Goalies: Adjusted Playoff Save Percentage (1984-2022)

[Hockey Outsider]

As requested, here's the list of adjusted save percentage, adjusted to the regular season (rather than playoff) average. In the interest of my limited spare time, I'm only posting the overall summary table showing career adjusted save percentage. This table is current through 2012.

* updated for 2012
* I'm not updating this post for 2016 as the difference between this method and the previous method is immaterial. I may revisit this every few years.

Career Save Percentage - minimum 1,000 shots

Goalie|Shots|Saves|Sv%
Patrick Roy* | 7218 | 6646 | 92.1%
Olaf Kolzig | 1446 | 1328 | 91.9%
Ken Wregget | 1767 | 1624 | 91.9%
Tim Thomas | 1524 | 1399 | 91.8%
Mike Liut | 1064 | 977 | 91.8%
John Vanbiesbrouck | 2039 | 1872 | 91.8%
Dominik Hasek | 3422 | 3140 | 91.7%
Ed Belfour* | 4641 | 4250 | 91.6%
Jean-Sebastien Giguere | 1546 | 1416 | 91.6%
Patrick Lalime | 1105 | 1011 | 91.5%
Kirk McLean | 2099 | 1916 | 91.3%
Grant Fuhr* | 3966 | 3618 | 91.2%
Curtis Joseph | 4044 | 3689 | 91.2%
Felix Potvin | 2186 | 1993 | 91.1%
Mike Richter | 2182 | 1988 | 91.1%
Miikka Kiprusoff | 1679 | 1529 | 91.1%
Martin Brodeur | 5439 | 4950 | 91%
Cam Ward | 1137 | 1034 | 91%
Tom Barrasso | 3521 | 3199 | 90.9%
Bill Ranford | 1536 | 1396 | 90.9%
Ryan Miller | 1448 | 1315 | 90.8%
Dwayne Roloson | 1478 | 1342 | 90.8%
Don Beaupre | 1538 | 1396 | 90.8%
Chris Osgood | 3246 | 2944 | 90.7%
Nikolai Khabibulin | 2155 | 1954 | 90.7%
Kelly Hrudey | 2531 | 2293 | 90.6%
Ron Hextall | 2632 | 2383 | 90.5%
Greg Millen | 1336 | 1209 | 90.5%
Henrik Lundqvist | 1565 | 1415 | 90.4%
Roberto Luongo | 1796 | 1623 | 90.3%
Marty Turco | 1345 | 1215 | 90.3%
Jon Casey | 1789 | 1615 | 90.3%
Mike Vernon | 3493 | 3154 | 90.3%
Evgeni Nabokov | 2170 | 1957 | 90.2%
Jose Theodore | 1730 | 1559 | 90.1%
Reggie Lemelin | 1147 | 1033 | 90.1%
Andy Moog | 2655 | 2386 | 89.9%
Ilya Bryzgalov | 1123 | 1007 | 89.7%
Arturs Irbe | 1513 | 1357 | 89.7%
Brian Boucher | 1078 | 963 | 89.3%
Sean Burke | 1101 | 983 | 89.3%
Marc-Andre Fleury | 2046 | 1824 | 89.2%
Antti Niemi | 1267 | 1128 | 89%

The biggest gainers (compared to the previous version, where I adjusted based on the PO rather than RS average) are Liut (+0.8%) and Beaupre (+0.7%). The biggest losers are Roloson (-0.4%), Thomas (-0.6%), Niemi (-0.7%) and Burke (-0.9%). In total, 37 of the 43 goalies experienced no more than a 0.3% change in their save percentage, and more than half (25 of 43) experienced a change of no more than 0.1%.

 

[Doctor No]

I like the update - thanks for putting it together. I'm still torn about the merits of which benchmark to use, but it's interesting to see them both.

 

[Hank4Hart]

There's no way to account for this, but depending on who the coach is some goalies get pulled early in the game (maybe giving up 3 goals on less than 10 shots or something), while other coaches would leave them in there no matter what (see Luongo 8 goal game or Roy 9 goal game), obviously for the goalies who didn't get pulled, their save percentage could have been higher

 

[Doctor No]

There's no way to account for this, but depending on who the coach is some goalies get pulled early in the game (maybe giving up 3 goals on less than 10 shots or something), while other coaches would leave them in there no matter what (see Luongo 8 goal game or Roy 9 goal game), obviously for the goalies who didn't get pulled, their save percentage could have been higher

On the other hand, there are goalies who give up 3 goals on 8 shots, are left in the game, and stop the last 22 shots that they face. It works both ways.

 

[MadLuke]

There's no way to account for this, but depending on who the coach is some goalies get pulled early in the game (maybe giving up 3 goals on less than 10 shots or something), while other coaches would leave them in there no matter what (see Luongo 8 goal game or Roy 9 goal game), obviously for the goalies who didn't get pulled, their save percentage could have been higher

We have a lot of shot for sample size and those stats and we will need a bigger pattern than a specific games to change the final results first decimals.

Luongo in florida was often pulled when the game was out of hand for the panthers, to make him start the very maximum of games, Luongo playing a game being the biggest hope of winning and making the playoff for them.

 

[Hockey Outsider]

Updated for 2013

 

[hatterson]

I'd be very interested to see how the results change if you include only even strength play.

Do you have your data hosted somewhere? I'd be willing to take a look at it and throw in some ES data from NHL.com which would at least get us back to 98.

 

[seventieslord]

Good work, HO!

 

[Hockey Outsider]

Updated for 2014.

 

[Hockey Outsider]

Updated for 2015. The single biggest takeaway is how great Henrik Lundqvist`s numbers look in an all-time sense.

