Even-Strength Goaltender Shot Quality since the Lockout

by Hawerchuk on Oct 10, 2009 3:42 PM EDT in Shot Quality

Shot Quality is an idea that's been around for quite a while - Alan Ryder's original study is available here (PDF). I found a little bit more data than Alan had originally, so I''ve written about it more over the last couple of years: you can look at the 2007-08 study and 2008-09 study or check out this handy sortable chart for both seasons.

I went through and filled in the goalies for 2005-06 and 2006-07, so we can take a look at how this metric works over a number of years for which we have semi-reliable data. This time, I corrected for arena biases like the ones I described at Madison Square Garden - only Henrik Lundqvist is substantially affected, but it's in there.

So which goaltenders performed the best relative to expectations?

Every goaltender who faced 1000 shots since the lockout is listed in this table. Act refers to a goaltender's actual save percentage, and Exp is his expected save percentage based on the model. The delta is the difference between the two - positive numbers are better.

Goalie	Δ	Act	Exp	Goalie	Δ	Act	Exp
J. Hiller	17.0	952	935	J. LaBarbera	0.7	939	938
Y. Danis	14.9	951	936	P. Budaj	0.2	941	941
P. Rinne	12.9	949	936	N. Khabibulin	-0.1	934	934
T. Vokoun	12.0	950	938	M. Turco	-0.5	941	942
D. Hasek	11.9	952	940	A. Auld	-0.9	940	941
E. Ersberg	10.6	951	940	D. Roloson	-1.8	940	942
C. Huet	9.6	948	938	T. Conklin	-2.2	940	942
J. Halak	9.6	945	935	S. Clemmensen	-2.5	939	941
H. Lundqvist	9.0	948	939	K. Weekes	-2.5	937	939
S. Mason	8.6	947	938	F. Norrena	-3.0	939	942
M. Fleury	8.4	948	939	A. Raycroft	-3.1	936	939
C. Anderson	8.4	945	936	B. Johnson	-3.1	937	940
J. Giguere	8.3	947	939	E. Nabokov	-3.4	941	944
W. Dubielewicz	7.9	944	936	M. Biron	-3.5	938	941
T. Thomas	7.7	947	940	C. Osgood	-3.6	934	938
K. Lehtonen	7.3	945	938	A. Niittymaki	-4.0	937	941
C. Price	7.3	946	939	J. Harding	-4.0	936	940
R. Luongo	7.2	949	942	O. Kolzig	-4.3	937	942
D. Ellis	5.9	943	937	J. MacDonald	-4.7	934	939
R. DiPietro	5.9	945	939	J. Theodore	-4.8	933	938
N. Backstrom	5.7	950	944	S. Burke	-4.9	934	939
M. Kiprusoff	5.2	947	941	D. Aebischer	-5.4	934	940
M. Garon	4.8	943	938	D. Sabourin	-6.0	931	937
R. Emery	4.8	944	940	K. Ramo	-6.1	933	939
C. Ward	4.5	941	937	C. Joseph	-6.7	933	940
M. Gerber	4.1	941	937	J. Hedberg	-7.0	930	937
R. Miller	4.0	945	941	P. Leclaire	-7.5	936	943
J. Quick	3.9	943	939	J. Holmqvist	-7.6	932	940
M. Brodeur	3.5	946	942	P. Lalime	-9.2	928	937
M. Smith	3.5	944	940	H. Toivonen	-10.6	926	936
I. Bryzgalov	3.5	945	941	B. Boucher	-10.8	926	937
M. Tellqvist	3.1	943	940	J. Grahame	-13.5	926	939
M. Legace	3.0	943	940	J. Aubin	-14.6	922	937
M. Fernandez	2.1	944	941	M. Denis	-16.6	925	942
C. Sanford	1.6	942	941	J. Markkanen	-17.0	925	942
V. Toskala	1.4	941	940	R. Esche	-17.1	926	943
E. Belfour	1.4	939	938	J. Thibault	-18.8	922	940
C. Mason	0.9	940	939	D. Cloutier	-19.0	921	940

It's amazing to see just how well Jonas Hiller has performed so far in his NHL career - this is not to say that he's performed at his actual level of ability; this is merely what we've observed. It's impressive to see a late-career Dominik Hasek at the top of the list, and Tomas Vokoun again shows up as the most unknown star goalie in the NHL (though his $5.3 million salary says someone appreciates him.)

