NHL teams slowly embracing advance statistics
Far removed from punishing hits on the ice, the real crunching in the NHL these days is being done in front offices around the league with the numbers involved in the complex, lengthy calculations of analytics.
The ”Moneyball” approach popularized in baseball has slowly become as much a part of the NHL as the breakaway. More and more teams are turning to the same kind of analytics that have taken over Major League Baseball when they assess talent, players and performance.
Never heard of Corsi and Fenwick statistics? And you call yourself a fan?
It’s a new era in the NHL and – much like in baseball – there’s a still a divide between the new school thinkers and the hockey lifers stewing at the thought that newfangled stats could ever replace gut feel in building a Stanley Cup championship roster.
Take Philadelphia, for example.
The franchise known for decades as the Broad Street Bullies now has more use for an extra set of pocket protectors than rough-and-tumble goons.
”Analytics is where we’re going,” general manager Ron Hextall said. ”You can’t overvalue it, but in my mind it’s going to become more and more and more valuable, I think in all sports. It’s another tool. Why not use every tool available? You still need eyes on hockey players. You need that. I don’t think that will ever change, but the analytics, I wouldn’t say it’s a huge part, but it’s going to get bigger and bigger.”
The Flyers, Toronto, Buffalo, Columbus, the Los Angeles Kings and others are leading the charge in using a new lens at scoping out the way players are judged. The key thought is, there are other ways to scout a player than the traditional means of goals, saves, plus/minus ratio and puck possession time.
Here are some of the stats that are becoming part of the lexicon:
– Fenwick Percentage: The percentage of unblocked shots (on goal or missed) taken by the player’s team; also known as FF%.
– Corsi: Named for former Buffalo Sabres goaltending coach Jim Corsi, this stat tracks shot attempts for and against taken by a team or player. It’s the sum of a team or player’s goals, shots on net, shots that miss the net and shots that are blocked.
– League-Wide Success Rate: The league-wide shooting percentage from that area of the ice in the time frame selected.
– PDO: The sum of a player’s on-ice shooting percentage and on-ice save percentage.
Sam Ventura, a 26-year-old Ph.D candidate at Carnegie Mellon, co-founded the analytics blog War On Ice. Ventura has become such a fan of NHL math, he has tinkered with creating his own advanced stats, working on a metric called zone transition times.
He used zone information (which zone a hit or shot may come from) and measured how long it took for each team to transition between zones.
”Over the long run, we should see the better teams holding the puck in the offensive zones longer and getting the puck out of their defensive zones faster,” he said. ”That’s sort of what I found. The metric I created correlates pretty highly with the number of points in the standings.”
Ventura’s stat could become the next big thing in the NHL. Or it could take years for some teams to adapt.
”Hockey’s such a free-flowing game that it’s hard to determine automatically where each player is at each point of the game,” he said. ”It’s not surprising that hockey has been slower in adopting analytics.”
The hard-liners agree that fancy math should go the way of Fox’s glowing puck.
”There are guys that leave people on the ice in bad situations and don’t get punished for it in terms of the numbers,” Washington Capitals coach Barry Trotz said. ”What analytics doesn’t show for me is when the game’s on the line, when games or important situations happen, it doesn’t show me who’s going to win them.”
The Sabres were at the forefront of analytics with stats developed by Corsi. He has moved on to St. Louis, leaving the heavy lifting in the front office to Jason Nightingale. The Pittsburgh Penguins hired Jason Karmanos as vice president of hockey operations in June to be their analytics guru. The Blue Jackets turned to Josh Flynn. The Capitals hired Tim Barnes. The Flyers use Ian Anderson. The Maple Leafs hired Kyle Dubas, a twentysomething former player’ agent without any previous NHL experience.
Teams have largely refused to make the analytics experts available to the media for fear they’ll expose classified ideas.
Most advanced stats debunk the idea that the oldest stats are still the most reliable. Ventura said he found in his research that hits and blocked shots – bread and butter for many NHL general managers – tend to be overrated.
”If you hit someone, that means you didn’t have the puck before. Not having the puck is bad,” Ventura said.
And blocked shots?
”Not that it’s bad to block shots, but if you have a lot of blocked shots, it means your team rarely has the puck when you’re on the ice,” he said.
There’s really no stopping the movement. The Stanley Cup champion Kings serve as a blueprint for finding undervalued players and consistently ranking among the league leaders in FenClose (the percentage of unblocked shot attempts a team takes in a game when the score is within one goal or tied).
It’s up to a team’s stats whiz to convey what’s important in clear terms to the guys on the bench.
”If I cross the blue line with possession of the puck, I don’t need to be a math major to know that the percentage of shots that I get are going to be higher,” Trotz said. ”Every coach in the league wants to enter the zone with possession of the puck, they really do. And for us, when it gets thrown in a coach’s face, you go, `Yeah, I get that.”’