Greatest Single-Season Goalkeeping Performances

[Steve Holroyd]

Borrowing a few concepts from baseball and The Hockey Compendium, I've come up with a quick list of the greatest single-season performances by goalkeepers in U.S. Division 1 leagues.

Statistics can be deceptive in any scenario, but they seem particularly useless when evaluating soccer goalkeepers. Unlike hockey, where a goaltender is regularly involved in a game, a soccer keeper can see no shots and still garner the clean sheet. Meanwhile, those who remember Portland's Mick Poole know he labored with a lot of bad defensive teams and often topped people's "he'd be great if he was with the Cosmos" lists (in hockey this was known as the "Gilles Meloche Syndrome.") Looking at his stats today, though, you wouldn't be impressed. However, he was great.

What I've done is hardly rocket science, and is barely touching the surface of trying to provide relative meaning to goalkeeper stats. Using the GAA stat, I've compiled a list of the best performances as far as the difference between the leading goalkeeper and the second-best goalkeeper, the theory being that a 0.80 GAA doesn't mean much if the runner-up came in at 0.89--obviously, both benefitted from a low-scoring year. However, someone coming in at less than a goal a match while the runner up let in 1.30 a game obviously had a great year, going above-and-beyond the "average" generated by league scoring as a whole.

Using nothing more than the difference, here's the top 8:

1969 Manfred Kammerer (Atlanta Chiefs) 1.07 (-.50)

1983 Tino Lettieri (Vancouver Whitecaps) 0.86 (-.40)

1973 Bob Rigby (Philadelphia Atoms) 0.62 (-.38)

1984 Paul Hammond (Toronto Blizzard) 1.16 (-.34)

1925-26 Fall River Marksmen 1.16 (-.32)

1971 Mirko Stojanovic (Dallas Tornado) 0.73 (-.32)

1924-25 Findlay Kerr (Fall River Marksmen) 0.86 (-.30)

1979 Phil Parkes (Vancouver Whitecaps) 0.96 (-.30)

Some notes:

Rigby's GAA was the NASL record. Obviously, it was not earned in a weak season for scoring, making it all the more impressive.

The 1925-26 figure comes from the combined GAAs of Findlay Kerr and Tommy Blair. I have not broken that season down yet. Kerr played 28 games, while Blair split time between Boston and Fall River. This could drop from the list, especially since I get the feeling Kerr and Blair finished 1-2 that year.

As a frame of reference, keep in mind Tony Meola's stunning 2000 season (16 shutouts and a league-leading 0.92 GAA) was only .08 ahead of the runner-up, Kevin Hartman.

For the greatest season ever, one has to go to a Division "2.5" league--a young Kasey Keller (F.C. Portland) led the Western Soccer Alliance in GAA with an 0.38 figure, 0.62 ahead of his competition, in 1989.

I'd like to check these numbers against the average goals per game league-wide in a given season; I've long felt that Kerr's 24-25 season is the greatest ever, as the American Soccer League was a very high-scoring league (he put up a 0.86 GAA and 20 shutouts the same year Archie Stark scored 67 goals in 44 games). I have all the ASL stats from the 1920s, but it will take some time.

I recognize this still doesn't save the Mick Poole's of the world. But I'm working on it. ;)

Anyway, this is a start. Whaddya think?

[NoSix]

Given the stats available, it may be the best you can do. I still feel that GAA is more a measure of team defense than goalkeeping ability. If I were going to rate keepers statistically I would look at a combination of save percentage and catch/punch percentage. Unfortunately, MLS doesn't seem to track unsuccessful catch/punch attempts, so in practice that would be hard to do right now.

[ChrisE]

So, Steve, I don't really think that you're going about this the right way. With what you're doing, you get, at best, a record of what goalkeepers had individually great seasons. But, do you really think that Bob Rigby 1973 was really that much better than Bob Rigby 1974 or Bob Rigby 1972? I think that a large part of GAA is simply how good the defense is in front of you, so it's inevitable that what you're measuring here is mostly how good a defense mixed in with a little goalkeeping ability and statistical variance.