What's most significant about using this many years worth of data is the relative insignificance of shot quality in overall performance. The variance in expected save percentage is so low that 90% of a goalie's observed performance is due to other factors - primarily skill. If you've seen a goalie play for three or four seasons, you can be pretty confident that his actual save percentage is a good measure of his true performance. This is not to say that some goalies, like Nicklas Backstrom or Evgeni Nabokov, haven't benefited from a coach whose defensive system limits quality chances (or Nikolai Khabibulin, who faced more tough shots than expected), but for the vast majority of starting goalies, shot quality has not been a factor in their career performance at even-strength.

Tweet Comment 8 comments | 0 recs |

Do you like this story?

Comments

Well it passes one sniff test — Dan Cloutier is dead last.

Hiller 1st, Giguere 13th — We’ll probably be missing goaltending coach Francois Allaire pretty soon in Anaheim.

http://www.battleofcali.com/

by Earl Sleek on Oct 10, 2009 5:59 PM EDT reply actions

It’s amazing Cloutier go to face 1000 shots…

by Hawerchuk on Oct 10, 2009 6:11 PM EDT up reply actions

It looks about right. Vokoun’s an elite goalie and kept Florida afloat last year, so I’m not surprised to see him rank very high.

Looks like the heckling that Carey Price gets in Montreal may not be entirely warranted, but that the suggestions to play Halak instead might not be entirely unwarranted — although I doubt the difference between them in these statistics is all that significant. Huet and Danis both ex-Habs, are up there, too. (Heck, technically Vokoun went through the Habs system, too.)

I thought Toskala would score a bit worse than this, guess a weak Leaf defense is a factor in his performances.

Ottawa may not find Leclaire to be a savior if this list is to be trusted.

If shot percentage is of no major significance, would that tend to indicate that shots against are a very strong measure of a team’s defensive ability?

by MathMan on Oct 11, 2009 4:47 AM EDT reply actions

Two things at play here:

1) Yes, shot volume is usually the best indicator of a team’s defensive play. Now, some coaches will accept more shots if they can push them outside (Lemaire, Hitchcock are the best examples), but that’s a rare exception.

2) Over four years, shooting percentage is not driven by luck. But over a single season, luck is more significant. And “shot quality” – how close the other team is able to get to the net – is part of that luck. So while shot volume is the best predictor of future performance, “shot quality” has value in looking back at what happened.

Oh, and a third thing: goalie careers are short. For most players, their apparent skills come and go very quickly. So if we need four years of consistent performance, we’re unlikely to get it; we may not even be able to get four years of consistent ability.

by Hawerchuk on Oct 11, 2009 1:29 PM EDT up reply actions

Not only is Vokoun rated high (as he should be) but his back-up last season and new starting netminder in Colorado, Craig Anderson is up there on the list too. He looks like a steal so far for the young season.

"I know everyone has their own opinion, but your opinion is wrong. "

by Mike @ MHH on Oct 13, 2009 5:47 PM EDT up reply actions

It’d be interesting to see Kippers actual and expected SV% each season since the lock-out. Im guessing they both drop, but with the former intersecting the latter at some point.

Matchsticks and Gasoline
Fivehole fanatics

by Kent Wilson on Oct 11, 2009 1:05 PM EDT reply actions

Kent, you would be exactly right:

Here are the deltas:

05-06: +13.8
06-07: +5.5
07-08: +5.4
08-09: -3.7

Kipper has fallen apart, which was obvious no matter how you look at it.

by Hawerchuk on Oct 11, 2009 1:22 PM EDT up reply actions

That chart will also be a feature of any more trolls that hang at MHH saying we should have had Theo last year because he was SUCH an upgrade over Budaj.

"I know everyone has their own opinion, but your opinion is wrong. "

by Mike @ MHH on Oct 13, 2009 5:48 PM EDT reply actions