It’s been an exciting summer for analytics in hockey, as team after team has moved to hire analytics consultants. The most high-profile move was the Toronto Maple Leafs’ addition of Kyle Dubas as assistant general manager, but other teams have also made moves:
I can confirm the report about Sunny Mehta being hired by the Devils is true.— Kent Wilson (@Kent_Wilson) June 12, 2014
I'm pleased to announce that I'm going to be working part-time for an NHL team this year.— Eric T. (@BSH_EricT) July 14, 2014
Obviously, teams are moving toward analytics because they believe it will help their decision-making process. But there is a theory out there that the competitive advantage afforded by analytics will disappear as more teams embrace them.
The thinking goes that while early adopters benefit because they know something nobody else knows, that benefit disappears once all these teams know roughly the same thing.
It’s a problematic theory.
“Analytics” is an all-encompassing word, one which in hockey generally refers to any effort to dig beyond conventional statistics (goals, points, etc.). It covers efforts to identify clutch performers by focusing specifically on high-pressure points in a game; it includes work done based on scraping data from the NHL’s official play-by-play charts; it incorporates the efforts of people building shot-quality maps.
To that list we can add dozens if not hundreds of other approaches, from a simple look at giveaways and takeaways to the most complex black-box model.
Consider the Pittsburgh Penguins as an example. Under former general manager Ray Shero, the team made no secret of its interest in analytics. In 2013, The Globe and Mail’s James Mirtle reported on a presentation made by the team’s director of player personnel, Dan MacKinnon, where the executive credited the numbers as a driving force in the team’s acquisition of James Neal.
“I don’t think we’ve made an impact decision since then without consulting the analytics,” MacKinnon said. “I’ll put it that way.”
Fast forward a little more than a year to the hiring of new GM Jim Rutherford, and analytics were suddenly an area that Pittsburgh wasn’t doing nearly enough in.
“I’m also going to bring analytics into the organization,” Rutherford said at his introductory press conference. “I don’t think we’re up to speed here on the use of analytics and this is something that I’ve gotten used to over the last few years. The analytics, if used properly, are great to really check everybody’s opinion.”
There is an obvious contradiction between the comments made by MacKinnon and those made by Rutherford, and it reflects the way that “analytics” has become a catch-all term.
Not all work with statistics is of equal value. Not all organizations lean on analytics to the same extent (either now or in the future), and two different teams relying on two different schools of analytic thought can come to dramatically different conclusions.
The other complicating factor is that as the best public analytics voices get snapped up—and the duo of Tyler Dellow and Eric Tulsky have been two of the most innovative thinkers in the public sphere over the last few years—big advancements will be made behind closed doors.
NHL teams have the resources to get advanced data and likely also to push beyond the speed with which the public sphere generates ideas.
As they outpace what is publicly available (that they haven’t already done so is made obvious by the names of the people getting hired), they should be able to create a competitive edge that other teams will have difficulty in matching, even if they suddenly change their minds and embrace the numbers.
Certainly that’s what the teams making these hires are trying to do.
Jonathan Willis covers the NHL for Bleacher Report. Follow him on Twitter for more of his work.