A few years ago, Thomas Finholt, a professor and Dean of Michigan’s School of Information, had what he termed an “accidental conversation” with someone in the athletic department about a new device called Catapult.
Catapult is an Australian-based company that produces small wearable devices that track data metrics during athletic practices and competitions. The athletes generally wear the Catapult in their jerseys between the shoulder blades, while football players wear them in their shoulder pads. It can tell, for instance, how fast or how high an athlete is jumping, and how quickly the athlete is changing direction.
Several of Michigan’s teams have been using the devices since 2018, and Finholt’s conversation launched a collaboration between the athletic department and students interested in sports analytics.
“I realized that was a whole different lens into the process and the competition,” said Finholt, who hopes Michigan will eventually offer a degree in sports analytics. “I’ve tried to nourish that as much as I can, because I think it’s a perfect storm of Michigan’s historic interest in competition and winning and the talent level of our students to be able to contribute to that.
“It’s a no-brainer. You want to bring those two sides of the house together.”
Finholt and many of the students, who have since formed the Michigan Sports Analytics Society (MSAS), are working hand-in-hand with Michigan’s football, men’s and women’s basketball, field hockey, men’s and women’s lacrosse, men’s soccer and ice hockey teams. Students are assigned different sports and examine the Catapult data for specific information as requested by coaches and trainers.
“It helps them learn more about the team and the players and understand things on a more complex level than what was available, let’s say, in the ‘80s,” said Michigan student Ben Riela, the incoming president of MSAS. “The sport is still the same, the way the game is played is still the same, but now with this whole technology revolution, there’s a ton of untapped potential in learning more about how the players are moving.”
First-year Michigan quarterbacks coach Matt Weiss was hired from the Baltimore Ravens, where he was, among several coaching positions, a football analyst. More recently he was the Raven’s running backs coach, but he brought to Michigan an eager interest in evolving the team’s use of analytics. The Ravens are considered one of the NFL teams that has most aggressively embraced analytics and led the league in 2019 in fourth-down conversions.
“I definitely wouldn’t say I specialize in analytics,” Weiss said in a recent “In the Trenches” podcast with Jon Jansen. “I love good ideas. I love the culture of teaching and learning. (Ravens coach) John Harbaugh is like that, too. That’s something for sure we’re looking to bring here and do more of here.”
Weiss made clear the desire to work with Finholt and his students.
“There’s a great untapped resource here. The players and the coaches aren’t the only talented people here,” Weiss said. “We have a student body that’s really smart, really talented with computer science majors, math majors, a lot of people want to work in sports. There’s ways we can tap into that.
“(Finholt is) going to help us with the science of some of our projects, whether it’s fourth-down decision-making or looking at trends in college football or looking at recruiting analytics, who you should be going after and maybe who you shouldn’t. All those things are possible. it will be fun to grow it and see where it goes.”
Finholt and the students want to deepen the relationship between analytics and college football, with coaches, he said, who tend to be a “little more conservative.”
“Guys are bringing with them experience from the Ravens that is shaping what they want to do here at Michigan,” Finholt said, referring also to first-year defensive coordinator Mike Macdonald, who spent the last seven seasons with the Ravens. “It’s a great thing because I think that’s the way of the future in terms of, really, two things, how one approaches the preparation of the student-athletes for competition and ensuring they are fit and ready and will not be vulnerable to injury, and then, there’s all the X’s and O’s stuff, the tactics. Our principle involvement has been more of the former and less of the latter.”
Baseball and analytics have been working hand in hand seemingly since the late 1970s when Bill James began writing about the sport and statistics and his sabermetrics approach. Then there was Michael Lewis’ book Moneyball, later made into a movie, that explored Oakland A’s general manager Billy Beane and how he built a successful low-budget team using analytics.
College football coaches, as Finholt indicated, had been more reluctant to join the analytics table, wanting to rely on the feel of the game and intuition for decision making. The analytics, however, are not meant to overshadow and dictate but to enhance.
“That’s a really big thing,” said Tyler Fuelling the past president of MSAS and recent UM graduate with degrees in data and computer science. “I feel like somebody would get hung up on that, that analytics is removing the human element and at odds with the traditional aspect of coaching. It doesn’t have to be like that.
“It can be another piece of information to help you make a decision. You don’t have to do away with the experience, but more information is always better, and the analytics and the experience are both types of information. It’s just providing things that maybe might be harder to pick up on with the human eye. Analytics provides a different set of data to look into.”
Fuelling, who worked with Michigan’s football analytics, will soon begin work at Telemetry Sports in Indianapolis, which performs analytics consulting with NFL teams.
The data generated by Catapult essentially looks like a spreadsheet and the students label the practice or game intervals as requested by the coaches. The students analyze it and make it understandable.
“Not just hand them a gigantic sheet of a bunch of numbers that you can’t do anything with,” Fuelling said.
The Catapult data, when studied properly, can also track an athlete’s injury that may not be observable by the naked eye. A hockey player, for instance, returns from a groin injury and looks to be skating normally, but the tracker for acceleration can show if the left push is the same as the right. And because they’ll have a baseline for an athlete’s exertion and production in practice, if a trainer wants a rehabbing athlete to work out at 50%, the data can be a guide.
Riela has worked with men’s basketball and will continue with the team next season, attending practice and games, according to COVID-19 protocols, to assist with data-driven projects. He helped create an application where the team can optimize its lineups based on certain statistics, like 3-point shooting percentage.
“There’s 70-plus lineup combinations, so we can set it so it ranks all by 3-point percentage,” Riela said. “If the coaches wanted to find a lineup with the highest 3-point percentage, boom, you can find it in like two seconds.”
Finholt, a self-described “casual” football fan, said they have done studies of turnovers and precursors of turnovers.
“They are rare events and so if you want to try to understand what the impact of turnover is on the game, you don’t have enough of them, so you have to start to operationalize, ‘What is a near interception?’” Finholt said. “(The students) have gone out to operationalize what these near misses might look like. The quarterbacks that have a lot of near misses turn out to hurt their team relative to quarterbacks that have fewer near misses.
“That’s a thing you don’t see when you’re watching the action that you only see as a result of analysis. It’s another way of viewing the game. It’s a way of gaining another focus, but it doesn’t undermine the coach’s intuition. It doesn’t undermine the assessment on the sideline. We don’t need a lot of analysis to know we’re having bad day. That’s gonna make you sort of risk averse whereas when you’ve got it going, you start to say, ‘What the heck, everything else has worked for us.’”
That, Finholt said, is a way analytics offer another angle for coaches.
“There’s always going to be a place for that intuition and having a feeling for what’s happening, but you also have to recognize that a lot of these things probably are probabilistic,” Finholt said. “If you know that, it will shape your intuition.”
Riela intends to pursue a sports analytics job but for now feels fortunate to have this opportunity at Michigan. The advantages of analytics, he said, are limitless.
“The football team, with a younger staff, seems more receptive, and in five, 10 years from now, I can’t wait to see where this goes and how it evolves,” Riela said. “We’re just at the beginning in terms of Michigan sports really integrating analytics.”
Our special thanks to:detroitnews.com