HFBoards

Go Back   HFBoards > General Hockey Discussion > The History of Hockey
Mobile Hockey's Future Become a Sponsor Site Rules Support Forum vBookie Page 2
Notices

The History of Hockey Relive great moments in hockey history and discuss how the game has changed over time.

Adjusted goalie stats. Has it been done?

Reply
 
Thread Tools
Old
02-22-2012, 01:51 PM
  #1
Weztex
Registered User
 
Weztex's Avatar
 
Join Date: Feb 2006
Posts: 1,292
vCash: 500
Adjusted goalie stats. Has it been done?

Curiosity here.

There are countless methods and projects that try to regulate different eras of scoring. What i've yet to see is any adjustement to goalies numbers. Prorating seasons to 82 games certainly appears silly since we would end up with tons of goltenders boasting 82GP and 50-60 wins seasons. Has anyone ever bring up some sort of formula normalizing games played/wins from one era to another?

Thanks.

Weztex is offline   Reply With Quote
Old
02-22-2012, 01:55 PM
  #2
Bear of Bad News
Mod Supervisor
 
Bear of Bad News's Avatar
 
Join Date: Sep 2005
Location: Windsor
Posts: 4,421
vCash: 663
It's been done by some people - although the "big deal" goalie stats are rate stats anyhow.

The most important place to do it is in the "value" metrics - wins above replacement, goal differential, things like that. Iain Fyffe blogged about a reasonable way to do this about a month ago:

http://hockeyhistorysis.blogspot.com...altenders.html

Bear of Bad News is offline   Reply With Quote
Old
02-22-2012, 03:03 PM
  #3
seventieslord
Moderator
 
seventieslord's Avatar
 
Join Date: Mar 2006
Location: Regina, SK
Country: Canada
Posts: 28,374
vCash: 500
Quote:
Originally Posted by Weztex View Post
Curiosity here.

There are countless methods and projects that try to regulate different eras of scoring. What i've yet to see is any adjustement to goalies numbers. Prorating seasons to 82 games certainly appears silly since we would end up with tons of goltenders boasting 82GP and 50-60 wins seasons. Has anyone ever bring up some sort of formula normalizing games played/wins from one era to another?

Thanks.
I have seen this done. It helped to smooth out eras where platooning was more prevalent, for example.

But yeah, like The Taco Man said, itís all about rate stats. See hockey outsiderís adjusted save percentage. Itís great but it would be nice if someone with time went ahead and used the stats we have going back to expansion at least (not sure pre-expansion numbers are comparable in the same way Ė what does it mean to be 4th out of the best 6 goalies in the world in 1967? Technically itís below average but thatís not really fair when the next year the guy in 6th is ďabove averageĒ)

seventieslord is offline   Reply With Quote
Old
02-23-2012, 08:02 AM
  #4
tarheelhockey
Global Moderator
 
tarheelhockey's Avatar
 
Join Date: Feb 2010
Location: The Triangle
Country: United States
Posts: 45,239
vCash: 1020
Quote:
Originally Posted by seventieslord View Post
But yeah, like The Taco Man said, itís all about rate stats. See hockey outsiderís adjusted save percentage. Itís great but it would be nice if someone with time went ahead and used the stats we have going back to expansion at least
Going back before 1982-83 would require having average leaguewide save% for those seasons... are those numbers available somewhere?

tarheelhockey is online now   Reply With Quote
Old
02-23-2012, 09:17 AM
  #5
plusandminus
Registered User
 
Join Date: Mar 2011
Posts: 980
vCash: 500
Quote:
Originally Posted by tarheelhockey View Post
Going back before 1982-83 would require having average leaguewide save% for those seasons... are those numbers available somewhere?
I think I may have seen them at http://sports.groups.yahoo.com/group/HAG_list/
or if it was http://hsp.flyershistory.com/
At least, there are total shots at, and goals allowed, for each team. One can use those two stats to get save percentage, but should remember to exclude empty net goals and empty net shots.

