HFBoards

Go Back   HFBoards > NHL Eastern Conference > Metropolitan Division > Philadelphia Flyers
Mobile Hockey's Future Become a Sponsor Site Rules Support Forum vBookie Page 2
Notices

An Interesting Look into Goalie Statistics

Reply
 
Thread Tools
Old
05-26-2009, 03:38 PM
  #1
FlyLife
Nuthin but a G Thang
 
FlyLife's Avatar
 
Join Date: May 2008
Location: Livin the High Life
Country: United States
Posts: 2,390
vCash: 500
An Interesting Look into Goalie Statistics

I figured I'd share this with you all, considering it is the offseason, and I'm still dying to talk hockey.

I found this post in the Atlanta forum when I was trying to find some info on Lehtonen. It is some pretty interesting stuff, both the blog he posted and behindthenet.ca (which is a great site). Basically it's about the effects of Shot Quality from team to team, and how that effects the goaltender's SV%, GAA, etc.
Quote:
Originally Posted by flerd_trandle View Post
Last night I stumbled across a blog where shot quality was discussed. You can see the discussion here:

http://brodeurisafraud.blogspot.com/

The author, who is doing some playoff goaltending analysis himself linked to a study of 5 on 5 goaltender performance for this past year. You can see that here:

http://www.behindthenet.ca/blog/2009...rformance.html

This analysis ranks Lehtonen sixth in the league, showing that he's a good goaltender on a bad team. As for other highly touted goaltenders:
I suggest any of you who are interested take a read, its not bad, and you might actually learn something. There are a lot of pissing matches about who's better than who on this site, in the nhl and trade rumors sections and wherever else. I think if we get a little more educated on what we are talking about, it can help out the quality of the site.

The blog coincides with this table from behind the net which I set for goalies with >60 games to remove the backups (its 2 years of data). And two goalies who are possibly available this offseason in Lehtonen and Hiller are both in the top 5 in the Delta column (please Homer read this!). Basically it is the difference between the goalie's actual save percentage, and what the expected save percentage would be based on the shot quality (according to their analysis). Biron is 25th in the league in terms of that statistic.

Then they start looking into how much you can look into those statistics, by determining how much Shot Quality differentiates from team to team. The first 2 cover the league as a whole, and the 3rd one looks into teams like Minnesota and Atlanta who are the expected outliers.
http://brodeurisafraud.blogspot.com/...ens.html#links (the comments are interesting)
http://brodeurisafraud.blogspot.com/...t-2.html#links
http://brodeurisafraud.blogspot.com/...ers.html#links


Behind The Net also has lots of stats for Skaters, like +/- stats that actually mean something, quality of competition, quality of teammates. All very useful and interesting stuff.

FlyLife is offline   Reply With Quote
Old
05-26-2009, 03:58 PM
  #2
mm6492
Registered User
 
mm6492's Avatar
 
Join Date: Dec 2007
Posts: 8,519
vCash: 500
Very intersting read. i knew alot of this about Lehtonen, but there are some other interesting things. I though Biron would be around 25, not that low. And I did not know Hiller was that high either

If Lehtonen can stay healthy, he would be a perfect solution to the goalie problem.


As you said, please Homer, read this

mm6492 is offline   Reply With Quote
Old
05-26-2009, 04:53 PM
  #3
CanadianFlyer88
Moderator
Knublin' PPs
 
CanadianFlyer88's Avatar
 
Join Date: Feb 2004
Location: Van City
Posts: 15,790
vCash: 500
I applaud the effort but there are so many uncontrollable variables that take away from the legitimacy of the study. The author even notes that he can't account for what happens between the shooter and the goaltender for each shot.

The most interesting part of the study for me is that all but 8 of the 40 goaltenders had better GAAs than they were 'expected' to.

If Holmgren has to consult this study, I seriously question his position with the Flyers.