The fact, for example, that Tony Meola only gave up .82 gpg in 2000 isn't lessened by the fact that Kevin Hartman only gave up .9 that year. That simply means that both of those goalies had tremendous years (or tremendous defences). It's not like goalscoring was down across the board in MLS that year (cf. Nick Rimando, 1.84 GAA). I think the reason that all of your goalies are in the pre-MLS days is that the quality of teams varied much more widely back then, not because the goalkeepers were any better (e.g. Tim HOward doesn't make your list, but I'd bet my arms that he's a better goalkeeper than Findlay Kerr was).

I've got to side with NoSix on this, and say that the best way to evaluate goalies is by using shot percentage and catches/punches (I don't think that should be a percentage, but whatever). Beineke ran some interesting statistics about save% and offensive traps in this thread:

http://www.bigsoccer.com/forum/showthread.php?s=&threadid=74498

As a sidenote, where did you get your stats for these goalkeepers?

[Steve Holroyd]

Originally posted by ChrisE
But, do you really think that Bob Rigby 1973 was really that much better than ... Bob Rigby 1972?

Yes, to a point of absolute certaintude...1973 was Rigby's rookie year. ;)


As a sidenote, where did you get your stats for these goalkeepers?

I've compiled an all-time U.S. Division One goalkeeper database. The NASL and MLS stats are taken from media guides, etc. The 1920s ASL and 1894 APFL stats are from my own research (although I am indebted to Colin Jose for making my life a little easier in this regard).

I hear your criticisms about GAA being more a team defense indicator than a measure of GK effectiveness. The thought of using save percentage (I have the figures calculated in my database) has crossed my mind. It is an interesting statistic...for example, Shep Messing is often criticized as overrated and overhyped, yet he has the highest career save percentage out of any of his notable contemporaries (Mayer, Mausser, Rigby, DuBose, etc.) The problem is save percentage by itself is just as unrevealing as GAA...a guy with a .990 percentage saving 3 shots a game is not as impressive as one with a .900 percentage facing 15 shots.

The Hockey Cpmpendium raises the same concern for hockey goalies, and came up with an endurance formula which is appealing. Maybe I'll apply it to the NASL and MLS years and see what comes up.

[Steve Holroyd]

Originally posted by ChrisE
I think the reason that all of your goalies are in the pre-MLS days is that the quality of teams varied much more widely back then, not because the goalkeepers were any better (e.g. Tim HOward doesn't make your list, but I'd bet my arms that he's a better goalkeeper than Findlay Kerr was).

FWIW, I don't follow your logic on this...where the comparison is only between the #1 and #2 goalkeepers, how does the (admittedly) varied quality of NASL teams come into play? I can see this being a factor if I was comparing GAAs against the league goals-per-game average; then, the fact that everyone got to beat up on the 1968 Dallas Tornado would have some significance. But with the comparison being limited to #1 and #2, I think any league-wide disparity is less of an issue.

[ChrisE]

Originally posted by Steve Holroyd
Yes, to a point of absolute certaintude...1973 was Rigby's rookie year. ;)

touche




I hear your criticisms about GAA being more a team defense indicator than a measure of GK effectiveness. The thought of using save percentage (I have the figures calculated in my database) has crossed my mind. It is an interesting statistic...for example, Shep Messing is often criticized as overrated and overhyped, yet he has the highest career save percentage out of any of his notable contemporaries (Mayer, Mausser, Rigby, DuBose, etc.) The problem is save percentage by itself is just as unrevealing as GAA...a guy with a .990 percentage saving 3 shots a game is not as impressive as one with a .900 percentage facing 15 shots.


I don't know anything about the NASL, so I don't really have much to say about Shep Messing. However, I don't follow why a guy with a .99 save percentage against 3 shots is not as impressive as the .9 guy facing 15 shots.

Are you simply assuming that if a guy is facing 3 shots instead of 15, the 3 shots he will face will not be as difficult to stop? I don't think that this is obvious at all, but I would be happy to hear an explanation.

Actually, the offside trap stats would seem to indicate the opposite - although a goalie whose team heavily employs the offside trap will face fewer shots, many that he does face will be significantly higher % shots (e.g. failed trap leads to one-on-one opportunity).

(for the record, I would take the .99 guy over the .9 guy any day; for every 10 goals .9 gives up, .99 gives up one)

[kenntomasch]

Originally posted by ChrisE

I don't know anything about the NASL, so I don't really have much to say about Shep Messing. However, I don't follow why a guy with a .99 save percentage against 3 shots is not as impressive as the .9 guy facing 15 shots.