By the way. I posted data some weeks ago where I wanted to discuss how reliable save percentage is. To what extent do skaters affect save percentage? I wanted to adjust goals allowed when playing short handed, based on save percentage and, if useful, other stats. (I know that is probably not what this thread is about.)
People spend a lot of time focusing on +/-, and even adjusting it, but it seems that goaltending is seldom included in the analysis despite being a very big factor in "biasing" skater +/-. It is easier to get good +/- having a goalie like prime Hasek, or Roy, than it is to have below average goaltending. It is easier for the current players of NY Rangers and Boston to have great +/-, than for those on teams with worse goaltending.

plusandminus is offline   Reply With Quote
Old
02-23-2012, 10:19 AM
  #6
tarheelhockey
Global Moderator
 
tarheelhockey's Avatar
 
Join Date: Feb 2010
Location: The Triangle
Country: United States
Posts: 45,239
vCash: 1020
Quote:
Originally Posted by plusandminus View Post
I think I may have seen them at http://sports.groups.yahoo.com/group/HAG_list/
or if it was http://hsp.flyershistory.com/
At least, there are total shots at, and goals allowed, for each team. One can use those two stats to get save percentage, but should remember to exclude empty net goals and empty net shots.
If anyone is planning to calculate this data, I have crunched some of the season numbers since 1998 and it would appear that subtracting empty-net goals and shots has an effect of roughly ~0.0025 on the overall league save percentage.

To state that more clearly:
If you're calculating Overall Goals - Overall Shots / Overall Goals,
you can just subtract about 0.0025 from the result to adjust for empty nets.

This is of course the modern rate. It's pretty much impossible to know whether it translates to earlier eras until someone enters all the old box scores into a database.

tarheelhockey is online now   Reply With Quote
Old
02-23-2012, 10:31 AM
  #7
reckoning
Registered User
 
reckoning's Avatar
 
Join Date: Jan 2005
Country: Canada
Posts: 5,829
vCash: 500
Quote:
Originally Posted by tarheelhockey View Post
Going back before 1982-83 would require having average leaguewide save% for those seasons... are those numbers available somewhere?
http://hfboards.hockeysfuture.com/sh...d.php?t=698806

reckoning is online now   Reply With Quote
Old
02-23-2012, 10:35 AM
  #8
seventieslord
Moderator
 
seventieslord's Avatar
 
Join Date: Mar 2006
Location: Regina, SK
Country: Canada
Posts: 28,374
vCash: 500
Quote:
Originally Posted by tarheelhockey View Post
Going back before 1982-83 would require having average leaguewide save% for those seasons... are those numbers available somewhere?
Absolutely. I have it all. I have all the shots and saves for every goalie in each season, and the league average is just the sum of all saves over the sum of all shots.

seventieslord is offline   Reply With Quote
Old
02-23-2012, 11:03 AM
  #9
tarheelhockey
Global Moderator
 
tarheelhockey's Avatar
 
Join Date: Feb 2010
Location: The Triangle
Country: United States
Posts: 45,239
vCash: 1020
Well I'll be damned. Thanks SL.

That makes the league-average save% pretty easy then. Here you go.

SeasonTotal GoalsTotal SavesSave %
1967-68243826951.910
1968-69268929331.908
1969-70261929780.912
1970-71336434769.903
1971-72329233760.902
1972-73403348737.896
1973-74393637842.896
1974-75485744222.890
1975-76486044096.890
1976-77471243076.891
1977-78468143093.889
1978-79468639903.883
1979-80509443613.882
1980-81569246798.876
1981-82581046448.873


Last edited by tarheelhockey: 02-23-2012 at 12:16 PM.
tarheelhockey is online now   Reply With Quote
Old
02-23-2012, 11:33 AM
  #10
seventieslord
Moderator
 
seventieslord's Avatar
 
Join Date: Mar 2006
Location: Regina, SK
Country: Canada
Posts: 28,374
vCash: 500
.882
.876
.873

those are the 1980-1982 league averages when all goalies are included.

seventieslord is offline   Reply With Quote
Old
02-23-2012, 12:17 PM
  #11
tarheelhockey
Global Moderator
 
tarheelhockey's Avatar
 
Join Date: Feb 2010
Location: The Triangle
Country: United States
Posts: 45,239
vCash: 1020
Thanks, I went back and edited the table.

