"Clutch" Goaltenders

[KaseMeOutside]

giguere. he did very well in playoff OT

 

[Big Phil]

To me it's hard to look solely at this kind of data because the on-ice performance tells so much with the naked eye. Take the 1987 Canada Cup with Grant Fuhr. He allowed 6, 5 and 5 goals in that three game series against the Russians. He allowed some soft goals in that series actually. However, typical of Fuhr, when he would allow a soft goal it was always at the beginning of the game. This way he let his team back in the games and then when it was close again he always managed to shut the door. Game #2 in overtime is a perfect example. He was stellar and gets less credit than he should. When the game was on the line Fuhr slammed the door shut, but he allowed 5 goals in an extremely important game. That being said, these were wide open end to end games that were going to have a lot of goals either way.

The opposite I think can be said of Curtis Joseph. His GAA might be better than Fuhr's but he was the anti-thesis for a guy you 100% trusted when the chips were down. Let in some untimely soft goals in his career and that doesn't always show up on the stat sheet.

 

[Doctor No]

That's fine to promote the "naked eye" narrative, although the naked eyes also lie to us (quite frequently).

Comparing Fuhr's and Joseph's GAA (in different eras) doesn't really tell you anything. If you told me the years that each goaltender played, I'd expect Joseph to have a lower GAA than Fuhr (and a higher save percentage). That's part of what this thread is about.

 

[The Zetterberg Era]

Curious how Chris Osgood actually stacks up with some of his peers here. He is a lot of times brought up as the guy that should fall short of the HHOF though his numbers seem to put him with a lot of elite company. Would have very likely won the Smythe had Detroit found a way to close out back to back titles against Pittsburgh. Just curious what he looks like in terms of this, would it help bolster his case or hurt it?

 

[mrhockey193195]

Great stuff! One small note, I believe you're missing a mutual elimination game for Lundqvist - 5/12/2012 game 7 against Washington (Rangers win 2-1).

 

[West]

Curious how Chris Osgood actually stacks up with some of his peers here. He is a lot of times brought up as the guy that should fall short of the HHOF though his numbers seem to put him with a lot of elite company. Would have very likely won the Smythe had Detroit found a way to close out back to back titles against Pittsburgh. Just curious what he looks like in terms of this, would it help bolster his case or hurt it?

Plus one on this.

 

[Doctor No]

Great stuff! One small note, I believe you're missing a mutual elimination game for Lundqvist - 5/12/2012 game 7 against Washington (Rangers win 2-1).

Ah, right you are - thanks! I had it listed incorrectly for both Lundqvist and Holtby; I've updated the overall table and Lundqvist's.

 

[Doctor No]

Olaf "Godzilla" Kolzig!

Kolzig is an interesting one - for those of you that saw my post a few weeks ago in the History forum, Kolzig has the greatest number of postseason games (since 1983) without playing in a mutual-elimination game:

Situation | GP | W | L | Actual SV% | Expected SV% | S+/30
All Games|45|20|24|0.927|0.902|+0.77
Can Eliminate|5|3|2|0.939|0.909|+0.91
Can be Eliminated|7|1|5|0.919|0.896|+0.70

Kolzig looks awfully good in general over the playoffs, but doesn't show any evidence of raising his game in the more "clutch" situations (granted, he's starting from a pretty lofty place).

 

[Doctor No]

This is great! Thanks for sharing.

I'm surprised that the results are not statistically significant. Are they still non-significant if you bin all the elimination games together and ask whether the goaltender is better in elimination vs. non-elimination? I'm just thinking that the sub-categories really cut down on the sample size so if you just make two categories (elimination and non-elimination) you might find that the results are significant. Maybe you have already tried this? Also, what are you using as your significance test? Z-score with binomial distribution for the variance?

Really good question - for better or for worse, I'm using a z-test assuming a binomial distribution on save percentage. This, of course, assumes that save percentage is a perfect binomial, which presents its own problems, but it's the best that I've thought of so far.

I look at a goaltender's overall playoff performance, and my null hypothesis is that (in a given type of elimination game) the goaltender will perform similarly to his overall level.

