clock menu more-arrow no yes mobile

Filed under:

Save % over time in the NHL

How dramatic is the recent change?

Dallas Stars vs New Jersey Devils, 2000 NHL Stanley Cup Finals
Ed Belfour playing for the Stars in 2000.
Set Number: X60553 TK1 R5 F19

There is a genre of tweet that shows up every so often that I like to call: fun with endpoints — Hockey Reference style. Someone who wants to make a wow sort of point, clips a section of the Hockey Reference League Averages page that’s carefully designed to make some fairly normal fluctuation in seasonal results seem amazing! Wow! And away they go.

Absolutely crazy!! Look at that massive difference, that’s so unprecedented. Actually I’m noticing how high the PPO (power play opportunities) is so late into the season this year. I wonder if... no, never mind trying to come up with reasons, it’s just crazy!!

Also, isn’t it interesting that the SA (Shots Against, which are naturally the Shots For as well) are identical to last year? We know from more sophisticated analysis that Expected Goals are up league wide, so it seems like we’re seeing real changes in the way teams pla- no, it’s just crazy!!!

Okay, enough of dunking on that. I’m moving on to victimizing the y-axis fundamentalists. Save % is my favourite thing to make charts of, even though I think it’s a basically useless stat. Even in this context, you could get the same idea from goals and shots or even GAA, which our cropper cut off. It’s currently 3, which it hasn’t been in all the years cropped and put in that tweet. But Save % is fun because it brings out the y-axis complaints if you don’t start it at zero. Let’s do it!

Huh.

That’s all the years the NHL measured SOG, and therefore Save %. What a drop, eh?

Okay, that’s obviously a little misleading, but also educational about the scale of the changes here. But what is the right scale on that y-axis?

You can tell from this one that the last time Save % was lower than this year’s number (.904 to date) was about 20 years ago.

Is that the right scale?

Whoa! Look at that volatility.

We looked at charts like this last season, as we examined the early-season surge in PPO and then watched it drop back to be, hilariously, identical to the year prior, so some of this isn’t new to any of us who remember last fall. This scale for the change in Save % shows so many different things. You can fairly clearly see the advent of modern shooting techniques and new stick technology driving it down and then the response in goaltending techniques driving it back up.

What you can learn from the numbers, if you don’t chop them off to do a wow tweet, is that Save % and PPO have an inverse relationship in the years when modern goaltending has taken hold, but that relationship stopped dead a few years ago.

Save % cratered at .901 in the post-lockout year of 2005-2006, and PPO was 5.85 stemming from the obstruction crackdown. Since then it’s fallen to below three, and might be rising again just this year. But right around 2017, or 2018, Save % began its decline while PPO kept dropping. It’s not “crazy”, it’s really easy to understand what happened. If I tell you, you’ll realize you’ve been watching it unfold in front of you ever since Auston Matthews made the Leafs good.

Shots have increased. First, almost entirely due to more power play time, but then, while PPO was in its trough, unchanging for years at the lowest historical levels ever, the shots went up a lot. And barring the strange pandemic year where the games got really dull — it wasn’t your imagination — they’ve kept on going up. This is true for Corsi, as well, so it’s not some function of measuring SOG or of players getting more shots through. The NHL just shoots more and it started in 2017-2018.

A fun project for someone is to figure out why shooting went up? Is it all players or just forwards? Is it all teams, or a rising quality at the bottom? Is it Gary’s parity or something else. Whatever it is, it coincided with a period where more shots ended up in the net.

It’s not “crazy” that Save % fell, by which I assume the author means inexplicable and/or remarkable. What’s interesting is the cause, but also the history of the game beyond carefully cropped bits of it. The push and pull of skill development and counter changes, technology and technique has been going on for over a hundred years.

The mistake is to think the NHL is unchanging, to imagine it’s a game you can know and understand and then never think deeply about again because you have mapped it all out in your mind. It will change! And we can’t expect shots to keep rising, offence to keep improving. There’s only so close to perfect you can get, which is part of the reason why Save % should have been expected to fall.

We don’t know what will happen next. What if the NHL makes a safe net that can withstand goalie pressure higher up the post? What if goalies can push off with their whole body? Will that make Save % go back up? What changes will come with new style defenders? What about the kids today, who aren’t so willing to get debilitating injuries for sporting glory? What about a changing view of what kind of person plays hockey just fine, thank you very much?

Sometimes the unknowable nature of hockey is really disheartening. Leads aren’t safe, standings lie, you never know what the playoffs will bring, and it’s a hard game to win a Cup in. But sometimes the mystery is exciting. The future of the game is unpredictable, and you don’t want to get left behind by thinking it’s the same game it always used to be to the point you find change shocking.