Interesting, if trivial, to note that short-term goalies have a net -.002 effect on overall save percentage.

tarheelhockey is online now   Reply With Quote
Old
02-23-2012, 12:44 PM
  #12
seventieslord
Moderator
 
seventieslord's Avatar
 
Join Date: Mar 2006
Location: Regina, SK
Country: Canada
Posts: 28,374
vCash: 500
Quote:
Originally Posted by tarheelhockey View Post
Thanks, I went back and edited the table.

Interesting, if trivial, to note that short-term goalies have a net -.002 effect on overall save percentage.
Indeed. I noticed that too. It is probably because most of those sub-1000 minute goalies were "replacement level".

seventieslord is offline   Reply With Quote
Old
02-23-2012, 05:45 PM
  #13
plusandminus
Registered User
 
Join Date: Mar 2011
Posts: 980
vCash: 500
Quote:
Originally Posted by tarheelhockey View Post
If anyone is planning to calculate this data, I have crunched some of the season numbers since 1998 and it would appear that subtracting empty-net goals and shots has an effect of roughly ~0.0025 on the overall league save percentage.

To state that more clearly:
If you're calculating Overall Goals - Overall Shots / Overall Goals,
you can just subtract about 0.0025 from the result to adjust for empty nets.

This is of course the modern rate. It's pretty much impossible to know whether it translates to earlier eras until someone enters all the old box scores into a database.
That is one of the things I have spent time doing. So I have the number of EN goals per game, or per team, or per season, and lots of more things, stored somewhere in my database.

Regarding save percentage, I would think that seasons with many powerplays would generally be bad for the save percentage. (It is possible to do formulas taking that into consideration. I myself have spent time doing it for more recent seasons.)

Regarding hockey stats in general, I'm still confused about some things:
* why doesn't people co-operate more into producing as reliable (factual) stats as possible? (I spend "enormous" amounts of time trying to correct and complete stats, and am constantly surprised it hasn't already been done by others.)
* why isn't there more research being done on different things (like "how much does skaters affect save percentage?", "how important are faceoffs?", "how much of team success can be attributed to goaltending?", etc., etc?)?
* why aren't readers interested in communicating about things like that? (This HOH section is generally very stats oriented, but for some reason interest seem low when it comes to researching things depper.)

For goalies:
* how consistant are single goalie save percentage from season to season?
* how much do save percentage change when goalies change team?
* how similar are save percentages for goalies on the same team?
* how accurate/reliable is save percentage stats anyway for older seasons?
etc...
Wouldn't it be interesting to try to find out? (I myself have not studied it. But if I would do, any result or presentation likely would get unnoticed, misunderstood or - rightly or wrongly - criticized. I'm the wrong person to do it, as basically no one here thinks I'm of any positive use whatsoever. Like yesterday, when I posted about face offs... it just got ignored. If I'm good at anything here, it might be being good at doing bad.)

I think it has basically never happened that anyone have asked me about anything, or shown any interest in my data or "research". Being a person having spent this much time, having built up all this data, and having spent relatively much time analyzing things, it's amazing how extremely poorly I seem to have performed on this board. It's almost as if I don't even exist.
Obviosly I'm procuding "products" of low quality or for which there is no demand, or are bad at "selling" them, or try to sell them in the wrong environment or at the wrong time. Maybe it would be easier if I would focus on writing/asking something about, for example, which is best of Lidstrom or Bourque. )

I've spent on average about 15-25 hours per week during the last year on hockey data, including about 50 hours per week during the summer vacation. I have followed hockey for at least 30 years, and am a professional programmer which makes it possible for me to do advanced things with stats, but here I've learnt that I'm basically a nobody and I suppose nothing will likely ever change that. When I for example communicate with former elite level players IRL, I get a much different response, i.e. they listen and confirm that I know what I'm talking about (but don't have the time that most people writing here seem to have). I have tried to get some feedback or co-operation here regarding studies, but to no interest. My "self-esteem" or "self-confidence" here is at an all-time low. Not being more fluent in the English language frustrates me too.

plusandminus is offline   Reply With Quote
Old
02-23-2012, 06:22 PM
  #14
seventieslord
Moderator
 
seventieslord's Avatar
 
Join Date: Mar 2006
Location: Regina, SK
Country: Canada
Posts: 28,374
vCash: 500
Quote:
Originally Posted by plusandminus View Post
Obviosly I'm procuding "products" of low quality or for which there is no demand, or are bad at "selling" them, or try to sell them in the wrong environment or at the wrong time.
this might be it.

seventieslord is offline   Reply With Quote
Old
02-23-2012, 07:45 PM
  #15
McGuillicuddy
Registered User
 
McGuillicuddy's Avatar
 
Join Date: Sep 2005
Country: Canada
Posts: 1,011
vCash: 500
Quote:
Originally Posted by tarheelhockey View Post
Well I'll be damned. Thanks SL.

