Hockey analytics will soon pass hobbyists by


The debate surrounding analytics and their usefulness in hockey is over. But now the question is: How best to use them in front offices and in media?

In the not too distant future, every NHL team will be translating the rapidly advancing analysis of hockey into actionable and direct instruction for coaches and players. Fans should expect efforts to be implemented at every level to improve play. Those are the lessons being taught by the analytics movement in other sports such as baseball, basketball and football.

For the moment, it appears that the fertile ground of hockey analytics has produced elite results consistently for franchises considered to be early adopters. As the rest of the league plays catch up to the likes of Stan Bowman’s analytically inclined Blackhawks, gaining—or keeping—an edge will come from emerging areas of analysis.

Some of those new areas were covered at the Pittsburgh Analytics Conference held at Carnegie Mellon University last weekend. It was an interesting mix of introductory lessons directed towards a largely neophyte audience, salted with some of the leading research in the field. The event was hosted by CMU professor Andrew Thomas, who has a Physics degree from MIT and PhD in Statistics from Harvard, and Sam Ventura, a PhD student in the CMU Statistics department. They are also the founders of the statistical resource site

Topics covered at the conference that are still in the ongoing stages of exploration included:

• Research into shot-location data informing defensive and offensive impacts of individual shooters/defenders
• Regression Models seeking to explain possession results and control for contextual factors in order to isolate the possession impact of individual skaters
• Shot-suppression analysis as a correlate to longterm success for NHL franchises
• Time-based analysis of Zone Transitions
• Adjusted Goaltender SV% controlling for a variety of relevant factors
• Wins Above Replacement based on valuation of game events and probabilistic outcomes
• Implementation and Interaction between analysts, management, coaches, players and media

Thomas grew up a Leafs fan in Scarborough, Ont., and, like so many of his peers, was hooked by the early ’90’s Doug Gilmour-led teams. His interest in sports analytics developed as he completed his undergrad degree at MIT and moved on to Harvard for his graduate work. At Harvard he studied under the pioneering sports statistician Prof. Carl Morris, who makes an appearance in Michael Lewis’s Moneyball thanks to his work on baseball.

A project that Thomas undertook tracking games for the Harvard Men’s Hockey Team in 2004 eventually culminated in a paper published in 2006 in the Journal of Quantitative Analysis in Sports. The paper was fittingly titled: The Impact of Puck Possession and Location on Ice Hockey Strategy. Hockey researchers are still exploring this exact topic, and Thomas’s work has since been referenced by a laundry list of analytics luminaries. In recent years, he has taken part in hockey analytics panels alongside the likes of Brian MacDonald—the Director of Hockey Analytics for the Florida Panthers and former professor at West Point—as well as Michael Schuckers, who founded Statistical Sports Consulting, is a professor at St. Lawrence University and developed of Total Hockey Rating—THoR—which has been presented at the SLOAN sports analytics conference.

Thomas and Ventura’s background has situated them well this year to expand the reach of statistically sound research in the greater hockey discussion. Leveraging the relatively central location of Pittsburgh and networking via social media, the CMU professors pulled together an interesting day that was broadcast live and touted in the lead up by local media, many of whom attended the event.

One of the issues in the academic statistical community has been spreading the message to the masses. Most fans are tangentially aware of the analytical work that’s filtered its way through the blogosphere and into the larger conversation of hockey circles in recent years. But less media attention has been paid to the academic work that’s been running parallel to, and largely ahead of, the work being performed by hobbyists.

The number of SABERmetricians in baseball with PhDs in math, statistics and physics has increased drastically over the years, particularly as money, attention and interest has poured into the field. The aforementioned Prof. Morris is the Faculty Advisor to the Harvard Sports Analysis Collective—an undergraduate club that explores various analytical questions in sports, mainly baseball, football and basketball. The club’s blog postings have attracted attention in the past from interested parties such as Dallas Mavericks owner Mark Cuban. This sort of analytical work is no longer flying under the radar in the other three major North American pro sports.

It isn’t under the radar in hockey anymore either. Physicists, economists, engineers and lawyers have been publishing work as hobbyists studying hockey for more than a decade, but hockey is likely to see a trend similar to the other pro sports: As the level of analysis required to break down the game progresses beyond counting shots and calculating simple percentages, the amount of effort required to collect, analyze and break down data requires an increasing amount of time, money and knowledge. University provides a convenient forum for discussion and debate of these concepts in a relatively neutral and unbiased setting.

The cross pollination of ideas between the leading minds of the NHL, NHL fans and skilled analysts remains a fundamentally key driver of the development of the sport, though. And bringing the message to the masses of interested fans is a way to further the overall discussion and adoption of new ideas and ways to explore hockey. The Pittsburgh conference followed on the heels of a similar effort organized by the hockey analytics community in Alberta this past summer. Discussion has been raised about similar conferences in locales such as Washington D.C. or Toronto. Don’t be surprised if one pops up in a town near you soon, and if you’re curious, don’t think twice about attending, just go, because the way the sport is going, there’s no turning back.

For links to PowerPoint presentations (including Burtch’s on dCorsi) and the conference’s YouTube channel, click here.

When submitting content, please abide by our submission guidelines, and avoid posting profanity, personal attacks or harassment. Should you violate our submissions guidelines, we reserve the right to remove your comments and block your account. Sportsnet reserves the right to close a story’s comment section at any time.