Skip to Main Content
The Athletic - March 29, 2019

Inside how NHL teams are approaching the upcoming player tracking data revolution

By: Craig Custance

Paul Power is an artificial intelligence scientist for the sports intelligence company STATS. Based in Leeds, England, his focus is primarily on soccer but you can hear the excitement in his voice, even over a spotty Skype connection, about the possibilities player tracking brings to hockey.

He broke the potential impact of player tracking data in hockey into three key components:

1. Athlete monitoring: Power suggests that gathering data like players’ skating speed over the course of several games will help build a better understanding of the physical load a player can exert in a game. It will help teams make their distribution of minutes more efficient. “If you can get your best players out there more than you thought, that’s a huge advantage,” Power said. He shared an example from rugby, where a team discovered that a player who started the game after the first 10 minutes had elapsed could play longer because the first 10 minutes are so intense. “They could maximize his performance,” he said.

2. A better understanding of context: “A huge issue with the current statistics in hockey is it’s very reductionist,” he said. He used passing data as an example. There may be data now as to whether a player completed a pass but there isn’t a wealth of context around it. “Was the guy under pressure? Did you have an overload on a certain part of the ice?” Power explained. “Being able to identify these situations, it allows you to measure how a player copes in these situations and test scenarios.” The next step is to then take the tracking data and simulate how defenders will respond in order to draw up the most effective strategy in real time. “In basketball, which is similar to hockey, imagine the last 20 seconds, you’re down by one and you have to come up with your play,” Power said. “You can draw a play that you want and because you can use all the tracking data, we can learn what you’re drawing and simulate what’s going to happen with the defense and what’s the most likely outcome. They’ll draw something on the fly, imagine being able to test that scenario quickly.” It sounds like something way off in the future, but Power said it’s already being done by the most progressive teams in soccer.

3. Roster construction: Once you’re able to use the tracking data to build models, you can better profile players who fit the needs of the roster and the specific tactics of your coach. “You can better profile players if you can look at movement patterns that your coach implements,” Power said. “Particularly if you have the league-wide data, you can pick out the players who best fit into your style. And then the opposite – who doesn’t fit?”

Soccer is certainly ahead of hockey in using this data and Power has observed a bit of a tipping point over the past 12 to 18 months where teams are seeing the impact of the data and actually believe what can be done. The sport has moved beyond the questioning phase.

Hockey, as the NHL builds out its player tracking platform, is very much entering the questioning phase. And therein lies the potential advantage for teams that capitalize.

Read the Full Article Here