That makes the league-average save% pretty easy then. Here you go.

SeasonTotal GoalsTotal SavesSave %
1967-68243826951.910
1968-69268929331.908
1969-70261929780.912
1970-71336434769.903
1971-72329233760.902
1972-73403348737.896
1973-74393637842.896
1974-75485744222.890
1975-76486044096.890
1976-77471243076.891
1977-78468143093.889
1978-79468639903.883
1979-80509443613.882
1980-81569246798.876
1981-82581046448.873
This is awesome. Do we have the numbers from 82-83 to present collated somewhere?

Edit: Never mind

http://www.hockey-reference.com/leagues/stats.html


Last edited by McGuillicuddy: 02-23-2012 at 07:54 PM.
McGuillicuddy is offline   Reply With Quote
Old
02-24-2012, 10:34 AM
  #16
tarheelhockey
Global Moderator
 
tarheelhockey's Avatar
 
Join Date: Feb 2010
Location: The Triangle
Country: United States
Posts: 45,239
vCash: 1020
Quote:
Originally Posted by plusandminus View Post
Regarding hockey stats in general, I'm still confused about some things:
* why doesn't people co-operate more into producing as reliable (factual) stats as possible? (I spend "enormous" amounts of time trying to correct and complete stats, and am constantly surprised it hasn't already been done by others.)
* why isn't there more research being done on different things (like "how much does skaters affect save percentage?", "how important are faceoffs?", "how much of team success can be attributed to goaltending?", etc., etc?)?
* why aren't readers interested in communicating about things like that? (This HOH section is generally very stats oriented, but for some reason interest seem low when it comes to researching things depper.)
Basically you have to be lucky enough to have other people online at the same time who know how to do it, and want to do it, and are able to do it. That's not easy. A lot of us only meet 1 or 2 of those qualifications. Personally I'm not good enough with numbers to feel trustworthy producing a database or doing deep stat analysis. I'm better off digging through reams of old newspapers to find a first-person account, and that's what I like to do, so I'm of little use to you here. I'm sure others on the forum will give you variants of that answer. You may in fact be the best qualified person available.

As to your self-esteem issue, just keep in mind that the majority of the users who read your work are lurkers who never say a word in response. That doesn't mean they aren't using it and building on it, or that they aren't appreciative. The rest of us are a bunch of blowhards and you shouldn't care what we think.

tarheelhockey is online now   Reply With Quote
Old
02-24-2012, 11:32 AM
  #17
tarheelhockey
Global Moderator
 
tarheelhockey's Avatar
 
Join Date: Feb 2010
Location: The Triangle
Country: United States
Posts: 45,239
vCash: 1020
Ok, I took a first run at calculating adjusted save% for 1968. The results are below.

Please please double-check these numbers to be sure I'm doing it right! I don't mind doing the leg-work on the rest of them, but I really would hate to put out a bunch of faulty information. And as I said before, this kind of analysis isn't really my thing.

I used the formula described by Hockey Outsider here: http://hfboards.hockeysfuture.com/sh...d.php?t=718221

Quote:
METHODOLOGY

To calculate adjusted games played:

1. Start with games played.
2. Adjustment #1: pro-rate games played to an 82 game schedule.

To calculate adjusted shots against:

1. Start with shots against
2. Adjustment #1: pro-rate shots against to an 82 game schedule
3. Adjustment #2: pro-rate shots against to a 29.0 shot per game environment

To calculate adjusted saves:

1. Start with shots against
2. Adjustment #1: pro-rate saves to an 82 game schedule
3. Adjustment #2: pro-rate saves to a 29.0 shot per game environment
4. Adjustment #3: pro-rate saves to a 90.5% save percentage environment

To calculate adjusted save percentage:

1. Divide adjusted shots against by adjusted saves

EXAMPLE

Letís look at Ed Belfourís 1993-94 season.

