A C Sports Analytics Intelligence Lab and the Department of Statistics and Operations Research welcomed Tim Chartier for March MATHness on March 4.
Every March, millions of people make predictions and draw up brackets for the NCAA basketball tournament. Chartier spoke to UNC students about the mathematics involved in creating media and how they can improve their predictions using analysis and mathematics.
“Bringing people in to speak not only to SAIL, but to the UNC community as a whole, about sports analytics, being such a popular topic, we thought it was a worthy endeavor,” said Conor, founder of SAIL. Kerr said.
Chartier is a professor of mathematics and computer science at Davidson College and specializes in sports analysis. A subject he teaches students bracketology – the process of creating mathematical formulas and models to predict tournament winners.
Using bracketology, Chartier and his students at Davidson created computer code that can help predict the outcome of March Madness tournament games.
Chartier pointed out that the code does not give a perfect bracket. Instead, he characterizes the predictions as “cloudy crystal ball” because the data cannot take into account possible hazards that arise during the tournament.
Chartier uses two methods: The Massey method and the Matrix Collie.
THE Massey The method examines game scores. It rewards blowouts and can be adapted to use kills only. THE Collie The method examines wins and losses. Both The methods examine the strength of the team’s schedule. Chartier encouraged students to try each method.
Students and people from anywhere can use these methods on the March MATHness website. Here, people can choose to rank the men’s or women’s tournament and choose the Massey or Colley method. Then people can adjust the ranking by weighting certain factors. Users can increase the weighting of home, away, or neutral-court games and even select a time of the season that they believe will be most predictive for their bracket.