I haven't tried putting the two types of elimination games together yet (I think that they'd be different kinds of "clutch" situations), but it's certainly something that would be straightforward to try.

There are some results that could be considered statistically significant. For instance, when a goaltender can eliminate an opponent:

Andy Moog (556 expected saves, 574 actual saves, p value 0.010)
Jose Theodore (219 expected saves, 228 actual saves, p value 0.018)
Ed Belfour (661 expected saves, 675 actual saves, p value 0.030)

When a goaltender can be eliminated:
Henrik Lundqvist (589 expected saves, 609 actual saves, p value 0.0011)
Tim Cheveldae (188 expected saves, 197 actual saves, p value 0.023)
Frank Pietrangelo (143 expected saves, 150 actual saves, p value 0.028)

In mutual elimination games:
Henrik Lundqvist (159 expected saves, 166 actual saves, p value 0.024)
Dwayne Roloson (92 expected saves, 97 actual saves, p value 0.030)

Neat to see Chevy on that middle list.

 

[Doctor No]

How about Ed Belfour?

Always loved Ed Belfour, even though he may have cost my Avs a Cup or two in their heyday. I gave away a bit above, but Belfour does appear to step it up when he's got a team on the ropes:

Situation | GP | W | L | Actual SV% | Expected SV% | S+/30
All Games|161|88|68|0.920|0.900|+0.60
Can Eliminate|24|18|5|0.939|0.901|+1.14
Can be Eliminated|21|10|10|0.928|0.901|+0.82
Mutual Elimination|6|5|1|0.921|0.902|+0.58

(Unlike the above table, these results form do not form a partition. Mutual Elimination games are included in all three "elimination" categories)

In mutual-elimination games, Belfour plays about as well as he always does, although the results are quite good:

Game | Result | Score | Shots | Exp Saves | Actual Saves
4/30/1990 vs. St. Louis|W|8-2|28|24.5|26
5/19/1995 vs. Toronto|W|5-2|24|21.9|22
6/4/1999 vs. Colorado|W|4-1|19|17.1|18
5/27/2000 vs. Colorado|W|3-2|33|29.7|31
4/22/2003 at Philadelphia|L|1-6|36|33.0|30
4/20/2004 vs. Ottawa|W|4-1|37|33.4|36

The one stinker may be bringing him down disproportionally (although it was a game seven).

 

[Doctor No]

Please,Luongo is the most clutch goalie.

Lou is still above average in the playoffs in general (but then again, most goaltenders are):

Situation | GP | W | L | Actual SV% | Expected SV% | S+/30
All Games|64|32|31|0.916|0.911|+0.16
Can Eliminate|14|6|8|0.912|0.911|+0.06
Can be Eliminated|7|3|4|0.911|0.907|+0.12
Mutual Elimination|3|2|1|0.931|0.909|+0.63

(Unlike the above table, these results form do not form a partition. Mutual Elimination games are included in all three "elimination" categories)

Luongo's mutual-elimination numbers look great up there, although the sample is small (on the other hand, he's got a few more years to show what he can do):

Game | Result | Score | Shots | Exp Saves | Actual Saves
4/23/2007 vs. Dallas|W|4-1|20|18.2|19
4/26/2011 vs. Chicago|W|2-1 OT|32|29.0|31
6/15/2011 vs. Boston|L|0-4|20|18.2|17

 

[Doctor No]

giguere. he did very well in playoff OT

Giguere certainly did well in playoff overtime, although most of those games weren't elimination games:

Situation | GP | W | L | Actual SV% | Expected SV% | S+/30
All Games|52|33|17|0.925|0.907|+0.54
Can Eliminate|10|7|2|0.918|0.909|+0.28
Can be Eliminated|6|3|3|0.911|0.905|+0.18
Mutual Elimination|1|0|1|0.880|0.917|-1.12

(Unlike the above table, these results form do not form a partition. Mutual Elimination games are included in all three "elimination" categories)

Given what I remember about Giguere, I was surprised that he only has one mutual-elimination game to his credit:

Game | Result | Score | Shots | Exp Saves | Actual Saves
6/9/2003 at New Jersey|L|0-3|25|22.9|22

 

[Doctor No]

Curious how Chris Osgood actually stacks up with some of his peers here. He is a lot of times brought up as the guy that should fall short of the HHOF though his numbers seem to put him with a lot of elite company. Would have very likely won the Smythe had Detroit found a way to close out back to back titles against Pittsburgh. Just curious what he looks like in terms of this, would it help bolster his case or hurt it?