Games played: Eddie played 66.6 games. We pro-rate that to 82 games (the season was actually 84 games), so Belfour gets credit for 66.6 * 82 / 84 = 65.0 games.

Shots against: the Eagle faced 1,892 shots. We pro-rate that to 82 games (the season was actually 84 games), and we pro-rate that to 29 shots against (the season average was actually 29.9 shots), so Belfour faced 1,892 * 82 / 84 * 29 / 29.9 = 1,794 adjusted shots against.

Adjusted saves: the Billionaire made 1,714 saves. We pro-rate that to 82 games (the season was actually 84 games), we pro-rate that to 29 shots against (the season average was actually 29.9 shots) and we pro-rate that 90.5% save percentage (the season average was actually 89.5%), so Belfour made 1,714 * 82 / 84 * 29 / 29.9 * 90.5% / 89.5% = 1,643 adjusted saves.

Adjusted save percentage: 1,643 adjusted saves / 1,794 adjusted shots against = 91.6%.

Goalie data was pulled from seventieslord's chart here: http://hfboards.hockeysfuture.com/sh...d.php?t=698806


And seasonal averages were calulated upthread.

If numbers are one digit off, it may be because they were calculated with fractions out to the 7th or 8th decimal, then rounded for the sake of making this chart readable.

Name Mins GP (mins/60) Adj GP GA SA Adj SA GAA Sv% SA/60 SV Adj SV Adj SV%
Bower 2237 37.28 41.31 84 1278 1353.62 2.25 0.9343 34.28 1194 1257.70 0.929
Gamble 2203 36.72 40.69 85 1292 1368.44 2.32 0.9342 35.19 1207 1271.39 0.929
Villemure 200 3.33 3.69 8 121 128.16 2.40 0.9339 36.3 113 119.03 0.929
Favell 2192 36.53 40.48 83 1208 1279.47 2.27 0.9313 33.07 1125 1185.02 0.926
Parent 2248 37.47 41.52 93 1247 1320.78 2.48 0.9254 33.28 1154 1215.56 0.920
Worsley 2213 36.88 40.87 73 934 989.26 1.98 0.9218 25.32 861 906.93 0.917
Giacomin 3940 65.67 72.77 160 1876 1987.00 2.44 0.9147 28.57 1716 1807.54 0.910
Vachon 2227 37.12 41.13 92 1068 1131.19 2.48 0.9139 28.77 976 1028.07 0.909
Martin 1552 25.87 28.66 67 770 815.56 2.59 0.913 29.77 703 740.50 0.908
Hall 2858 47.63 52.78 118 1339 1418.22 2.48 0.9119 28.11 1221 1286.14 0.907
Maniago 2877 47.95 53.13 133 1492 1580.28 2.77 0.9109 31.12 1359 1431.50 0.906
Dejordy 2838 47.3 52.41 128 1435 1519.91 2.71 0.9108 30.34 1307 1376.76 0.906
Bassen 1299 21.65 23.99 62 678 718.12 2.86 0.9086 31.32 616 648.86 0.904
Cheevers 2645 44.08 48.85 125 1343 1422.46 2.84 0.9069 30.47 1218 1282.98 0.902
Smith, G 1130 18.83 20.87 60 642 679.99 3.19 0.9065 34.09 582 613.05 0.902
Hodge 3310 55.17 61.13 158 1673 1771.99 2.86 0.9056 30.33 1515 1595.82 0.901
Binkley 3141 52.35 58.01 150 1572 1665.01 2.87 0.9046 30.03 1422 1497.86 0.900
Dryden, D 1268 21.13 23.42 69 698 739.30 3.26 0.9011 33.03 629 662.56 0.896
Gill 270 4.5 4.99 13 130 137.69 2.89 0.900 28.89 117 123.24 0.895
Simmons 300 5 5.54 13 127 134.51 2.60 0.8976 25.4 114 120.08 0.893
Rutledge 2443 40.72 45.12 117 1143 1210.63 2.87 0.8976 28.07 1026 1080.73 0.893
Johnston 1525 25.42 28.16 73 707 748.83 2.87 0.8967 27.82 634 667.82 0.892
Sawchuk 1937 32.28 35.77 99 914 968.08 3.07 0.8917 28.31 815 858.48 0.887
Caron 60 1 1.11 4 37 39.19 4.00 0.8919 37 33 34.76 0.887
Crozier 1729 28.82 31.93 95 852 902.41 3.30 0.8885 29.57 757 797.38 0.884
Gardner 533 8.88 9.84 32 284 300.80 3.60 0.8873 31.97 252 265.44 0.882
Bauman 1294 21.57 23.90 75 656 694.81 3.48 0.8857 30.42 581 611.99 0.881
Edwards, R 2178 36.3 40.22 126 1095 1159.79 3.47 0.8849 30.17 969 1020.69 0.880
Wetzel 269 4.48 4.97 18 142 150.40 4.01 0.8732 31.67 124 130.62 0.868
Norris 334 5.57 6.17 22 168 177.94 3.95 0.869 30.18 146 153.79 0.864
Caley 30 0.5 0.55 3 20 21.18 6.00 0.850 40 17 17.91 0.845


