“It is not based on any team’s record this year, but only on the seed each team is assigned by the selection committee, which is assumed to capture – as well as it can be captured – the quality of a team. Wins and losses are the result of chance with the probabilities determined from two statistics: the record of upsets in first-round matches over the past 30 years, and the percentage of tournaments that have ended with a 1-seed as the final champion,” Richards said.
Designed as a way to show students that ideas of physics can be applied to nontraditional models, Richards said this project started as an introduction to physics research for a high school student and has since been used by a Canadian middle school teacher.
“This began as an introduction to physics research for Michael McHugh, who was then in high school and is now a student at the University of Colorado,” Richards said. “Last year, Karen Slattery, a teacher in Ontario, Canada, used this model with some of her middle-school students, who were excited to use math, computers and scientific thinking on a problem already of interest to them.”
Although this model is not designed to help fans win their brackets, it certainly illustrates the use of physics in the sporting world.
“We didn’t design this model to help folks win, but to help get our students interested in physics and to practice object-oriented programming,” Richards said. “In fact, it can be demonstrated that the model that statistically mimics the historical record most accurately is not the best option for bracket competitions. However, changing one of the two parameters can generate results comparable with assuming the favorite always wins – a boring choice that nevertheless beats most bracket competitors.”
To learn more about Richards’ NCAA Tournament model, e-mail richardsh@marshall.edu or call 304-696-6466. For more information on programs in the Marshall University School of Physical Sciences, visit www.marshall.edu/cos.