It's weird to say this as an Avalanche fan, but I'm a big fan of Osgood's. He *was* the goalie when my Thunderbirds reached the Memorial Cup (admittedly as host):

Situation | GP | W | L | Actual SV% | Expected SV% | S+/30
All Games|129|74|49|0.916|0.900|+0.49
Can Eliminate|30|14|14|0.920|0.900|+0.60
Can be Eliminated|15|6|9|0.890|0.897|-0.20
Mutual Elimination|6|2|4|0.901|0.900|+0.04

(Unlike the above table, these results form do not form a partition. Mutual Elimination games are included in all three "elimination" categories)

Ozzie seems to lose a bit of steam when on the ropes (fortunately for him, the Red Wings weren't on the ropes a lot). His game sevens:

Game | Result | Score | Shots | Exp Saves | Actual Saves
4/30/1994 vs. San Jose|L|2-3|17|15.0|14
5/16/1996 vs. St. Louis|W|1-0 2OT|29|26.5|29
4/30/2002 at Toronto|L|2-4|29|26.1|26
4/22/2003 at Vancouver|L|1-4|32|28.9|28
5/14/2009 vs. Anaheim|W|4-3|27|24.4|24
6/12/2009 vs. Pittsburgh|L|1-2|18|16.1|16

 

[charlie1]

Really good question - for better or for worse, I'm using a z-test assuming a binomial distribution on save percentage. This, of course, assumes that save percentage is a perfect binomial, which presents its own problems, but it's the best that I've thought of so far.

I look at a goaltender's overall playoff performance, and my null hypothesis is that (in a given type of elimination game) the goaltender will perform similarly to his overall level.

I haven't tried putting the two types of elimination games together yet (I think that they'd be different kinds of "clutch" situations), but it's certainly something that would be straightforward to try.

There are some results that could be considered statistically significant. For instance, when a goaltender can eliminate an opponent:

Andy Moog (556 expected saves, 574 actual saves, p value 0.010)
Jose Theodore (219 expected saves, 228 actual saves, p value 0.018)
Ed Belfour (661 expected saves, 675 actual saves, p value 0.030)

When a goaltender can be eliminated:
Henrik Lundqvist (589 expected saves, 609 actual saves, p value 0.0011)
Tim Cheveldae (188 expected saves, 197 actual saves, p value 0.023)
Frank Pietrangelo (143 expected saves, 150 actual saves, p value 0.028)

In mutual elimination games:
Henrik Lundqvist (159 expected saves, 166 actual saves, p value 0.024)
Dwayne Roloson (92 expected saves, 97 actual saves, p value 0.030)

Neat to see Chevy on that middle list.

Thanks a lot for the detailed response.

Not sure if you're looking for alternative significance tests but a bootstrap approach is a possible alternative to the binomial assumption.

Amazing to see how much Lundqvist steps it up in clutch situations. If I'm understanding that correctly, he has prevented what should have been 30 goals in those 19 elimination games. Really incredible.

 

[charlie1]

Ozzie seems to lose a bit of steam when on the ropes (fortunately for him, the Red Wings weren't on the ropes a lot).

Interesting, I guess that explains (from your website): "WEAKNESSES: Confidence. Bowman doesn't seem to want to rely on him."

 

[Ventana*]

Could you do Curtis Joseph? Like the stats by the way. Thank you.

 

[BraveCanadian]

The opposite I think can be said of Curtis Joseph. His GAA might be better than Fuhr's but he was the anti-thesis for a guy you 100% trusted when the chips were down. Let in some untimely soft goals in his career and that doesn't always show up on the stat sheet.

