There are plenty of theories out there as to why hockey has been so slow to embrace advanced analytics the way other sports have in recent years.
The Globe and Mail asked four statistically minded people working in the NHL, either as employees or independent analysts, for their opinion as to why it’s taken so long for the trend to emerge in front offices around the league.
BEHIND THE TIMES
One of the more common theories is that NHL teams are very traditional organizations, with teams often run by large groups of former players and in much the same way they’ve been for decades.
“They’ve all grown up without these types of numbers and have to relearn the way they think about statistics,” said one staffer on a Western Conference team that is beginning to dabble in advanced statistics.
“I think what’s standing in the way of things fundamentally changing in the NHL is the backgrounds of guys who are general managers,” statistician Gabriel Desjardins added. “I think something like 20 of them played in the NHL. Just by its very nature, that’s not going to bring a lot of new ideas into the process.”
TOO HARD TO MEASURE
Others point to the complexity of the sport, with so many shift changes, different numbers of players on the ice in different situations and even randomness (or luck) factoring into every game.
“Hockey’s very complex,” acknowledged Kevin Mongeon, an economics professor at the University of New Haven who founded the Sports Analytics Institute with data analyst Mike Boyle. “In order to extract information that is usable on a regular basis and teams can actually make actions on it, it takes a very high, sophisticated level of analysis to be done on it.
“It’s not as simple as baseball, therefore it takes more of an investment of effort and time for the teams. That has held it back.”
DON’T KNOW WHERE TO TURN
Mongeon also believes there’s also a lot of confusion among teams as to where to go for analysis.
Combined with an unwillingness to invest resources into what’s a relatively new field, that helps explain why so many teams have ignored the area entirely.
“Some teams want to use analytics, but they don’t know specifically what that means,” he said. “There’s a problem in the market with what is a good analytics service to buy.
“Teams aren’t for all intents and purposes investing in major statistical groups to build that intellectual capital [the way they have in other sports]... They don’t have that intermediate person inside the organization. And they’re not really willing to pay for it. But they’re becoming more willing to pay for it.”
Also, statistics that are perceived to be the most useful can rapidly change, something that causes confusion as to which analytics are the best for teams to turn to.
“The signal-to-noise also hurts with adoption of the newer numbers in that the perspectives can be overwhelming,” the Western Conference staffer said. “And it creates an uncertain landscape.
“A few years ago we had [statistician]Alan Ryder’s work, then everyone jumped on the Corsi bandwagon and the online community keeps chipping away at that. You don’t want to invest in a system if it won’t be the standard two years from now.”
A NEW FRONTIER
Some of those working for NHL teams, however, argue that the quality of analysis is only now getting to the point to be really useful.
“It’s not that everyone’s been blind to this stuff forever,” said Pittsburgh Penguins director of player development Dan MacKinnon, who used SAI’s metrics extensively last season and was one of the members on the first hockey panel at MIT’s Sports Analytics Conference this year. “It’s more that, it hasn’t been available until fairly recently. The NHL didn’t even track this kind of [advanced data]until 2006 or 2007.”
Other than the Penguins, MacKinnon points to the Boston Bruins, Buffalo Sabres, Calgary Flames and Tampa Bay Lightning as teams that have begun to invest in this type of analysis.
Many of those in front offices looking at analytics are also from a younger generation, one that grew up reading about Bill James, Moneyball and some of the advanced analysis being done in other sports.
“Honestly, probably the greatest obstacle in sport is that it’s hard to strip away all the emotion and make completely rational decisions,” MacKinnon said. “That’s exactly what analytics is trying to do.
“The reason why we don’t mind being a little open about [what we’re doing]is we’re hoping to take this to the next level where we can get this data from the other leagues [like the AHL] It’s going to take a larger community than just the Penguins to do it. If we get more teams asking for it, I think we can get that.”