Analytics Mailbag: What will be impact of NHL’s player tracking data?


Tampa Bay Lightning defenseman Victor Hedman (77) celebrates his goal against the Pittsburgh Penguins during the third period of an NHL hockey game. (AP Photo/Chris O'Meara)

It’s analytics mailbag time again, where we answer your questions with the best data hockey has to offer.

As always, if you asked a question that requires a very in-depth answer, it will likely become its own article sooner rather than later!

Daniel! Why you gotta do me like this? It’s the analytics mailbag and you’re asking what we’re not capturing? Mean, but I’ll do my best to answer this one.

One thing I think hockey analytics struggles with as a whole is developing a robust measure of the situational nature of an individual’s play. This isn’t throwing shade at anyone who has attempted to do so, every year we get better and more robust metrics from people like the Evolving Wild twins and Micah Blake McCurdy, hockey is just an extremely difficult sport for separating player performance from teammate performance.

Some areas that are tough to account for are things like how consistently a player is on the ice with the same teammates versus how often the lines are blended. How much does getting used to the tendencies of your teammates impact performance? Does that have cascading impacts on things like pass completion rates? We’re going to get there, there are too many smart people involved.

While it may not look like an analytics question, I’m going to answer it anyway. I always try to use team colours in my graphics and to get it right I use a reference website called Team Color Codes. This shouldn’t be surprising to many people, but the NHL isn’t a very imaginative league when it comes to colours, and a lot of teams use the same shades.

Edmonton, Florida, Nashville, Washington, and Winnipeg all use the same navy blue (RGB code 4, 30, 66), while Buffalo and Columbus share a similar but different navy (RGB code 0, 38, 84).

Because I work with NHL colours so often, I get super bored of the teams that all copy each other and really appreciate the unique ones that brighten things up a little bit. Because of that, I really like New York Islanders blue (RGB code 0, 83, 155). It’s brighter, and a bit more pale than anything any other team does, and I welcome the variety.

You thought you were asking a troll question and ended up getting a real answer!

I think a lot of people are focused on things like player skating speed or positioning, but a lot of that data already exists in other forms. I know Sportlogiq has data on defensive gap distance, and how much space players can make for themselves on shots, adding in how much pressure players are under while taking shots to give a better idea of which players find that dead space offensively.

What I think we’re going to see come out of player tracking has more to do with body positioning relative to the puck as opposed to maintaining space on the ice. Which players rotate their bodies the right way to get set for shots? How often does a defenceman have their back to the play?

Localized body movement tracking is going to give us more insights that are more actionable from a skills development and coaching standpoint, but I’m not sure how well that sort of data will stack up as predictive of future performance. It could be used by NHL teams to more quickly identify problematic behaviours during a cold streak for a player, in order to work with them to get back on their A-game.

I’m sure there are tons of insights to be had that I haven’t even imagined yet, it’s an exciting time.

Oh boy, Sean is trying to get me in trouble. Not to call anyone out of course, but as I mentioned earlier player speed is often highly focused on but not that great of an insight.

One thing I could see being misused a lot is gap distance. Often when talking about defencemen maintaining their gap while defending off the rush, it’s spoken about in fairly simplistic terms. Keeping a tight gap is generally seen as good, while a wide gap is seen as bad. The truth is this is a very situational stat, both in who a defending player is engaging and the strengths and weaknesses of said defensive player.

For example, the second-widest average gap distance of any regular defenceman in the NHL this season is kept by Victor Hedman. Is he too conservative in his defensive approach? Probably not, he’s a giant with a long stick so he can afford to hang back and take up more space, and he also has the skating ability to close that gap quickly when he chooses.

On the other side of the scale there are players who are too aggressive and get beat because of it. You don’t really want to maintain a super tight gap on Connor McDavid, he’s just going to make you look foolish. So I hope there’s a lot of attention to detail when this kind of data is widely available.

Of course they are! Faceoffs are one of the only situations that coaches have a significant amount of direct control. Choosing who is on the ice, where they’re positioned and what the set play is following relatively few possible outcomes. A team that is good at both winning faceoffs and decisively executing set plays will lead to some success. Similarly a team that is well prepared to respond as a unit when losing a faceoff will have success.

The question isn’t really whether faceoffs are worth anything so much as how much do we overvalue them simply because coaches get to have so much impact on them. A faceoff is just a puck battle with set rules, but there are about 500 loose pucks available for possession changes at even strength every game in the NHL, which is more than 10 times the number of faceoff in the busiest faceoff games. A team that plays a well-structured style without the puck can very easily make up for losing most of their faceoffs.

When submitting content, please abide by our submission guidelines, and avoid posting profanity, personal attacks or harassment. Should you violate our submissions guidelines, we reserve the right to remove your comments and block your account. Sportsnet reserves the right to close a story’s comment section at any time.