Why don’t passing statistics exist?

(This post is part of a fundraiser for 826 Boston, a non-profit tutoring and writing center. For a donation of $50 or more, readers can get me to write anything about hockey they want, so donate today to make me say stuff by clicking here. I probably don’t believe the nonsense below, which was requested by Joey.)

One of the big criticisms you often hear about “advanced” stats is that you can’t measure hockey like that because of “the flow of the game” and the speed of it and so on. One thing that those people don’t like to keep in mind about this is that we can certainly develop those statistics, and pretty easily. We just haven’t yet.

Passing stats should, in theory, be pretty easy to come up with. In much the same way that you record who’s on the ice for every hit, turnover, shot attempt, penalty, and so on, you also just track that data for passes. Along the way you also add in data like who was the passer and who was the receiver, approximate length of the pass, whether the pass was successful, and so on.

It probably takes a few extra people in the press box to do this, but it would start to tell you a lot pretty quickly: The league-wide or team or individual completion rates by pass distance, which players connect for the most passes, which passes lead directly to shot attempts, etc. This would add probably a few million data points per season to our understanding of the sport.

And look, people have tried it for individual teams. Does it necessarily tell us something to know how good any handful of Islanders are at passing? Probably not, but if things were collected at the league level, that’s a whole different story. Right now it’s hobbyists, just like the people who tracked zone entries for so long before a few got scooped up by NHL teams. But it’s all about to be moot.

That’s because all current player-tracking data we have is from observation, and when the league implements RFID chips that track players and puck alike instead, we’re not going to have to sit here and compile this data ourselves by watching the same games over and over again (assuming, that is, the league makes its tracking data public). Instead we’ll be able to more or less instantly see all the things being discussed above, for every second of every shift in every game all season long. This really is an insane new frontier and if nothing else it’s going to give us more data to evaluate the game. Maybe we’ll find out, for instance, that Tyler Bozak might not be good at driving possession or putting up points, but he’s comparably elite when distributing to Phil Kessel on passes of between 15-20 feet when transitioning through the neutral zone at 22 miles per hour. (We won’t find that out, of course, because he’s not, but in theory we could.)

That is the level of data we are about to receive (maybe). Hockey’s a fluid game? You bet it is. And maybe you can’t measure every aspect of it. But what if the NHL’s player tracking gives us something like this? Won’t the sport’s analytics movement have been worth it?

Thank you for reading and supporting 826 Boston. Don’t forget to donate!

Leave a Reply

Your email address will not be published. Required fields are marked *