CanadianFlyer88 is offline   Reply With Quote
Old
05-26-2009, 05:10 PM
  #4
zarley zelepukin
Registered User
 
zarley zelepukin's Avatar
 
Join Date: Oct 2008
Location: Norristown, PA
Country: United States
Posts: 1,687
vCash: 500
Quote:
Originally Posted by CanadianFlyer88 View Post
I applaud the effort but there are so many uncontrollable variables that take away from the legitimacy of the study. The author even notes that he can't account for what happens between the shooter and the goaltender for each shot.
Those were my thoughts as well. It seems that the numbers don't take into account the velocity of the shot or its location, did it come off of an odd man rush, was it a one-timer, was it screened? There are too many other things to look at besides shot location to give this study much weight.

zarley zelepukin is offline   Reply With Quote
Old
05-26-2009, 05:41 PM
  #5
FlyLife
Nuthin but a G Thang
 
FlyLife's Avatar
 
Join Date: May 2008
Location: Livin the High Life
Country: United States
Posts: 2,390
vCash: 500
Quote:
Originally Posted by CanadianFlyer88 View Post
I applaud the effort but there are so many uncontrollable variables that take away from the legitimacy of the study. The author even notes that he can't account for what happens between the shooter and the goaltender for each shot.

The most interesting part of the study for me is that all but 8 of the 40 goaltenders had better GAAs than they were 'expected' to.

If Holmgren has to consult this study, I seriously question his position with the Flyers.
They actually talk about how they can't account for other factors in the article. But yea, its obviously not a perfect system, because there are things like screens on the goalie, and deflections that aren't taken into account. But when you are looking at that large amount of data, you can find trends like say that you will score 51% of the time with a shot from 1 foot outside the crease...whether it is a deflection, a wrist shot, or a backhand.

I think it can be similar to Sabremetrics in a way, where you can predict results based on large amounts of data (say 3+ years of data), because eventually every players stats will theoretically get closer to their 'norm.' Obviously Sabremetrics are far more advanced at this point, but when they started it was ideas like this...so I don't think you can just throw it out the window.

Also, there are only 8 of the 40 goaltenders with better GAA's because it is restricted to 60 games played. Here it is with all of them http://www.behindthenet.ca/2008/5_on...ort=10&mingp=1

Anyways, I didn't really take much effort for me to do this, just stumbled upon it and found it pretty interesting. And it is something to talk about besides Lupul, Briere and the Salary Cap.

And I wouldn't mind at all if Holmgren looked into different ways of analyzing goalies. They seem to be a crap shoot anyways, why not look into patterns that could mean something? Especially when we have had absolutely ZERO success doing so for ****ing ever.

FlyLife is offline   Reply With Quote
Old
05-26-2009, 06:37 PM
  #6
BobbyClarkeFan16
Registered User
 
BobbyClarkeFan16's Avatar
 
Join Date: Nov 2005
Location: London Ontario
Country: Canada
Posts: 8,213
vCash: 500
There's just so many variables with regards to goaltending that it's hard to get a good read on things. However, I will say this. The one thing that really seems to be a good indicator of how good a goalie is seems to be save percentage. I've seen goalies with good gaa's but poor save percentages. Save Percentage, I think, is probably the safest stat that you can measure in terms of what makes a good goalie. You might have a goalie with a bad GAA, but if he has a good save percentage, that usually goes a long way in terms of whether he becomes a good goalie or not.

BobbyClarkeFan16 is offline   Reply With Quote
Old
05-27-2009, 11:42 AM
  #7
kicksave27
Registered User
 
kicksave27's Avatar
 
Join Date: Nov 2007
Posts: 2,268
vCash: 500
I'm glad at least somebody agrees with me about Brodeur.

The problem with Kari is not his stats its that he can't stay healthy. I just don't think the flyers can significantly upgrade the goalie position over biron without pulling some insane trade for a top tier goalie.

kicksave27 is offline   Reply With Quote
Old
05-27-2009, 12:44 PM
  #8
GoneFullHextall
bring on the draft
 
GoneFullHextall's Avatar
 
Join Date: Feb 2003
Location: Somewhere in NH
Country: United States
Posts: 35,558
vCash: 50
If Biron decides to price himself out of Philly then they may not have a choice but to at least explore the possibilty of brining in Lehtonen or a guy like Harding.
What other options are out there? Khabibuilin? to expensive. Manny Fernandez? terrible and hes more fragile then anone.