Hopefully this is right, or close enough to be a useful start.

tarheelhockey is online now   Reply With Quote
Old
02-24-2012, 01:50 PM
  #18
seventieslord
Moderator
 
seventieslord's Avatar
 
Join Date: Mar 2006
Location: Regina, SK
Country: Canada
Posts: 28,374
vCash: 500
I find his formula over-complicated.

Personally I would do this - It wouldn't make the results worse, they would either be the same or better, but I am not sure which:

- calculate the league average error rate (inverse of sv%)
- calculate the goalie's error rate
- divide goalie's error rate by league average - this represents what "percentage" better than average he was
- convert that back into a sv%

seventieslord is offline   Reply With Quote
Old
02-24-2012, 01:55 PM
  #19
seventieslord
Moderator
 
seventieslord's Avatar
 
Join Date: Mar 2006
Location: Regina, SK
Country: Canada
Posts: 28,374
vCash: 500
I guess I should actually demonstrate this rather than just explain.

1994, Ed Belfour.

league average: .895. Or a 10.5% error rate. belfour's sv% is .905, or a 9.5% error rate. 9.5/10.5 = .902. So you could say Eddie was letting in goals at just 90.2% of the average rate. If the normalization number is .905 (9.5% error rate) and eddie was at .902 of that, his adjusted error rate is .857, or .86, leaving him with a .914 adjusted sv%.

I wonder which one is more mathematically sound?

seventieslord is offline   Reply With Quote
Old
02-24-2012, 02:00 PM
  #20
tarheelhockey
Global Moderator
 
tarheelhockey's Avatar
 
Join Date: Feb 2010
Location: The Triangle
Country: United States
Posts: 45,239
vCash: 1020
Quote:
Originally Posted by seventieslord View Post
I guess I should actually demonstrate this rather than just explain.

1994, Ed Belfour.

league average: .895. Or a 10.5% error rate. belfour's sv% is .905, or a 9.5% error rate. 9.5/10.5 = .902. So you could say Eddie was letting in goals at just 90.2% of the average rate. If the normalization number is .905 (9.5% error rate) and eddie was at .902 of that, his adjusted error rate is .857, or .86, leaving him with a .914 adjusted sv%.

I wonder which one is more mathematically sound?

For those who haven't visited the other thread, Hockey Outsider had it at .916.

tarheelhockey is online now   Reply With Quote
Old
02-24-2012, 03:07 PM
  #21
seventieslord
Moderator
 
seventieslord's Avatar
 
Join Date: Mar 2006
Location: Regina, SK
Country: Canada
Posts: 28,374
vCash: 500
Someone with number knowledge (overpass?) should come in here and comment. These results are close but still different. and long-term it could affect matters.

I'm leaning towards my method just because mine is based on division and not on a "points above or below the average" method (i.e. .920 vs. .910 is not the same as .900 vs. .890). but I also admit I don't fully understand HO's method to say much more than it's awfully complex.

seventieslord is offline   Reply With Quote
Old
02-24-2012, 04:59 PM
  #22
reckoning
Registered User
 
reckoning's Avatar
 
Join Date: Jan 2005
Country: Canada
Posts: 5,829
vCash: 500
Slightly off-topic, but...

