As the NHL season gets underway, Jake Hurlbut ’15 will be paying much closer attention to the statistical side of the game thanks to his St. Lawrence University fellowship experience with Associate Professor of Statistics Michael Schuckers.
A few years ago, Schuckers created Total Hockey Ranking (THoR), a comprehensive statistical rating of NHL forwards and defensemen based upon all on-ice events. The work he’s done with THoR, which several St. Lawrence students have had the opportunity to help with, has earned national recognition at the MIT Sloan Sports Analytics Conference, known as the “Super Bowl of Sports Analytics,” drawing attention from professional teams and agents.
In spring 2014, Schuckers was looking for another student to work on THoR with him over the summer. It just so happened that Jake’s interests aligned perfectly with what Schuckers needed.
“When Dr. Schuckers approached me, it was a dream come true for me because I was able to combine two of my favorite things – computer science and hockey,” Jake says.
Most of Jake’s time as a University fellow was spent working on the code for THoR in order to make it faster and more efficient. Before Jake started working on the project, it took over a day to compute a season’s worth of statistics. Now, it takes about four minutes.
“We can now test and design new models in a way that allows us to determine the results in hours rather than in weeks,” Schuckers says. “This will greatly speed up the research process. It will also allow for us to investigate a range of models that will be similar to THoR.”
Jake’s summer work with Schuckers is coming at an interesting time in professional hockey. According to Schuckers, there’s been increased chatter among bloggers, journalists and a few academics who have been looking to apply quantitative methods to hockey over the last several years despite the belief that hockey is too hard to analyze. Leading into this NHL season, several teams have said they are interested in looking at these methods, while some are even hiring analysts. Between Jake’s summer work and hockey knowledge, he sees why something like THoR could be so valuable to the game.
“Hockey is a fast-paced game, so to be able to use these numbers to make decisions with the freshest data is really important,” Jake says. “THoR takes into account every event that happens on the ice, then looks at them over a long period of time. Something like this might not be usable for an in-game scenario, but for long-term strategizing, having these numbers could be very useful a quarter- or half-way through the season for making trades and moving personnel around the team.”
As the NHL finds itself at the brink of this new analytical era, Jake is taking his summer experience and also looking ahead.
“This was the first time I was able to devote all my time to the programming involved in one project,” Jake says. “Being at St. Lawrence last summer and focusing on this research project allowed me to be completely immersed in it. I understood what was going on in the program on a much deeper level. That’s exactly the kind of experience I need to get ready for what I’ll be doing after I graduate and go out into the real world.”