I know Fuhr has the reputation of shutting the door when it matters most and having seen much of his career I somewhat buy into it but on the other hand you have to keep in mind that his teams had to be good enough to overcome those soft early goals as well..

Joseph is horrendously underrated on the boards here because he fairly often dragged teams that had no business making it beyond the first round farther than they deserved and then they got torched by even better teams.

Also, his teams couldn't score for him.

But hfboards just writes that off as "he was the best first round goaltender ever" nonsense..


Cool look at some of these stats btw.. interesting to see how they match up with our perceptions of some of these guys.

 

[TheDevilMadeMe]

To me it's hard to look solely at this kind of data because the on-ice performance tells so much with the naked eye. Take the 1987 Canada Cup with Grant Fuhr. He allowed 6, 5 and 5 goals in that three game series against the Russians. He allowed some soft goals in that series actually. However, typical of Fuhr, when he would allow a soft goal it was always at the beginning of the game. This way he let his team back in the games and then when it was close again he always managed to shut the door. Game #2 in overtime is a perfect example. He was stellar and gets less credit than he should. When the game was on the line Fuhr slammed the door shut, but he allowed 5 goals in an extremely important game. That being said, these were wide open end to end games that were going to have a lot of goals either way.

The opposite I think can be said of Curtis Joseph. His GAA might be better than Fuhr's but he was the anti-thesis for a guy you 100% trusted when the chips were down. Let in some untimely soft goals in his career and that doesn't always show up on the stat sheet.

I definitely think the data would be a better measure of "clutchness" if you isolated it further to something like "save percentage in the third period of tied or 1 goal playoffs games plus overtime," but that really kills the sample size.

 

[BraveCanadian]

I definitely think the data would be a better measure of "clutchness" if you isolated it further to something like "save percentage in the third period of tied or 1 goal playoffs games plus overtime," but that really kills the sample size.

TCG did something along those lines a while back:

http://brodeurisafraud.blogspot.ca/2009/01/which-goalie-was-best-in-clutch.html

 

[Doctor No]

That's exactly right - I would find it interesting to review TCG's work independently, but it would require me to have my database set up with additional information (unfortunately, I only store shots and saves in total for each goaltender, not by period, and I only track the final game score).

TCG used to post around here quite a bit; it'd be fun to have him back more now that we have this subforum.

 

[Doctor No]

Thanks a lot for the detailed response.

Not sure if you're looking for alternative significance tests but a bootstrap approach is a possible alternative to the binomial assumption.

Amazing to see how much Lundqvist steps it up in clutch situations. If I'm understanding that correctly, he has prevented what should have been 30 goals in those 19 elimination games. Really incredible.

I like the bootstrap approach (although I haven't used it rigorously since my grad school days) - it's something that I will have to look at further here. Thanks!

Regarding Lundqvist, it's about 27 goals prevented beyond average in the 20 "can be eliminated" games.

Since Lundqvist himself is better than average (in all playoff games), it's about 20 goals prevented beyond "Lundqvist average" in the 20 games.

Both quite impressive.

 

[Doctor No]

Could you do Curtis Joseph? Like the stats by the way. Thank you.

I've always been a Joseph fan - my first goal mask was actually a copy of his Edmonton mask in my colors (maroon and gold). I'm staring at it on my wall as I type this.

Here are CuJo's overall playoff numbers by elimination situation (would "Elimination Situation" be a great name for a progressive rock band, or what?):

Situation | GP | W | L | Actual SV% | Expected SV% | S+/30
All Games|133|63|66|0.917|0.901|+0.44
Can Eliminate|20|11|7|0.905|0.897|+0.25
Can be Eliminated|24|10|12|0.922|0.903|+0.58
Mutual Elimination|9|4|3|0.889|0.899|-0.30

(Unlike the above table, these results form do not form a partition. Mutual Elimination games are included in all three "elimination" categories)

The mutual elimination game totals are close to statistically significant (p-value of 6.7%). They certainly don't paint a good picture. The raw totals:

