UC Irvine student uses analytics to predict NFL playoffs

The San Francisco 49ers are a proud NFL franchise, with five Super Bowl wins to their credit.

However, it’s been 25 years since they’ve held up the Vince Lombardi trophy.

If a 25-year-old UC Irvine graduate student is right, the 49ers will taste glory again this postseason.

Matthew Littman, who is pursuing a Master of Science in Business Analytics degree at UCI, created a predictive analytics model to predict the Super Bowl LIV champion. Before the playoffs began, Littman’s model predicted (in order) the 49ers, Kansas City Chiefs, Tennessee Titans and Green Bay Packers as the teams with the greatest chance to win.

So far, the model is spot on. San Francisco hosts Green Bay in the NFC Championship game Sunday, while Kansas City hosts Tennessee in the AFC Championship.

“I’m in a couple of leagues where we try to pick the winners of games, and it always seems pretty random to me,” Littman said. “Being a business analytics student at UCI, I kind of wanted to use those skills and see if I was actually able to use some data and predict who’s going to win the Super Bowl.”

It started as a class project with fellow UC Irvine MSBA students Addy Dam and Vedant Kshirsagar. Once the class was over, Littman honed the model, which uses seasonal data from each Super Bowl season since 1966. Some defensive statistics from earlier years were missing, so Littman said he used a technique called Multiple Imputation by Chain Equations (MICE) to predict that missing data.

Littman’s findings were published on the digital platform Towards Data Science.

“Every rendition of my model has predicted San Francisco, dating all the way back to Week 13,” he said. “It didn’t matter what methods I tried or what models I was coming up with. They all predicted San Francisco.”

Key factors included the number of quarterback sacks and the turnover differential. These numbers were lower than expected for the Baltimore Ravens, who earned the No. 1 seed in the AFC playoffs with a 14-2 record. But Littman’s model predicted the Ravens to lose in the divisional round, and they did, 28-12 at home against the Titans.

The Boston-born Littman said his background as a football fan also helped him with his model, although his beloved New England Patriots were upset by the Titans in the wild-card round.

“It’s always inspiring to see our business analytics students take their skills outside the classroom to work on projects they are passionate about,” said Bryan Muñoz, UC Irvine MSBA associate director of recruitment and admissions, in a release. “Predicting a Super Bowl winner is tricky business. This is a very ambitious project.”

Littman has put his money where his mouth is, so to speak. He said he has made bets on the 49ers to win Sunday’s NFC Championship game against the Packers, as well as Super Bowl LIV.

“[The 49ers and Chiefs] will make a fantastic Super Bowl if they both end up there,” said Littman, who hopes to develop a user-friendly app for his model in the future. “That’s kind of what I’m hoping for, so that my model’s correct, but also because I want to see a good game.”
::

Support our sports coverage by becoming a digital subscriber.

For more sports stories, visit latimes.com/socal/daily-pilot/sports or follow us on Twitter @DailyPilotSport.