Trio of University of Toronto students win NFL’s annual analytics competition | CBC Sports

A trio of University of Toronto students has won what you could call the Super Bowl of analytics competitions.

The group beat out 300 other teams from around the world and walked away with $20,000 US at the NFL’s annual Big Data Bowl, held in conjunction with the league’s scouting combine, set in Indianapolis.

Hassan Inayatali, a third-year student in the engineering science department, has never played a down of contact football in his life or experienced the crush of a defender’s helmet.

But along with fellow students Aaron White and Daniel Hocevar, he was able to develop a new way to measure one of the most important things in football: getting pressure on an opposing team’s quarterback.

Currently in football, pressure on a quarterback is measured by stats like sacks, hits or hurries.

“Those outcomes are measured at the end of the play and they don’t really speak to how, for example, pressure evolves over the duration of the play,” Inayatali explains. “So what we wanted to do was essentially develop a metric that provides continuity in the sense that you can tell what the pressure is on a quarterback at every frame of the play.

“There’s still pressure when a defender is four yards away. So what we wanted to do is essentially quantify that so that teams could have a better sense of what’s going on over the duration of the play rather than just the results of the play itself.”

The NFL provided the teams with detailed player tracking data from games played during the first eight weeks of the 2021 season. Inatayatali’s group used 4,911 plays as the basis for their analysis.

Hocevar says the project’s colourful graphics that clearly demonstrated how quarterback pressure evolves over the course of a play was attractive to both teams and the event’s judges.

“Our goal with the project was to create something that’s highly interpretable,” Hocevar says. “We kind of created this colourful heat map that shows every location on the field and how much pressure there is at each location. And that seems to be something that was really popular.”

The team was able to present its project to more than 250 analytic staffers from a number of NFL teams, who recognized how the winning entry could be effective in both measuring players and on-field planning and scheming.

“One of the interesting things we found is that there are a lot of players, especially defensive lineman, that traditional metrics don’t value very much,” Hocevar said. “But when you look at everything they’re doing throughout the course of the whole play, it becomes very evident that they’re really good players and contributed a lot to their team.”

Using analytics to increase engagement

This new way of measuring quarterback pressure also translates well on to television and could soon be a part of broadcasts.

Julie Souza is the head of Sports with Amazon Web Services, one of the NFL’s broadcast partners. She was also one of the event’s judges.

“We wanted to make sure we were developing something complex that measures the intricacies of what happens on a football field,” Inayatali adds. “We also created something that football fans and coaches could see and be like, ‘OK, that makes sense with what I’m used to seeing.'”

The group, who met through the school’s sports analysis club, hope their win could lead to bigger things as the sports world continues to embrace new ways of tracking and measuring success. Inayatali has already landed a job in the analytics department of Chicago’s NHL team.

“There’s a lot of directions you could take different projects and they always present unique challenges,” Hocevar says.

“There’s so much data available and it’s really a great way of putting to the test everything I’ve learned in school and push my skills as a developer and a data scientist to the limits.”