When he was 10 years old, Craig Buntin’s dream was to compete in the Olympics. By the time he was 26, he had teamed up with Valérie Marcoux to win three-straight Canadian national championships in figure skating and appear at the Torino 2006 Winter Olympics. Buntin barely missed a return trip to the 2010 Games with a new partner and retired soon thereafter.
“At 30 years old,” Buntin recalled thinking then, “I’m looking at this going, what am I going to do with the next 20 years of my life? I need a goal that is as big as going to the Olympics. The study really was a blank slate—what do I do next?”
First, Buntin founded and sold a coffee-tea fusion company called Teabean. He bypassed undergrad and went straight to McGill University’s Desautels business school for an MBA. For his final project, he conducted an independent research study that would—a few iterations later—turn into Sportlogiq, an AI-powered sports analytics company that can glean insights from a single broadcast camera.
Sportlogiq now works with 25 NHL teams, a handful of NBA and NFL teams, a few soccer clubs, and five major broadcast partners. The Swedish Hockey League recently hired Sportlogiq to be its official data partner. None of that, though, seemed a likely offshoot of his initial MBA project.
Buntin, who is the only student McGill ever allowed to skip a bachelor’s program, surveyed global challenges, using studies from the World Economic Forum as a guide. He was impressed by a lot of academic research in neural nets, machine learning, and computer vision and sought a good market application for them. His first idea was to pursue a self-driving car.
Upon completion of his degree, Buntin linked up with an incubator in Montreal called TandemLaunch that assured funding if he could better refine his project. He toured some two-dozen universities looking for a business idea, even going overseas to the U.K. Eventually, he ended back where he started.
“Everything from cold-calls, emails, to walking around the hallways looking at posters,” Buntin said. “In one of the many poster tours I did, there was one poster up in McGill, of all places—it was actually in my backyard—and it was on surveillance video technology that did anomaly detection.”
He came across the work of then-Ph.D. student Mehrsan Javan, who was applying computer vision to security. If someone dropped a bag, jumped a fence, or did anything else out of the ordinary, the computer system would flag the aberration.
Buntin thought there were broader applications. He and Javan began collaborating on projects, using whatever video they could scrounge up. One test clip was an NFL game. Football was a perfect case study: players wore uniforms, had numbers on their backs, and ran on a field already delineated by yardage. That experiment made clear the potential in sports, where a lot of early tracking information was done manually. Thousands of workers would review footage and tag video clips, a process that was slow and rife with inaccuracies.
The mission Buntin, now CEO, relayed to TandemLaunch was this: “We’re going to build a technology that can see and understand and describe the game the same way that a human would.”
The first product drew on Buntin’s past. Sportlogiq created VeriSkate, which TSN used to track figure skating at a Canadian competition in 2015. The idea was to provide insights like step sequences, jump height, distance covered, and ice spacing while hoping to quell—or even ignite—controversies over the sport’s subjective judging component. Sportlogiq raised its seed round of financing on the heels of that product. (Javan is a fellow co-founder and CTO.)
VeriSkate far exceeded the analytics from Buntin’s own athletic career. Back then, at the suggestion of famed coach Richard Gauthier, he had written down his daily training in a notebook, logging the number of attempts of each element and rating his performance. He’d turn those subjective metrics into charts to track his progress.
Sportlogiq’s first team client was an NHL franchise, which shipped hard drives of video cash-on-delivery. (The startup had to cover the $100 shipping cost.) After that footage was ingested and analyzed, Sportlogiq sent the results back. Buntin noted an immediate impact in the trades that team made and its style of play, all for the better.
Three-quarters of the league has now hired Sportlogiq, whose core technology remains the anomaly detection system. Sportlogiq has 10 full-time Ph.D’s in its employ and collaborates with 17 university researchers doing projects across eight labs and five schools. CB Insights tabbed it one of the top 100 private artificial intelligence companies in the world.
“We’ve really, I believe, hit the primary pain point that these teams have, which is the ability to see the game everywhere, very quickly filter through the data, and have the data answer the questions that they have,” Buntin said.
The Swedish Hockey League was seeking a centralized solution rather than have each franchise and the league office all use different products. After a 20-game test last season, SHL came away impressed with the accuracy of the data and partnered with Sportlogiq for the season now underway. While some basic event stats will still be logged manually, Sportlogiq will dispense tracking statistics via several avenues.
“We’re looking to integrate our current system with this system, so we have one database,” said Erik Strandmark, the SHL’s director of media rights, before later adding: “We will know everything where the players are every second of the game, both the broadcast for media and for fans.”
The statistical models and algorithms have transferred beyond hockey to other sports with a flowing style of play. The system’s roots remain in hockey, however, so perhaps Sportlogiq might figure in the NHL’s plans to introduce a tracking system that combines optical monitoring of player movement with a hardware solution for the puck.
Buntin declined to comment on any discussions Sportlogiq may have had with the NHL, saying only that the NHL’s plan is “one of the most forward-thinking, innovative, interesting approaches in any league I’ve seen so far. I really think that the insights these guys generate, based on the ideas they have so far, could really, really change the sport.”