Would it be a better idea to somehow convert save percentage to a whole number that would represent the amount of goals a particular goalie prevented for his team? It might be a good way to compare say a goalie who is .929 in 40 games with somebody who is .925 in 70 games. A forward who had the highest PPG in a season, but only played half the year, would never be considered for an All-Star spot. But a goalie in the same situation would. It might be a better way to look at career numbers as well, since career save percentage obviously benefits goalies who retire early.

reckoning is online now   Reply With Quote
Old
02-24-2012, 05:00 PM
  #23
TheDevilMadeMe
Registered User
 
TheDevilMadeMe's Avatar
 
Join Date: Aug 2006
Location: Brooklyn
Country: United States
Posts: 45,406
vCash: 500
Quote:
Originally Posted by reckoning View Post
Slightly off-topic, but...

Would it be a better idea to somehow convert save percentage to a whole number that would represent the amount of goals a particular goalie prevented for his team? It might be a good way to compare say a goalie who is .929 in 40 games with somebody who is .925 in 70 games. A forward who had the highest PPG in a season, but only played half the year, would never be considered for an All-Star spot. But a goalie in the same situation would. It might be a better way to look at career numbers as well, since career save percentage obviously benefits goalies who retire early.
Someone correct me if I'm wrong, but isn't this what goals versus threshold tries to do?

TheDevilMadeMe is offline   Reply With Quote
Old
02-24-2012, 05:10 PM
  #24
Hockey Outsider
Registered User
 
Hockey Outsider's Avatar
 
Join Date: Jan 2005
Country: Canada
Posts: 3,796
vCash: 500
My formula is very simple once you ignore the part about the games played & shots against adjustment. Essentially it's (goalie's save percentage) multiplied by (all-time average save percentage) divided by (save percentage for that year).

In the example above, it's 90.6% (Belfour's sv% that year) * 90.5% (historical norm) / 89.5% (average save percentage for the 1993-94 season). This works out to 91.6%.

To clarify - it isn't based on points above/below average. My formula would (quite appropriately, I think) see a 90.0% save percentage in a year that featured an average of 89.0% to be superior to a 92.0% save percentage in a year that featured an average of 91.0%.

The rest of the formula is simply to put goalies onto the same page in terms of games played and shots against. For example, a goalie who played in 42 of his teams 84 games in 1993-94 would be scaled down to playing in 41 of his team's games in a hypothetical 82 game season. Similar adjustment for shots against.

(For the record I don't have a strong preference for my formula posted above - if someone can do it better, I welcome it).

Hockey Outsider is offline   Reply With Quote
Old
02-24-2012, 05:29 PM
  #25
Hockey Outsider
Registered User
 
Hockey Outsider's Avatar
 
Join Date: Jan 2005
Country: Canada
Posts: 3,796
vCash: 500
Quote:
Originally Posted by plusandminus View Post
That is one of the things I have spent time doing. So I have the number of EN goals per game, or per team, or per season, and lots of more things, stored somewhere in my database.

Regarding save percentage, I would think that seasons with many powerplays would generally be bad for the save percentage. (It is possible to do formulas taking that into consideration. I myself have spent time doing it for more recent seasons.)

Regarding hockey stats in general, I'm still confused about some things:
* why doesn't people co-operate more into producing as reliable (factual) stats as possible? (I spend "enormous" amounts of time trying to correct and complete stats, and am constantly surprised it hasn't already been done by others.)
* why isn't there more research being done on different things (like "how much does skaters affect save percentage?", "how important are faceoffs?", "how much of team success can be attributed to goaltending?", etc., etc?)?
* why aren't readers interested in communicating about things like that? (This HOH section is generally very stats oriented, but for some reason interest seem low when it comes to researching things depper.)

For goalies:
* how consistant are single goalie save percentage from season to season?
* how much do save percentage change when goalies change team?
* how similar are save percentages for goalies on the same team?
* how accurate/reliable is save percentage stats anyway for older seasons?
etc...
Wouldn't it be interesting to try to find out? (I myself have not studied it. But if I would do, any result or presentation likely would get unnoticed, misunderstood or - rightly or wrongly - criticized. I'm the wrong person to do it, as basically no one here thinks I'm of any positive use whatsoever. Like yesterday, when I posted about face offs... it just got ignored. If I'm good at anything here, it might be being good at doing bad.)