Game | Result | Score | Shots | Exp Saves | Actual Saves
4/30/1990 at Chicago|-|2-8|14|12.4|10
5/15/1993 at Toronto|L|0-6|36|32.1|30
5/19/1995 vs. Vancouver|L|3-5|21|19.0|17
4/29/1997 at Dallas|W|4-3 OT|41|37.0|38
5/4/1998 at Colorado|W|4-0|31|28.0|31
5/9/2001 at New Jersey|L|1-5|20|17.8|15
4/30/2002 vs. NY Islanders|W|4-2|33|29.7|31
5/14/2002 vs. Ottawa|W|3-0|19|17.2|19
4/22/2008 at San Jose|-|3-5|11|10.0|10

In the first, he was replaced by Riendeau; in the last, he replaced Kiprusoff.

 

[KaseMeOutside]

Giguere certainly did well in playoff overtime, although most of those games weren't elimination games:

Situation | GP | W | L | Actual SV% | Expected SV% | S+/30
All Games|52|33|17|0.925|0.907|+0.54
Can Eliminate|10|7|2|0.918|0.909|+0.28
Can be Eliminated|6|3|3|0.911|0.905|+0.18
Mutual Elimination|1|0|1|0.880|0.917|-1.12

(Unlike the above table, these results form do not form a partition. Mutual Elimination games are included in all three "elimination" categories)

Given what I remember about Giguere, I was surprised that he only has one mutual-elimination game to his credit:

Game | Result | Score | Shots | Exp Saves | Actual Saves
6/9/2003 at New Jersey|L|0-3|25|22.9|22

see i think playoff overtime should be considered a clutch statistic, we've all sat through them and there really isn't a much more intense situation IMO

 

[Doctor No]

see i think playoff overtime should be considered a clutch statistic, we've all sat through them and there really isn't a much more intense situation IMO

Agreed completely. See the posts a few above this one, discussing the work already done by TCG.

Hence, I chose a different definition - performance in elimination games - for my testing. Post #1 describes it all.

 

[BraveCanadian]

Game | Result | Score | Shots | Exp Saves | Actual Saves
5/15/1993 at Toronto|L|0-6|36|32.1|30
5/19/1995 vs. Vancouver|L|3-5|21|19.0|17
4/29/1997 at Dallas|W|4-3 OT|41|37.0|38
5/4/1998 at Colorado|W|4-0|31|28.0|31
5/9/2001 at New Jersey|L|1-5|20|17.8|15
4/30/2002 vs. NY Islanders|W|4-2|33|29.7|31
5/14/2002 vs. Ottawa|W|3-0|19|17.2|19

You're right that it doesn't look that good on the face of it but I'm just going to add in:

1993: Cujo was spectacular that series. Blues were an 85 point team. The Leafs were a 99 point team and outshot the Blues all series. He faced 121 shots in the first two OT games alone. Cujo is the only reason it went to 7 in the first place. Leafs finally blow it apart in game 7 to perpetuate the 1st round goalie rap.

1995: Cujo didn't play well all series imo.

1997: Cujo played out of his mind that series. 81 point EDM team vs 104 Stars. Reward? 107 point defending Cup champion Avs team in round 2 to perpetuate his 1st round goalie rap.

1998: Cujo played out of his mind that series. 80 point EDM team vs. 95 point Avs team and Patrick Roy. Cujo allows one goal in the final THREE games to win the series. Reward? 109 point Stars team in round 2 to perpetuate his 1st round goalie rap.

1999: Just throwing this in here that I think he was pretty mediocre against Buffalo.. then again this was Hasek at his best too.

2001: The Leafs are outshot all series. 206-148 for the series. Cujo is a main reason it goes to 7 but his teammates only manage 16 shots on Brodeur in Game 7. Devils rightly win the series (and go on to the Cup finals).

2002: Won't bother with the wins but in the series he lost in 2002 against Carolina the Leafs scored a total of 6 goals in 6 games. The Leafs actually had 9 more points in the regular season but didn't do anything.

It just seems to me that more often than not Cujo was a very strong goaltender on pretty poor teams and the times he was on an at least good team.. they collapsed. Cujo personally only has a few sub-par playoffs and he set his own bar pretty high imo.