Student predicts race outcomes

Eric Andreoli is a self-proclaimed lover of Excel models.

He knows it’s kind of dorky, but the Kelley MBA student doesn’t care.

Last spring, he spent hours building tables of Little 500 data. Andreoli wanted to create a database of Little 500 results to build a race day prediction. He used the data from four Little 500 events: Individual Team Trials, Qualifications, Team Pursuits and Race Day. Miss ’N Out results were not included. Since times aren’t taken during that Spring Series event, the data cannot be quantified. 

The idea was first conceived during a Kelley School of Business class on Excel with professor Wayne Winston. Winston has compiled stats for nine years for the Dallas Mavericks, owned by Indiana alumnus Mark Cuban.

Andreoli said the combination of Winston’s class and a roommate who was a former Little 500 rider sparked his interest in creating the predictions.

“My roommate and I always talked about the race,” Andreoli said. “We’re just big race fans. I just wondered if there was enough data to do Winston-eqsue predictions.”

It turned out there was enough data — 10 years worth. Listed on the website for the Indiana University Student Foundation, the student organization that runs Little 500, was data Andreoli could use. However, the PDFs listed results differently every year, leaving Andreoli to do a few hours of extra work to clean up the data.

“Sometimes it would be ‘team name, dash, rider name, dash, time,’” Andreoli said. “Sometimes ‘team name’ and ‘rider name’ would switch and sometimes they’d be in three separate columns.”

After a few weeks, Andreoli was able to put his data to the test when he made his race day predictions for the men’s race.

“In the actual (results) versus the predicted, the downside is it only got one completely right,” Andreoli said. “The good is within one spot, it got 11 teams right and within two spots, it got 17 teams right. So it does a good job predicting the area where teams will be.”

Andreoli’s next step is to take the formula to a higher level and see if the race is affected by the number of rookies versus veterans on a team while working to build a database for the women’s side.

“The database can’t account for things like (the Cutters’ Eric) Young being on the last lap,” Andreoli said. “But it does a good job with a generalization of how deep a team is and most likely where they’ll end up.”