 

[Hockey Outsider]

Updated for 2016. Four things stand out:

1. Braden Holtby`s playoff numbers are phenomenal. Yes, he only has four playoff appearances, and has never been out of the second round, but his numbers are consistently great.

2. Henrik Lundqvist had a rough spring, which hurts his numbers. Still, a single trip to the conference finals would probably put him 5th in wins.

3. Patrick Roy was a pretty good goalie in the playoffs. I know this has nothing to do with my 2016 update, but I`m still blown away by his numbers. His ability to maintain an extremely high level of play over an enormous number of games is without parallel among playoff goalies. He has three of the top seven, and four of the top sixteen, individual postseasons.

4. Goaltending is less of a difference-maker these days. Seven of the ten worst performances from a Stanley Cup winning goalie have occurred since 2007. (Most of those were objectively good performances, but "worst" relative to other Cup winning goalies). There are notable exceptions of course (Thomas in 2011, Quick in 2012), but it seems like goalies don't need to steal many games any more. They just need to be quietly dependable. As goaltending continues to improve, the gap between the best netminder and an average one continues to shrink.

 

[Doctor No]

Holtby's playoff career has been underrated, certainly.

In the period where we have save percentages (either actuals or good estimates), Holtby's postseason support-neutral winning percentage is 62.6%, which (among goalies with 20 or more games played) is second highest - behind John Davidson.

Great work on the update - as always!

 

[Hockey Outsider]

I also calculated Holtby having a 62.6% win percentage (see post #6), so good to see we're getting the same results.

 

[MoonDragn]

I still feel that not enough data is collected on how Brodeur's puck playing skills have led to an overall better defense in NJD and how his SV% numbers really doesn't reflect how much impact he has on his team. There is a reason he won so many games and while some of it had to do with his teammates, a lot of it was because of him as well.

I feel that Goalies like Brodeur are hard to fit into your neat SV% ranking because he performs so much outside of the box that it is hard to quantify. I find it funny that people are always using statistics to prove how bad he was, but not realizing that the statistics doesn't show everything that was unique about Brodeur.

 

[Doctor No]

People can present things as research without them necessarily proclaiming them to be infallible or the be-all end-all.

I don't normally feel comfortable speaking for other people, but I know Hockey Outsider well enough to confidently tell you that he's aware of the limitations of save percentage.

In fact, he says as much in the thread starter: "My purpose isn’t to present one number which is a perfect representation of a goalie’s performance. Rather, I want to improve on what has already been quantified in conventional statistics."

 

[MoonDragn]

He does say that, but then he added this bit:

I’ve stated before that save percentage is, in my opinion, the single best statistic to measure goalie performance.

I realize it is all opinion, I just think that you can't adjust stats that easily as there are many many factors that can affect each one's stats. and as pointed out, all it takes is one anomaly to skew the results.

 

[Hockey Outsider]

I appreciate the comments. I said "save percentage is the single best statistic" in order to explain why I did the research in the first place. For example, I could have used another statistic (like wins, shutouts, or GAA) but save percentage is less influenced by factors beyond the goalie's control (such as offensive support, or number of shots faced). I've written about that extensively in other threads, so won't repeat the reasoning here. Of course, every statistic in hockey is influenced by team context - the question is a matter of degree.

Dr. No is absolutely right. My intention is to improve the statistics that already exist. It's better to adjust statistics for known factors, rather than use unadjusted numbers. (Otherwise, Patrick Roy looks like a very good, rather than outstanding, playoff goalie, and Grant Fuhr looks fairly poor by modern standards). It's better to correct for a known problem (even if there are other problems that we can't yet quantify). The only problem with this approach is if someone says that after X adjustments, they have a perfect, definitive statistic (which I haven't done).

I wrote a long post several years ago (maybe 2010?) when I tried to quantify Martin Brodeur's two primary advantages that aren't captured in save percentage - his durability, and his puckhandling skills. I concluded that, based on some plausible assumptions, Brodeur probably matched Roy in terms of regular season performance. There's definitely value in those two qualities that aren't captured in save percentage, so an analysis focusing solely on that would underrate Brodeur.

 

[Hockey Outsider]

Updated for 2017 (see comments in the posts on the first page).

 

[Hockey Outsider]

With the exception of post #26 (I can't find my data source for that), this thread has been updated after the data migration on HFBoards. Let me know if you see any errors/problems.

 

[Hockey Outsider]

Updated for 2018 (see comments in the posts on the first page).

 

[Hockey Outsider]

Updated for 2019 (see comments in the posts on the first page).

 

[Hockey Outsider]

Updated for 2020 (see comments in the posts on the first page).

 

[Leafmealone11]

Here's a chart showing the distribution of adjusted save percentages. I'm only using goalies with 300+ minutes; that threshold is arbitrary, but there needs to be some kind of arbitrary threshold or else we'd be looking at statistically meaningless results for goalies who faced only a few dozen shots.



The largest single category (90% to 91%) straddles the mean and features nearly 25% of all observations. Although there are more observations above the mean than below, the distribution appears to follow roughly a bell curve.

If the average save percentage was 90.5%, but most goalies were actually stopping say 85% but Patrick Roy and a few other superstars were stopping 95%, than another distribution (and therefore another method to evaluate goalies) might be more meaningful. Let me know if I've misunderstood your post.

Noob question, is this average based on goalies that played 300+ minutes only or is it goalies that played 300 minutes vrs league average?

 

[Hockey Outsider]

Noob question, is this average based on goalies that played 300+ minutes only or is it goalies that played 300 minutes vrs league average?

Definitely a good question. When I calculate the league average for each year's playoffs, all goalies are included. In terms of what I present, I usually filter the data, but every shot is included in figuring out the average each year.

 

[Bouboumaster]

Very interesting stuff! Thanks good sir!

Patrick Roy is a monster!