Eric Tulsky knows better than anyone that hockey is a complex sport.
The 10 skaters on the ice – 10 different variables, from a mathematical perspective – are constantly interacting with each other in fluid and unpredictable ways, and the identities of those variables are oft-changing as players shuffle on and off the bench. For most of its existence, those dynamics have limited the scope of statistics for hockey.
But the hockey analytics movement has begun at full blast in the 21st century, and Tulsky, one year into his full-time employment as the Carolina Hurricanes’ data analyst, now finds himself at the movement’s forefront.
“If you have a game that’s already solved, it’s not a fun game to analyze – nobody wants to do tic-tac-toe analysis,” Tulsky said. “Being in a field that’s hard to analyze means that there’s a lot of places ... we can still explore.”
Tulsky, 41, grew up a hockey fan but didn’t enter the industry until well into his career. Having double-majored in chemistry and physics at Harvard and then having attended graduate school at California-Berkeley, he worked in nanotechnology for over a decade while studying and blogging about hockey on the side.
He thinks he entered the analytics field at the right time – while more quantifiable sports like baseball had already experienced analytics revolutions, the groundwork had been laid for hockey, but the industry was still in its infancy.
“It was clear what direction the field was going, but when I got into it ... there was still a lot of low-hanging fruit left,” Tulsky said. “There were a lot of things that I looked around and said, ‘How come nobody’s done that before?’ ”
The Philadelphia native eventually began working part-time for the Hurricanes in 2014-15 and was hired full-time last summer, spurring him to move from California to Raleigh.
“My job is to collect data on anything I can and use that data to make recommendations anywhere where I can provide insight,” Tulsky said.
Given the exclusivity of the still-growing hockey analytics industry, his research for the Hurricanes is secret.
Not secret, however, is Tulsky’s work for his SB Nation blogs, Broad Street Hockey and Outnumbered, which he wrote for from 2011 to 2014.
My job is to collect data on anything I can and use that data to make recommendations anywhere where I can provide insight.
Canes analytics expert Eric Tulsky
One of Tulsky’s most oft-cited findings during that era of publishing his research was that players reach peak offensive production at around age 24, debunking scout ratings from one team – he won’t disclose which, but says it wasn’t the Hurricanes – which most highly valued players at age 29.
The effects of that finding seem to be reflected in the Canes’ current roster makeup, as five of the team’s projected 2016-17 defensemen are 24 or younger and last year’s top scorer, Jeff Skinner, also turned 24 in May.
Another sizable portion of Tulsky’s published work focused on offensive zone entries, which are generally categorized into either “carry-ins” (the player skates the puck into the zone) or “dump-ins” (the player shoots the puck into the zone before he and his teammates pursue it).
Tulsky’s report on the subject, which showed that carry-ins produced more than twice as many shots on goal than did dump-ins, was eventually featured at MIT’s 2013 Sloan Sports Analytics Conference.
“A lot of the difference in how many offensive opportunities a team got came down to that neutral zone decision of whether to carry it in or dump it in,” said Tulsky, who can instantly elaborate on the topic. “When you’re focusing on offense, I think people naturally focus on the offensive zone play, and what that published work suggested was that a lot of the generation of offense occurs through the neutral zone rather than in the offensive zone.”
Tulsky also once wrote about the stark unpredictability of goaltenders and the difficulties that poses for teams’ contract decisions.
He declined to comment on anything Hurricanes-specific, but his findings on the subject are relevant to the Hurricanes’ recent decision to re-sign 32-year-old goalie Cam Ward, whose career .910 save percentage ranks 36th among 50 goaltenders with 200 or more appearances since 2005.
“If you watch a guy play 10 or 20 games and he faces a few hundred shots, a couple bounces here and there can be a big difference in his numbers,” he said. “The result is that it can just take a really long time to evaluate a goalie. Problem is, by the time somebody’s played a really large sample, they’re often past their prime, so it’s just a really tough market to analyze.”
Hockey analysts have been in the news especially frequently this offseason. Not every team employs someone to fill a role like Tulsky’s, and two that did – the Montreal Canadiens and Edmonton Oilers – both parted ways with their analysts recently.
The man fired by Montreal, Matt Pfeffer, then went public to The Hockey News about his opposition to the P.K. Subban-Shea Weber trade and spoke of a “pushback from people in the NHL” against analytics. He later apologized for his comments.
Hurricanes coach Bill Peters has made it clear that such a pushback was not the case in Carolina.
“(Tulsky’s) extension is 3614, I call him, I ask him to come down into the war room, and then when it’s time to get back to work ... I kick him out,” joked Peters about the closeness of their relationship on Sportsnet radio over the summer. “I give him projects, but he kind of scoffs at them because they’re so simple-minded. He likes the big, long, elaborate ones, but I like the information, and he gives us real good information.”
Tulsky, too, spoke highly of his role with the Hurricanes and said that, given his newfound access to the team’s data and tracking capacities and the time afforded by working full-time, he has progressed significantly beyond his published data of years past.
But as he prepares for his second season inside PNC Arena, he knows there will always be more he can do.
“The way you get a competitive advantage is in part by advancing the ball beyond what’s in the public domain,” Tulsky said. “Anything that I think might be useful, I’m continually working to improve.”