GoneFullHextall is offline   Reply With Quote
Old
05-27-2009, 01:07 PM
  #9
ELab2
Registered User
 
ELab2's Avatar
 
Join Date: Apr 2007
Location: Atlantic City
Country: United States
Posts: 5,348
vCash: 500
Quote:
Originally Posted by CanadianFlyer88 View Post
If Holmgren has to consult this study, I seriously question his position with the Flyers.
Not sure about that. This kind of statistical analysis is becoming MUCH MUCH more common in sports. I don't think it's everything but I certainly think it has a place.

ELab2 is offline   Reply With Quote
Old
05-27-2009, 01:18 PM
  #10
Jester
Registered User
 
Jester's Avatar
 
Join Date: Jul 2004
Location: St. Andrews
Country: Scotland
Posts: 34,075
vCash: 500
Quote:
Originally Posted by ELab2 View Post
Not sure about that. This kind of statistical analysis is becoming MUCH MUCH more common in sports. I don't think it's everything but I certainly think it has a place.
Yes, but you have to be very careful when you analyze such data and the problem that develops -- and I think is the case here -- is trying to get information out of data that simply isn't there (as it is currently collected).

This is particularly true when you're looking at the "danger" of a shot based purely on location. A shot taken from the point with no one between the shooter and the goalie is not terribly dangerous. A shot taken from the point with a lot of traffic can be extremely dangerous. That scenario also ignores the fact that often defenseman are putting the puck on net looking for rebounds (thus creating a more dangerous shot from a closer location), etc. So, a goalie who is really good at controlling his rebounds is theoretically going to face less dangerous shots and, therefore, not look like he's doing such a good job based on the portions of that analysis that I saw. That's a big problem.

He is also attempting to bring a whole team element into this, which is also important to do...but evaluating along those lines isn't necessarily presenting a good story either. For example, certain teams might be better at keeping clear sight lines for their goalies than others. Those stats do not address that.

As BCF16 notes, SVPCT remains the most effective statistic for looking at goaltenders. It is not perfect, and there is most definitely a degree of difficulty to different shot locations...but it's going to be difficult in a sport like hockey to so easily present that data as there are so many variables that go into the danger level of the shot...not the least of which is the simple intention of the shooter.

Jester is offline   Reply With Quote
Old
05-27-2009, 02:38 PM
  #11
ELab2
Registered User
 
ELab2's Avatar
 
Join Date: Apr 2007
Location: Atlantic City
Country: United States
Posts: 5,348
vCash: 500
Quote:
Originally Posted by Jester View Post
Yes, but you have to be very careful when you analyze such data and the problem that develops -- and I think is the case here -- is trying to get information out of data that simply isn't there (as it is currently collected).

This is particularly true when you're looking at the "danger" of a shot based purely on location. A shot taken from the point with no one between the shooter and the goalie is not terribly dangerous. A shot taken from the point with a lot of traffic can be extremely dangerous. That scenario also ignores the fact that often defenseman are putting the puck on net looking for rebounds (thus creating a more dangerous shot from a closer location), etc. So, a goalie who is really good at controlling his rebounds is theoretically going to face less dangerous shots and, therefore, not look like he's doing such a good job based on the portions of that analysis that I saw. That's a big problem.

He is also attempting to bring a whole team element into this, which is also important to do...but evaluating along those lines isn't necessarily presenting a good story either. For example, certain teams might be better at keeping clear sight lines for their goalies than others. Those stats do not address that.

As BCF16 notes, SVPCT remains the most effective statistic for looking at goaltenders. It is not perfect, and there is most definitely a degree of difficulty to different shot locations...but it's going to be difficult in a sport like hockey to so easily present that data as there are so many variables that go into the danger level of the shot...not the least of which is the simple intention of the shooter.
Yeah that makes sense. I suppose that the website would be of negligible use, but if the data was collected in a more detailed and painstaking fashion I think it could be useful. This would pretty obviously have to be done by a team or the league.