I don't know that Steve meant literally .99 vs. .90 (nobody is going to stop 99% of the shots against him), but that all things being equal, the guy who faces more shots per game is being tested more often, and therefore his accomplishments are more impressive.


(for the record, I would take the .99 guy over the .9 guy any day; for every 10 goals .9 gives up, .99 gives up one)

The MLS career record was .756 by Brad Friedel coming into this year.

I don't know what the NASL record was, but Rigby's 1973 might be it. Here are the save percentages and SOG/90 minutes numbers for the NASL 'keepers Steve mentioned above:


1969 Manfred Kammerer (Atlanta Chiefs) - .789 (5.07 SOG/90 minutes)
1983 Tino Lettieri (Vancouver Whitecaps) - .864 (6.38 SOG/90 minutes)
1973 Bob Rigby (Philadelphia Atoms) - .907 (6.69 SOG/90 minutes)
1984 Paul Hammond (Toronto Blizzard) - .706 (3.95 SOG/90 minutes)
1971 Mirko Stojanovic (Dallas Tornado) - .892 (6.75 SOG/90 minutes)
1979 Phil Parkes (Vancouver Whitecaps) - .775 (4.29 SOG/90 minutes)

I looked at some other 'keepers, and about six shots on a goal a game is what you find from the guys with the higher percentages. I guess what happens is, if you're a defense that gives up more than six shots on goal a game, chances are your keeper isn't going to be saving a great percentage of them. Chances are you've got some problems defensively.

[Steve Holroyd]

My analogy about a guy stopping all of fewer shots being a lesser goalkeeper stopping a lower percentage of more shots may borrow too much from hockey concepts to be effective in soccer. This is the "endurance" factor the hockey book I referenced tried to address.

Having played some goal, I can empathize with your stats showing the impact the offsides trap can have on the "quality" of shots a keeper may be called upon to save. Offensive opportunities "breaking" a trap are often of the breakaway or "odd man rush" variety, forcing a keeper to come off his line earlier, etc. Because of the size of a soccer net, the "quality" of shots faced becomes more relevant than simple "endurance" (i.e., the number of shots faced). So, again, maybe my borrowing from hockey concepts leaves a lot to be desired.

Because only Americans bother to quantify soccer goalkeeper performance (a small percentage of Americans at that) and have only borrowed from hockey concepts to do so, a lot of this is underdeveloped. Much like QB ratings, maybe a figure combining GAA, the difference between the GAA and the league GAA (a plus/minus type of stat maybe), something that figures the offside trap and the frequency of its employ (a "strength of schedule" concept), balls touched/punched/saved, total shots faced per game, etc. etc. might reveal a GK rating that has some actual quantitative use.

One of the reasons I started this thread was to generate just this type of discussion.

You don't know anything about the NASL? Wow...thanks for making me feel old. ;)

[ChrisE]

Originally posted by Steve Holroyd
My analogy about a guy stopping all of fewer shots being a lesser goalkeeper stopping a lower percentage of more shots may borrow too much from hockey concepts to be effective in soccer. This is the "endurance" factor the hockey book I referenced tried to address.

I understood the analogy, I just don't know how reliable it is.



Because only Americans bother to quantify soccer goalkeeper performance (a small percentage of Americans at that) and have only borrowed from hockey concepts to do so, a lot of this is underdeveloped. Much like QB ratings, maybe a figure combining GAA, the difference between the GAA and the league GAA (a plus/minus type of stat maybe), something that figures the offside trap and the frequency of its employ (a "strength of schedule" concept), balls touched/punched/saved, total shots faced per game, etc. etc. might reveal a GK rating that has some actual quantitative use.


I still don't trust the ranking a goalie by GAA, simply because it is so reliant on how good his defense was, and what kind of game his team is playing. For example, this year, Zach Thornton (1.22), Kevin Hartman (1.12), Nick Rimando (1.12), and Pat Onstad (1.03) had GAA's well below Tim Howard's (1.33). However, I don't think that there's many people who would claim that any of these guys were better GK's this year than Tim. Meanwhile, of those guys, only Hartman (.796) and Rimando (.778) had better save %'s than Tim (.777), and none had a higher C/P per 90.

