ShareThis Page
Pitt professor’s model predicts Patriots Super Bowl win; he hopes it’s wrong |

Pitt professor’s model predicts Patriots Super Bowl win; he hopes it’s wrong

Aaron Aupperlee
| Wednesday, January 31, 2018 3:06 p.m
Konstantinos Pelechrinis, an associate professor at University of Pittsburgh's School of Computing and Information and head of the Network Data Science Lab (Photo from University of Pittsburgh)
Getty Images
ST PAUL, MN - JANUARY 29: Tom Brady #12 of the New England Patriots speaks to the media during Super Bowl LII Media Day at Xcel Energy Center on January 29, 2018 in St Paul, Minnesota. Super Bowl LII will be played between the New England Patriots and the Philadelphia Eagles on February 4. (Photo by Hannah Foslien/Getty Images)

A University of Pittsburgh professor really hopes his statistical model got its Super Bowl prediction wrong.

But he’s pretty certain it’s going to be right.

Konstantinos Pelechrinis, an associate professor at Pitt’s School of Computing and Information and head of the Network Data Science Lab, said his model picked the New England Patriots to beat the Philadelphia Eagles, 25 to 22.

“I really hope I’m wrong,” the Pittsburgh Steelers fan said Wednesday.

The model isn’t too far off the Las Vegas odds, which favor the Patriots by 4.5 points in a 48-point game.

Pelechrinis’ research looks at the importance of factors such as turnovers, penalties, passing yards, rushing yards and the balance between them. To determine who might win, the model considers how each team did against teams similar to its opponent. While evaluating how the Patriots stack up against the Eagles, the model put more stock in the outcome of New England’s game against Pittsburgh than its game against the New York Jets, Pelechrinis said.

The model then simulates the game several times. By several, Pelechrinis means 10,000, each time changing the performance indicators such as passing yards and penalty yards for each team.

Pelechrinis said his model gave the Patriots a 61 percent probability of winning the game, a number he feels very confident is right.

“That doesn’t mean they will win,” Pelechrinis said.

Pelechrinis came to Pitt in 2010. A sports fan, he wanted to use data to get a different look at the game. Pelechrinis also thought teaching data analytics using sports would keep the attention of his students. He has developed similar models for basketball.

“Data can give you a different perspective of the game,” Pelechrinis said.

Pelechrinis uses data to gain a deeper understanding of more than just sports. His work with metrics and models extends to cities and people. Pelechrinis is working on the developing models and algorithms for smart cities and looking at what social networks can tell us about the behavior of people.

He co-authored a study showing that systems like Pittsburgh’s Healthy Ride bike-share program contribute to neighborhood property value increases of about 2.5 percent and worked on a research project that showed Healthy Ride decreased car trips and the demand for parking spaces. He’s worked on algorithms that can map the shortest and safe route to walk home. And he’s written about what technology cannot do for smart cities.

Pelechrinis and fellow professor Alexandros Labrinidis lead the Pitt Smart Living Project. The project seeks to use data and information to improve transportation and cities. Large screens around the city showing real-time transportation times and options are one result of the project.

His football model is far from perfect. It correctly predicted 90 percent of the games during the 2016-17 playoffs and only 70 percent of the games during this year’s playoffs. It didn’t get the Pittsburgh-Jacksonville game correct, but it was closer than other models. Pelechrinis said his model predicted the Steelers had a 61 percent probability of winning. Other models, ones used by ESPN and other national outlets, gave the Steelers an 80 to 90 percent chance of winning.

And speaking of that game, Pelechrinis used his model to provide a few retrospective tips to the Steelers staff. Remember the last big play of the game, when wide receiver JuJu Smith-Schuster scampered down the sideline and cut back toward the middle of the field to gain more yards before being tackled? The inbounds tackle forced the Steelers to use a timeout, a precious commodity in Pelechrinis’ model. If Smith-Schuster would have continued down the sideline and been tackled out of bounds, stopping the clock, saving a timeout but not gaining as many yards, the Steelers would have had a better shot at winning, Pelechrinis said his model determined.

“That would have given the Steelers a 5 percent (higher) win probability,” Pelechrinis said.

Pelechrinis said most teams now have people on staff running and developing statistical models to help predict outcomes. He has offered his model to teams in the past but found no takers.

Pelechrinis had one last bit of free advice for the Steelers from his model. He evaluated the likelihood of each play scoring points and found that every no-huddle play scores on average 0.7 points and every play preceded by a huddle scores 0.4 points.

So stick with the no-huddle next year.

Aaron Aupperlee is a Tribune-Review staff writer. Reach him at, 412-336-8448 or via Twitter @tinynotebook.

Aaron Aupperlee is a Tribune-Review staff reporter. You can contact Aaron at 412-320-7986, or via Twitter .

Categories: NFL
TribLIVE commenting policy

You are solely responsible for your comments and by using you agree to our Terms of Service.

We moderate comments. Our goal is to provide substantive commentary for a general readership. By screening submissions, we provide a space where readers can share intelligent and informed commentary that enhances the quality of our news and information.

While most comments will be posted if they are on-topic and not abusive, moderating decisions are subjective. We will make them as carefully and consistently as we can. Because of the volume of reader comments, we cannot review individual moderation decisions with readers.

We value thoughtful comments representing a range of views that make their point quickly and politely. We make an effort to protect discussions from repeated comments either by the same reader or different readers

We follow the same standards for taste as the daily newspaper. A few things we won't tolerate: personal attacks, obscenity, vulgarity, profanity (including expletives and letters followed by dashes), commercial promotion, impersonations, incoherence, proselytizing and SHOUTING. Don't include URLs to Web sites.

We do not edit comments. They are either approved or deleted. We reserve the right to edit a comment that is quoted or excerpted in an article. In this case, we may fix spelling and punctuation.

We welcome strong opinions and criticism of our work, but we don't want comments to become bogged down with discussions of our policies and we will moderate accordingly.

We appreciate it when readers and people quoted in articles or blog posts point out errors of fact or emphasis and will investigate all assertions. But these suggestions should be sent via e-mail. To avoid distracting other readers, we won't publish comments that suggest a correction. Instead, corrections will be made in a blog post or in an article.