ELab2 is offline   Reply With Quote
Old
05-27-2009, 02:50 PM
  #12
Jester
Registered User
 
Jester's Avatar
 
Join Date: Jul 2004
Location: St. Andrews
Country: Scotland
Posts: 34,075
vCash: 500
Quote:
Originally Posted by ELab2 View Post
Yeah that makes sense. I suppose that the website would be of negligible use, but if the data was collected in a more detailed and painstaking fashion I think it could be useful. This would pretty obviously have to be done by a team or the league.
Right, but the problem there is that it becomes terribly subjective, which is the problem with the "scoring chance" statistic. Different teams and different eyes will evaluate what a "scoring chance" is differently, which problematizes any analysis that tries to bring it all together.

The thing that is always important to remember about statistics is that they only tell part of the story, usually the result...rarely the context. Therefore, when you're talking about robust statistics (something like SVPCT where it is not narrowly defined) you can present some generally accurate "broad" statements about stuff...it's when you get into this nitty gritty that it gets immensely more complicated.

It's also why baseball is such a great statistical sport and others...not so much. Baseball is a series of individual events, it isn't a sport "in motion," which most other sprots are.

Jester is offline   Reply With Quote
Old
05-27-2009, 03:15 PM
  #13
CanadianFlyer88
Moderator
Knublin' PPs
 
CanadianFlyer88's Avatar
 
Join Date: Feb 2004
Location: Van City
Posts: 15,790
vCash: 500
Quote:
Originally Posted by ELab2 View Post
Not sure about that. This kind of statistical analysis is becoming MUCH MUCH more common in sports. I don't think it's everything but I certainly think it has a place.
Jester pretty much covered it but my problem isn't the statistical analysis; it's the model in its current form.

This study doesn't account for variables that directly impact a goaltender's ability to position himself to make a save, make a save, react to rebounds, etc.

My Holmgren comment was related to this particular study, not the use of statistical analysis in general.

CanadianFlyer88 is offline   Reply With Quote
Old
05-27-2009, 03:29 PM
  #14
Giroux tha Damaja
Registered User
 
Giroux tha Damaja's Avatar
 
Join Date: Apr 2009
Location: Mount Holly, NJ
Country: United States
Posts: 9,234
vCash: 500
Send a message via AIM to Giroux tha Damaja
I just had an epiphany: Statistics are like bikinis. The good ones show you a lot, but they'll never show you everything. And it is always the important stuff that you can't see.

Goalie stats are like one piece bathing suits with a wrap (they don't really show you ****) and hockey is a fat chick (not particularly suited for bikinis/statistics).

Giroux tha Damaja is offline   Reply With Quote
Old
05-27-2009, 03:42 PM
  #15
IrishSniper87
Registered User
 
IrishSniper87's Avatar
 
Join Date: Jun 2007
Location: Media, PA
Country: United States
Posts: 13,402
vCash: 500
Quote:
Originally Posted by I am The Mush View Post
I just had an epiphany: Statistics are like bikinis. The good ones show you a lot, but they'll never show you everything. And it is always the important stuff that you can't see.

Goalie stats are like one piece bathing suits with a wrap (they don't really show you ****) and hockey is a fat chick (not particularly suited for bikinis/statistics).
I love the analogy.

IrishSniper87 is offline   Reply With Quote
Reply

Forum Jump


Bookmarks

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off



All times are GMT -5. The time now is 08:51 PM.

monitoring_string = "e4251c93e2ba248d29da988d93bf5144"
Contact Us - HFBoards - Archive - Privacy Statement - Terms of Use - Advertise - Top - AdChoices

vBulletin Copyright ©2000 - 2015, Jelsoft Enterprises Ltd.
HFBoards.com is a property of CraveOnline Media, LLC, an Evolve Media, LLC company. 2015 All Rights Reserved.