I think it has basically never happened that anyone have asked me about anything, or shown any interest in my data or "research". Being a person having spent this much time, having built up all this data, and having spent relatively much time analyzing things, it's amazing how extremely poorly I seem to have performed on this board. It's almost as if I don't even exist.
Obviosly I'm procuding "products" of low quality or for which there is no demand, or are bad at "selling" them, or try to sell them in the wrong environment or at the wrong time. Maybe it would be easier if I would focus on writing/asking something about, for example, which is best of Lidstrom or Bourque. )

I've spent on average about 15-25 hours per week during the last year on hockey data, including about 50 hours per week during the summer vacation. I have followed hockey for at least 30 years, and am a professional programmer which makes it possible for me to do advanced things with stats, but here I've learnt that I'm basically a nobody and I suppose nothing will likely ever change that. When I for example communicate with former elite level players IRL, I get a much different response, i.e. they listen and confirm that I know what I'm talking about (but don't have the time that most people writing here seem to have). I have tried to get some feedback or co-operation here regarding studies, but to no interest. My "self-esteem" or "self-confidence" here is at an all-time low. Not being more fluent in the English language frustrates me too.
A few thoughts:

1. Don't take things so personally. I've posted a lot of quantitative studies over the years. Sometimes they've sparked pages of thoughtful debate and discussion. Sometimes they get angry responses from people who are unwilling to challenge conventional thinking. Sometimes they get no virtually no responses. Not everyone will think that your study is as brilliant as you do - that's just the way it is.

2. I share your frustration. When I started digging in-depth into hockey stats (I think this was in 2006), there was virtually no useable information online. For example, one time I wanted to analyze goalie stats from the Original Six era, and I had to manually enter it into Excel, line-by-line, from "The Hockey Compendium". It took me many hours (I shudder thinking about how many). Things are better today (look at the data accumulated here, on Yahoo's Hockey Analysis Group, and on hockey-reference.com). Still, a lot of very basic thing are lacking (such as a comprehensive database of goalie stats, broken down by team, from 1968-present). There are a lot of projects I'd like to work on, but now that I'm older and busier, I'm simply not willing to spend, say, 25 hours accumulating data to do five hours worth of analysis.

3. I think that people here generally have more respect for you if you're a "hockey guy" rather than a "stats guy". (Makes sense, seeing as this is a hockey forum, not a statistics forum). People listen to Overpass and Seventies (and me, at least when I posted on a regular basis) because they can step back and describe what the numbers mean. They understand the context and relevance of the stats. To be honest, I find that you post a lot of numbers but rarely step back and explain (or even consider) their implications. I don't mean to sound harsh, but this is my honest opinion.

For example, you once posted something about scoring streaks. When I asked why we would care about scoring streaks (which are, IMO, due to randomness explained by basic probability theory), you told me that you weren't trying to say that scoring streaks are important. This begs the question - why invest time and effort into studying something that has little if any relevance or practical implications? Nobody wants to read about numbers for the sake of numbers - they want numbers that can shed light on meaningful issues, like which goalie had the best peak and/or career? How does penalty killing impact Norris trophy voting? Who was at fault for the Chicago Blackhawks' dissapointing playoff performances in the 1960s? Was Dionne really as bad a playoff performer as people suggest? How important was Sakic to the Avalanche? Etc.

There is an audience for statistical studies but the key is to sell the topic/issue/problem, not the numbers!

4. I have no right to tell you how to spend your free time, but spending 800 - 1,300 hours a year on collecting hockey data seems awfully high! This is coming from someone who used to spend a lot of time doing the same thing (though never that much). Why not cut back from 1,200 hours a year to say 750 (which still seems really high - over two hours per day!) - that gives you time to watch a new movie each week, go to the gym for a 1-hour workout three times a week, and half an hour a day learning how to play a new instrument. Hockey is great, but I'd be willing to bet that this would make you a happier, more well-rounded person.


Last edited by Hockey Outsider: 02-24-2012 at 05:37 PM.
Hockey Outsider 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 11:23 AM.

monitoring_string = "e4251c93e2ba248d29da988d93bf5144"

vBulletin Copyright ©2000 - 2016, Jelsoft Enterprises Ltd.
HFBoards.com is a property of CraveOnline Media, LLC, an Evolve Media, LLC company. ©2016 All Rights Reserved.