Here we go again.
— Hawerchuk (@behindthenet) October 28, 2013
Much of the problem with the debate over shot quality, it seems to me, is the words. Just saying shot quality is too general and that is because people mean two different things when they say it. First they mean the probability of an individual shot and, second, they mean the average shot probability. Somehow quality has snuck in where probability belongs.
Individual Shot Probability does matter and this has been illustrated time and time again. There’s no doubt about it. The most recent example is the distance analysis by Michael Parkatti, Different shots have different probabilities of going in and there are plenty of factors that influence these probabilities. These include x and y coordinates as well as the type of shot matter. Here are some heat maps to emphasize that.
What has not been shown to matter much, to my mind, is Average Shot Probability (ASP), either for shots that a goalie has faced or that a team has faced or that a team has generated over a long period of time. It might be there but the consensus (yes, David, I see your hand is up) is that it is not. I’ve tried to look for it. It, ASP, matters but not a ton. I’ve got plans to take another look at it again this winter. But there is little denying that where we are right now is that we lack evidence for the value of long term repeatable ASP. Somewhere there’s a fourteen year old kid with mad R skills and a great idea on how to model these data and, perhaps, they’ll find that shot quality exists. It’s just that right now we don’t have enough evidence for it.
In the next couple years, we’re going to get a deluge of more and better data. That data will include things like speed of the shot, better x and y coordinates and additional context like whether or not the goalie’s line of sight is obstructed. It may change much of what we know and value right now. I’m excited to dig into it.