tang ’15, Seth Kindig ’14. (Bill Giduz/Davidson College)

Connecticut is basking in the glow of its NCAA basketball championship last week, while the rest of us are mostly licking our wounds from some bad March Madness bracket picks. Before the tournament, we talked to Davidson College math professor Tim Chartier about his annual March Mathness math-driven bracket challenge. How did he and students fare? Here’s his epilogue:

By TIM CHARTIER

We’ve definitely had a wild ride this March. Most brackets struggled, unless you were already a UConn fan, and such is the nature of modeling the real world. Whether using math, college allegiance, or even favorite color, one is making modeling decisions regarding the strength of teams in predicted games.

Tim Chartier

Inherently, one team, at best, has a certain probability of beating another. For instance, Nate Silver predicted that Florida had a 70 percent chance of beating the UConn Huskies to advance to the championship game. Yet, Florida lost. Was he wrong? Probably not. How?

Let’s look at it this way. If you have a coin that will flip heads 70 percent of the time and you flip 8 times and get tails all 8 times, my response will likely be, “Wow! That’s unlikely.” Yet, I will still predict that you will flip heads 70 percent of the time.

Sports contain more randomness and chaotic behavior than flipping a coin. That’s why we watch. And that’s what attracts applied mathematicians like me and the Davidson students I teach to research in it.

WE GOT PEOPLE THINKING MATH

So, let’s get to the key question – how’d we do? First, I win – big time! Our code, which ran our math methods on this year’s data, was run over 16,000 times in over 25 countries. How’s that winning? The code inherently has people thinking mathematically! That’s the MATHness in March MATHness.

But, the more pressing question is how’d our methods do? Not surprisingly, we did not have a method in the 99th, 97th or 96th percentile on ESPN’s online tournament. In my Finite Math class, the best bracket beat more than 87 percent of over 11 million brackets, with several brackets at 80 and 77 percent. My pool of students in my uDemy course (a free online course on sports ranking) had two brackets that beat more than 92 percent of the brackets. Finally, a new research method beat 91 percent.

How’d we do this when none of our brackets predicted the Huskies? It was an unexpected result so many, many brackets didn’t predict the 7th seed winning the championship. Our methods were very strong in many rounds with one method being as good as any other method (the 100th percentile) in the ESPN pool.

So, for us, it was a great year. Many people got involved, we have new research ideas, and we have a lot of new ideas for next year. We didn’t win the billion and didn’t even land as high as some years. But then again, we are modeling March Madness and expecting it to always be different would be math madness!

Tim Chartier teaches math at Davidson College and annually promotes his “March Mathness” bracket challenge. You can find out more at http://marchmathness.davidson.edu/

RELATED LINKS

March 20, 2014, DavidsonNews.net/CorneliusNews.net, “For Davidson prof, students, it’s March Mathness.”

March 10, 2014, Davidson.edu, “Prof. Tim Chartier garners national media attention for his March Mathness method” – The college reports on the annual challenge and links to coverage from around the nation.

by