Going back to your .99/3, .9/15 analogy, I think that GAA is an even worse measure than save percentage in this case. These guy's respective GAA would be .03 and 1.5. The guy who only faces 3 shots is going to look 50 times better than the guy who faces 15, if you're just basing it on GAA, whereas based on save percentage, he's only 10 times better. I think it's pretty clear why GAA is far far too dependent on a player's defense to be used to evaluate a goalie. It punishes goalies for facing more shots, whereas you want them to be rewarded.

I don't know if you've read any Peter Hirdt's columns, but he had one that seemed to indicate that the Catch/Punch statistic was much more important than is usually assumed:
http://www.mlsnet.com/content/03/analyze0801.html



You don't know anything about the NASL? Wow...thanks for making me feel old. ;)

Well, I didn't know that Shep Messing played goal... :)

[NoSix]

Fundamentally, a goalkeeper has two (individual) defensive responsibilities - to save shots on goal and catch/punch passes by the opposing team. Therefore, I would propose to track:

Save Attempts (Shots on Goal Against) (SA)
Saves (S)
Catch/Punch Attempts (CPA)
Catches/Punches (CP)

Then I would use the weighted average of the save and catch/punch percentage as my goalkeeper rating (GR):

Save Percentage (SP): SP=S/SA
Catch/Punch Percentage (CPP): CPP=CP/CPA

Goalkeeper Rating (GR):
GR=(SA*SP+CPA*CPP)/(SA+CPA)=(S+CP)/(SA+CPA)

As an example, I calculated GR's for 2003 MLS goalkeepers. Since MLS fails to track CPA, I assume all keepers have a 100% CPP. This still serves the purpose of weighting the relative importance of shot blocking expertise and penalty area mastery for each keeper.

PLAYER SA S CP GR
Tim Howard 92 74 65 0.885
Pat Onstad 132 103 117 0.884
Jonny Walker 69 55 44 0.876
Kevin Hartman 184 149 81 0.868
Nick Rimando 129 100 71 0.855
Adin Brown 132 95 115 0.850
Zach Thornton 160 123 82 0.847
Jon Busch 139 104 88 0.846
Scott Garlick 121 83 97 0.826
Tony Meola 165 121 82 0.822
DJ Countess 173 115 88 0.778

According to this simple statistic, the top 3 keepers in MLS this year were Howard, Onstad, and Walker. I think it is a superior individual rating compared to GAA.

[NoSix]

List includes MLS keepers with > 2700 minutes played over the four seasons:

GOALKEEPER MIN SHTS SVS C/P Years GR
Joe Cannon 7101 451 338 292 00-02 0.848
Jon Busch 3430 224 167 143 02-03 0.845
Zach Thornton 10026 601 449 322 00-03 0.835
Kevin Hartman 7873 490 371 226 00-03 0.834
Tim Howard 6849 544 400 320 00-03 0.833
Tony Meola 8668 533 395 247 00-03 0.823
Nick Rimando 9208 619 449 338 00-03 0.822
Scott Garlick 9054 667 466 356 00-03 0.804
Adin Brown 6971 519 357 287 00-03 0.799
Matt Jordan 7729 484 324 293 00-02 0.794
Jeff Causey 2975 201 125 163 00-01 0.791
Tom Presthus 5415 408 287 149 00-02 0.783
Mike Ammann 3699 273 181 146 00-01 0.780
David Kramer 2999 200 128 127 00,02 0.780

By year the best performances were by Busch in 2002, Rimando in 2001, and Meola in 2000. As noted in the previous post, Busch's performance dropped off somewhat in 2003, so it will be interesting to see if he can maintain the fairly high level he has set for himself.

Not hard to see why Hankinson preferred Cannon over Garlick, but you certainly could (and did) question his timing!

[beineke]

Originally posted by NoSix
By year the best performances were by Busch in 2002, Rimando in 2001, and Meola in 2000. As noted in the previous post, Busch's performance dropped off somewhat in 2003, so it will be interesting to see if he can maintain the fairly high level he has set for himself.


This seems to suggest that a keeper's "performance" is highly correlated with the quality of the team in front of him.

