Laura Albert McLay

For sports fans, March Madness means filling out an NCAA college basketball tournament bracket. For Laura Albert McLay, associate professor of industrial and systems engineering at UW–Madison, it’s a chance to apply her work in data and analytics. This semester, McLay is teaching an upper-level course, Discrete Optimization and Modeling Techniques, while also exploring unique data topics — including bracketology — via her blog, Punk Rock Operations Research.

What mistakes do people make when they pick their teams?
The 7–10 and 12–5 matchups sometimes produce upsets more often than not. But don’t worry too much about the early rounds — the scoring rewards getting it right in the later rounds. Another common mistake is to pick too many No. 1 seeds in the Final Four. You can use some game theory [instead]: pick different teams in the Final Four than everybody else to differentiate your bracket. It’s hard to win if everyone picks the same teams.

What’s your number one tip for filling out brackets?
My top tip is to check out various rankings methods based on math modeling. My favorite ranking model is the Logistic Regression Markov Chain model, but others are good, too. Math-based ranking methods will help get you in the ballpark and will identify some likely upsets to pick. Math modeling has limitations, but you can then supplement your bracket picks with other information, like preseason rankings and injury reports.

Has bracketology work been an engaging way to connect to students?
It’s been a great way to connect material we are learning about in class to a real-world application. For many students, sports analytics is a great avenue for introducing how to do mathematical modeling and make data-driven decisions. Sports can really help a topic “click” for certain students. And it’s a lot of fun for me.

What kinds of projects are you and your students working on?
My research studies discrete optimization applications for the public good. I’ve worked on homeland security and emergency-response problems, anything from ambulances and fire engines responding to 911 calls to post-disaster emergency response. We live in a world with limited resources, and effectively using those scarce resources is sometimes the difference between life and death.

What sorts of topics do you enjoy exploring in your blog?
One of my favorite posts is about vampire population growth inspired by Markov chain modeling I was teaching in class. The theoretical solutions to the model were that either the vampire population dies off or it explodes. I was a little bit skeptical about the seemingly stable vampire populations in movies, so I wrote about that, and that post went viral. Other popular posts have been about zombies, election forecasting, being struck by lightning, and how to optimally snow-blow your driveway.

Interview conducted, edited, and condensed by Daniel McKay x’16.

Published in the Spring 2016 issue