Using adjusted save percentage, let’s find the NHL’s best goalie

Last summer, I took a crack at trying to standardize and evaluate goalie performance using something I called SV+, or Adjusted Save Percentage.

Modeled after stats like ERA+ and OPS+ in baseball, it’s expressed in such a way that anything above 100 means a goalie is posting an above-average save percentage and anything below 100 means they aren’t. The number, in short, expresses how many goals a league-average goalie would allow in the number of shots it would take a given goalie to allow 100. I won’t go too much into the specifics in this post, so check that post for further definition/explanation.

Before we look into this year’s version of the charts in the post linked above, a couple of quick notes that I didn’t mention the first time around.

–The thing about this number is that it assumes that changes in all-around goalie numbers over time are largely out of a goalie’s control. To put that one way: in 2011-12 the average save percentage in the NHL was .914, while this year it was down to .912. Is that because goalies simply got worse between last year and this one? Could be, but the fact that average goalie numbers have increased pretty consistently since the full-season lockout leads me to think that they didn’t all just get together before this season and decide to start being worse. I think there’s probably another explanation — maybe offense is up, maybe defenses were allowing higher-quality shots since they weren’t able to work together much during the lockout, maybe there were more power plays called, etc. (I have no idea if any or all of those claims are true, and that’s a project for someone else, but here it’s beside the point.) Essentially, I think league-wide trends outside of a goalie’s control do affect his performance, and SV+ reflects that.

–Last year I just used the save percentage numbers and kind of left them there for us to look at, but didn’t do much more with them. Looking back, that probably wasn’t the right way to go about it. Think about it — Brian Elliott and Cory Schneider, the top two names on the list, played 38 and 33 games in 2011-12 respectively. Is their performance actually more impressive than Mike Smith’s or Henrik Lundqvist’s, who put up fantastic numbers while playing 67 and 62 games? Maybe it’s just me, but I don’t think so.

So this year’s version includes a slight correction there, which we’ll call Goals Saved Above Average (abbreviated as GSAA through this piece). In simplest terms, it’s the difference in goals between what a goalie allowed and what a league-average goalie would allow in the exact same number of shots. It’s calculated based on save percentages, goals allowed, and shots faces, meaning that, for example, if two guys posted the exact same save percentage but one faced twice as many shots, his GSAA would be twice as high (or as low, if it was below-average). Fairly straightforward.

Cool? Cool. Here’s the 2013 season numbers, sorted by GSAA. (I arbitrarily decided to only include goalies that faced at least 100 shots.)

Names that stuck out to me, and other thoughts:

—Should of* kept?

–But seriously, all hindsight/trade arguments aside, it was a hell of a year for Sergei Bobrovsky, especially so when you consider that almost all of the difference there came just from the second half of the season. He’s probably going to win the Vezina and he’ll deserve it.

–Craig Anderson played for half of this season. That he put up the numbers he did in that sample is fairly incredible, though it also makes you wonder how much higher (or lower) that goals-saved figure would be if he hadn’t missed time due to injury.

–Last year, I noted that only five goalies — Lundqvist, Luongo, Vokoun, Lehtonen and Niemi — had played at least three seasons as a primary starter and had never finished any single year since the 2004-05 season with a below-average save percentage. That streak continued this year for four of them, and it would’ve been all five if not for a clunker from Luongo on the season’s final night. Pretty impressive consistency.

–Meanwhile, how about Pekka Rinne, Carey Price and Jonathan Quick ALL being below-average? Not a good year for goalies who recently signed super-long-term contracts.

–Since this is indeed a Flyers blog, I’m obligated to note that Ilya Bryzgalov being third-lowest on the list just makes me really sad.

–Meanwhile: I’ll laugh at Steve Mason until the sun comes up because he isn’t really very good even though 62% of Philadelphia currently thinks he’s The Savior, but his numbers this year between Columbus and Philadelphia were actually above league-average. No, this does not mean he should be handed the starting job for next season. Hell no. But for all of the shots I and many others have taken at him in the last month, it’s worth noting he did, in fact, do well for himself this year. Man, shortened seasons are nuts.

–If there’s one takeaway you have from this post, let it be these next two tables. Here are the top 10 goalies, by GSAA, from the 2011-12 season, with their 2013 totals side-by-side. (Tim Thomas was technically 10th, but since he didn’t play and all, Jimmy Howard got bumped up.)

So the guys who were on top of the world last season, collectively, fell off and were as a whole slightly above-average. Obviously there was some variation within the group – a few guys stayed strong, a few fell down closer to average, and one fell totally off the map and was the worst goalie in the league this year.

Now, here are this year’s top 10 goalies, with their 2011-12 numbers also side-by-side.

Stop me if it sounds like there’s an echo in here: the guys who were on top of the world this season were, collectively, slightly above-average last season. While you’ve got some mainstays from the previous chart (Lundqvist, Howard, Schneider), you’ve also got three guys who were downright bad last year (Bobrovsky, Crawford, Reimer) and others who were pretty average (Niemi, Anderson).

Basically: as a whole, the guys who are really good one year are not a guarantee to be really good again in another year. Everyone together, now: never ever ever ever ever react too much to one good/bad year by a goaltender. Take this advice and apply it to every single goalie you talk about this summer and every summer ever.

–One other thing you may have noticed: the top guys this year had more impressive numbers than the guys from last year, with a collective SV+ that was higher (121.75 compared to 117.88) than last year’s. That may also be a product of the shortened season — chances are those guys don’t all keep up their tremendous numbers at their current rates if this is an 82-game season.

Obviously, a lot of numbers in here, so feel free to ask in the comments for any clarification, and/or to point out anything else that seems odd or incorrect.

* Yes, I’m well aware that “should of” is improper phrasing of the term “should have”. Further reading here.