[NoSix]

Originally posted by beineke
This seems to suggest that a keeper's "performance" is highly correlated with the quality of the team in front of him.

An astute observation, and a testable hypothesis, perhaps you would be interesting in running some numbers? It seems reasonable to me to expect that an outstanding season by a goalkeeper would correlate with good results by his team.

Certainly the Metros performance over the first half of 2003 does not explain Tim Howard's stellar rating.

[NoSix]

All-Time MLS Goalkeeping Leaders by GAA and GR (minimum 3500 minutes):

GR Rank Goalkeeper GR GAA GAA Rank
1 Cannon, Joe 0.845 1.25 3
2 Hartman, Kevin 0.834 1.12 1
3 Thornton, Zach 0.830 1.23 2
4 Howard, Tim 0.829 1.49 8
5 Dodd, Mark 0.825 1.57 10
6 Campos, Jorge 0.822 1.37 4
7 Rimando, Nick 0.822 1.49 6
8 Meola, Tony 0.820 1.44 5
9 Jordan, Matt 0.813 1.51 9
10 Simpson, Mark 0.792 1.49 7

Sorry, but I'm not buying that Mark Simpson was a better goalkeeper than Tim Howard or Mark Dodd.

FWIW, career Goalkeeper Ratings for MLS (minimum 2700 minutes):
RANK GOALKEEPER MIN SA S CP GR
1 Friedel, Brad 3419 234 177 141 0.848
2 Cannon, Joe 9261 580 433 368 0.845
3 Busch, Jon 3430 224 167 143 0.845
4 Hartman, Kevin 13769 805 602 416 0.834
5 Hahnemann, Marcus 5847 439 317 293 0.833
6 Thornton, Zach 15201 891 653 511 0.830
7 Howard, Tim 7681 607 442 359 0.829
8 Dodd, Mark 7977 691 495 429 0.825
9 Rimando, Nick 9208 619 449 338 0.822
10 Campos, Jorge 4327 273 197 153 0.822
11 Meola, Tony 17517 1212 886 595 0.820
12 Garlick, Scott 14901 1071 763 594 0.815
13 Jordan, Matt 10358 660 458 421 0.813
14 Sommer, Juergen 4868 360 249 225 0.810
15 Zenga, Walter 4051 254 168 192 0.807
16 Brown, Adin 6971 519 357 287 0.799
17 Feuer, Ian 4032 279 183 198 0.799
18 Reis, Matt 3239 179 125 88 0.798
19 Simpson, Mark 4531 265 181 139 0.792
20 Presthus, Tom 8766 614 425 275 0.787
21 Cassar, Jeff 4161 293 191 180 0.784
22 Causey, Jeff 6885 487 307 345 0.784
23 Lagerwey, Garth 4242 316 212 149 0.776
24 Dougherty, Mark 10766 759 505 354 0.772
25 Kramer, David 8996 627 406 340 0.771
26 Ammann, Mike 11155 743 478 393 0.767
27 Oshoniyi, Bo 3294 259 169 125 0.766
28 Salzwedel, Dave 2847 198 125 105 0.759

[mpruitt]

Again from hockey. When I started reading this thread I couldn't really understand what was being talked about in terms of any sort of Replacement Goalie Average. I've been reading a little bit of http://www.puckerings.com, a hockey metric site to see if I couldn't understand some of this stuff beter. I found this article (http://www.puckerings.com/research/nwinpct.html) by Iain Fyffe to be very interesting. It talks about what they've termed the Neutral Winning Percentage, which is based off a similar theory in baseball. "Support-neutral records remove the effect of team offence from a pitcher?s won-lost record, thereby producing a result that more fairly evaluates a pitcher's performance."

"Each goaltender's season is broken down by the number of times which he allows zero goals in a game, one goal in a game, two goals, and so on. The records for all teams in the league are then broken down based on the number of goals they allowed in each game. We then use these league numbers to compute an expected record for the goaltender, based on the number of times he allowed each number of goals against.

As I can explain this they took league-wide numbers and looked at the number of times a goalie had given up 1 goal, 2 goals, 3 goals etc etc and compared that with the percentage of times that goalie won the game as a result of giving up that number of goals. From their example, in the period they looked at when a goalie posted a shut out his team won the game 92% of the time and tied 8% of the time, obviously never losing. When he gave up only one goal that average drops to 78% win 15% tie 7% loss.

