Follow @james_tanner123 One of the more interesting things to me about analytics in hockey is the shot-quality argument. I've seen anti-analytics people make so much fun of the idea that shot quality doesn't matter that listening to them, you'd think that there wasn't any doubt.
But there is.
Lots of people have written about this, and most of it is data heavy and very statistical, therefore I will eschew data here and just talk theory. I realize that this means I will prove nothing and convince no one, so instead my goal is just to make you think about the situation and leave it to you and your search-engine to find the proof, if you're interested. The up-shot is that multiple people have crunched the data and it is generally accepted -as far as I can tell - in the "analytics community" that over time it doesn't have a great effect on scoring.
I think that for something so counter-intuitive (of course shot quality matters, you idiot!!!) it might be worth it to try and explain or discuss it without a lot of numbers. So, based off of a post I made in the comments section yesterday, that is what I'm going to try to do today.
In talking about shot quality it's important to understand that any single game is just an anecdote in the big picture. In one game, shot quality can matter a lot. If you have more high-danger scoring chances than the other team, you are likely to win.
I believe, however, that in the NHL, it is so difficult to get those high-danger chances, and that the goalies will save enough of them, and conversely let enough non-dangerous chances in, that over time, that it all balances out.
Another reason we shouldn't be concerned with shot-quality is that if you were too look at a lot of data all at once, you could easily determine an average shot difficulty level. With a big enough sample size, and knowing how difficult it is in the NHL to get off un-obstructed high-danger scoring chances, I think you can reasonably assume that very few players would be able to out-perform the average, and even those who can only make up one or two out of 18 players on every team.
This is why shot volume is the most important factor. I just don't think you can reasonably expect that a team could take so many low or high quality chances that they would be able to deviate from the average-shot quality very much over time.
If there are 1230 games on the NHL schedule, and each game features roughly 50 shots , that is 61,500 shots, roughly, taken over the course of a full season.
The NHL is a league at nearly full parity. The best teams are only slightly better than all but the worst teams. Only a few teams are truly awful and no team is radically the best.
Same with players: the difference between Ovechkin and Tanner Glass is big, but probably smaller in reality that most realize. This is because at the professional level of anything, the best in the world are only marginally better than the other people who are also the best.
This is important because if players had a greater range of abilities and the difference between them was larger (like it is in lower levels of the game) then luck would be less of a factor. However, in the NHL it is so difficult to get high-quality shots/scoring chances and the goalies and defenses are so good that it is impossible (over enough time) to get enough of them to win regularly while being outshot every game.
Next is the fact that defense, goaltending, coaching and uncalled penalties all work to further shrink the natural edge more talented players have, which in turn makes luck more of a factor and shot-quality less of one.
If there was no defense or hitting, guys like Glass wouldn't keep up and luck would have no effect on the game. Ovechkin would score 140 goals every year and he could take less shots - just hold the puck for a better chance. But that isn't how the NHL works.
The effect of defense and parity mean that it is difficult to get off very good unobstructed scoring chances. In comparisons to the roughly 61 000 thousand total shots across all games in an NHL season, very few are "dangerous."
If it was easier to get really good scoring chances, then the best players would all score more goals and shot quality would be a very big factor. But in the NHL, its rare to get those chances.
Therefore, it makes more sense, when looking at the big picture, to just use the average danger of a shot to stand in for the value of all shots. Just like how you can calculate the average speed you drove somewhere, you can have an average shot danger.
One interesting example is Alex Ovechkin. He routinely gets a lot more shots than any other forward in the NHL, but he does not have a lot success in the shoot-out, and many worse players have higher career shooting percentages than he does. But Ovechkin's secret isn't that he is a better scorer than anyone else: it's his shot volume. If you look at the data since he's come into the league, most of the players just below him on the list of goal scorers score on a higher percentage of their shots. Ovechkin, however, has 1128 more shots than anyone else since he entered the league. (NOTE: That is insane.) The best scorer in the NHL is the best scorer because he understands that shooting the puck is the only thing that matters. Good shot, bad shot, it all equals out.
Given enough time, in the NHL, shot quality doesn't matter.
------------ Coyotes
The Coyotes are at home tonight hosting the Rangers. The Rangers and our old friend Keith Yandle who will be making his first return home since being traded last year.
Good times.
The Coyotes will start Smith in goal and be without Antoine Vermette, which isn't really the problem. The problem is why is Vermette an NHL team's second line centre at this point in his career and how can a team hope to ever win if that is the case?
If the Coytoes keep it close and try for a moral victory, then we should all go home happy.
Thanks for reading.