Will analytics take the beauty out of ice hockey?
In the midst of World Cup this past summer, . It was a staggering piece, not only for its depth, but simply because it’s still not all that common for soccer to be treated with that kind of analytical evaluation. There are ongoing attempts, sure, but trying to develop advanced statistics like those used in baseball for soccer usually leads to the conclusion (or, at least, the argument) that it’s very difficult to analyse such a fluid game in the same way as others, that too much is unknowable, too much is decided by the soccer gods.
You used to be able to say the same thing about hockey. In 2014, that changed.
This year marked a turning point for analytics. For a time, the field of hockey statistics that went beyond the box score figures was an undercurrent conversation built around a relatively small community of hockey devotees. It was mostly online, where blogs scraped the publicly available game data from National Hockey League games and interpreted it in new ways.
The blogging element is perhaps partly why hockey analytics were, generally, dismissed for so long. For a bunch of nerds on computers to tell wisened hockey veterans how to pick teams and predict success seemed to rub the establishment the wrong way. Even now, some are still sceptical of putting too much faith in advanced stats. Just last month at a sports statistics conference, Calgary Flames general manager Brian Burke called analytics “over-rated”. (.)
From now on, though, Burke is fighting for a losing side. 2014 marked the point where the hockey analytics movement hit the mainstream. No longer confined to the internet, analytics are everywhere. They’re on your TV, . If you read about hockey anywhere these days, you’re bound to come across them. And, more and more, just as they have in the NFL and NBA, analytics are taking up space in team head offices. What happened?
No single event can be regarded as changing everything. Online chatter certainly moved the yardsticks. The discussion around the most popular advanced stats, Corsi and Fenwick (both measurements of puck possession – , has been growing for a while, and NHL teams are increasingly aware of what’s being discussed by fans, and what they care about. High-profile columnists and analysts like and CBC’s Elliotte Friedman also helped things along, particularly in Canada, which remains home to the highest proportion of devoted hockey fans.
But if we had to point to one event, it would be when teams started hiring analytics bloggers to work in their front offices.
In mid-August, the Toronto Maple Leafs hired Darryl Metcalf, largely regarded as one of the best analytics sites (and which has now, sadly, shut down due to his contract with the Leafs), along with Cam Charron and Rob Pettapiece. The move came on , founder of (and one of the team’s loudest critics online). Not to be left behind or risk being less competitive – or perceived as such, anyway, , , and all made similar hires.
An important aside here: this is not to say that analytics have never been accepted anywhere in the NHL. Both have been referencing their own, internal, statistics for some time. Blackhawks general manager Stan Bowman joined the franchise in 2001 at age 29 with an analytical framework for players the team hadn’t seen before. However, the details of either team’s analytics remains a mystery – again, we can blame competition.
Still, the moves this summer were significant. In a single week, the Oilers and Leafs hirings catapulted analytics from a niche topic to a widely regarded legitimate way to evaluate hockey.
There’s still some way to go, however, in convincing the majority of the hockey world that analytics have their place. As the 2014-15 season has progressed, there has been a lot of debate over why, if teams like the Oilers and Panthers have all these stats wizards working for them, they’re still so lousy. The answer, provided again by analytical champion Mirtle, is a simple one: . We may yet see a team built entirely off analytical appraisals (that is, unless a team like the Kings has been and we just don’t know it). And by Friedman’s estimation, .
Overall, however, the numbers generally bear out. Last season when the Leafs were in the midst of a streak and heading for the playoffs, there was much scepticism, built on the analytics, they could continue their momentum. Sure enough, they faltered and missed the post-season. The Colorado Avalanche, too, seemed to defy the odds for longer last year, but . This year, Burke’s own Calgary Flames are consistently performing better than the statistics would suggest they should, and before the year is out, barring a major spending spree on new talent (which is a likely scenario), a correction is expected.
Watching trends develop over the long term, or even over the course of a game, isn’t much use if those same data can’t help you change things. It can help size up the competition, particularly if you notice that when Player X is on the ice, puck possession overall for that team is up. Or vice versa. It will likely affect trades and scouting. And in the more immediate term, analytics – one assumes – allow teams to notice their own anomalies or weak points, and perhaps propel new line combinations or new players entirely. Even individual players can use analytics to improve, measuring their performance against people like .
Whether all these new numbers really make a huge difference to the game, or just a difference to those watching it, is still unknown. Like in the NFL, their broad adoption and application is too new to properly gauge success in this realm – or even what success really looks like.
In any case, the question of whether advanced statistics can be applied to hockey seems to be settled, or on its way there shortly. The question that remains is whether it will take all the fun out of the game. When you watch hockey, a lot of the enjoyment is gained in not knowing the outcome, and believing that talent, chance, fate, luck or whatever else will determine things – in short, that the hockey gods will decide things. One could argue that seeing a graph charting win probability in real time limits that enjoyment somewhat. Perhaps that’s true. But I like to think of it this way: the hockey gods are still in control; advanced statistics simply makes it more obvious when they’re doing their best work.