This method could easily be applied to soccer.

He goes onto talk about some of the things which I think were similar to what was mentioned above.
In the essay entitled ?Goaltender Perseverance: a Useless Stat?, I demonstrate that the number of shots a goalie faces is a function of his team. Thus, a goaltender who faces a large number of shots is being unfairly penalized for playing for his particular team. This distortion must be removed to effectively compare goaltenders.

To control for this distortion, we should not evaluate a goalie based on the actual number of goals he allows, but on the number of goals he would allow when facing an average number of shots. In this was, we remove the bias resulting from facing a high or low number of shots.

Unfortunately, some distortion will remain. Team defence affects not only the number of shots faced, but the quality of shots faced. However, this distortion cannot be removed because we cannot determine the effect it has on a goalie?s save percentage. Still, this distortion is present in all methods currently used for evaluating goaltenders (including save percentage), so even if it is present in Neutral Winning Percentage, this method is still an improvement.

He has calculated that by
LgShotsPerGame x (LgShootPct/OppShootPct)

Where:

LgShotsPerGame is the total shots in the league divided by the total games played in the league

LgShootPct is the total goals in the league divided by the total shots in the league

OppShootPct is the opponent?s goals divided by the opponent?s shots
Which is something I'm having a little trouble wrapping my mind around at the moment.

[mpruitt]

ChrisE, that forumla really provides an interesting justiposition to the oen that you and beinke came up with. I wonder what you think of its merits vs. yours?

[NoSix]

Which is something I'm having a little trouble wrapping my mind around at the moment.

It looks to be a fairly straightforward strength of opponent adjustment. If team A concedes 20 shots in a match against team B, and team B averages 12 shots per game, then against team C, which averages 6 shots per game, team A would be expected to concede 20*6/12=10 shots . Take the actual shots conceded against team C, apply the ratio in reverse, and you can calculate an adjusted number of shots faced which takes into account the strength of the opponent.

[mpruitt]

It looks to be a fairly straightforward strength of opponent adjustment. If team A concedes 20 shots in a match against team B, and team B averages 12 shots per game, then against team C, which averages 6 shots per game, team A would be expected to concede 20*6/12=10 shots . Take the actual shots conceded against team C, apply the ratio in reverse, and you can calculate an adjusted number of shots faced which takes into account the strength of the opponent.

Oh I get it now because the second part of the equation gives the percentage of the relativeness of being better or worse then the league average of shots per game and how many of those shots should go in, yes?

EDIT: I think this answers my question. From the article:
To control for this distortion, we should not evaluate a goalie based on the actual number of goals he allows, but on the number of goals he would allow when facing an average number of shots. In this was, we remove the bias resulting from facing a high or low number of shots.

What do you think of the method in the article I linked above of using number of goals allowed as an expected win percentage for keepers? I think it does do a much fairer job of taking out team play then GAA but obviously it's still prevalent there because you're saying that if Goalkeeper A only gives up a certain amount of goals his team should win about X% of the time. What if you have a goalkeeper who's team is predominently horriable at getting goals. He's still being penalized because the way that breakdown works is that it's still using leaguewide averages? In the article he says:
This winning percentage is offence-neutral. The number of goals the goaltender’s team scores has no effect upon the percentage. The bias resulting from playing for a high- or low-scoring team is eliminated.
He refers to it as being 'eliminated' but I don't see how that is, it's marginalized but eliminated? How? Why? Is he off the mark?

[NoSix]

What do you think of the method in the article I linked above of using number of goals allowed as an expected win percentage for keepers?

I think it is a fundamentally flawed argument. Goalkeepers don't win or lose soccer games, teams do. Ideally what you want to do is separate individual goalkeeper performance from team performance (so that eventually you might understand how individual goalkeeper performance contributes to team performance). I think my goalkeeper rating stats do that pretty well. In principle, you could use a method analogous to the one proposed in this paper to calculate save and catch/punch percentages adjusted for strength of team and opponent, but that would be a hell of a lot work for perhaps not much marginal benefit. I'm guessing team and opponent strength tend to balance out over the length of a typical goalkeeper's career.

Here again is my ranking of All-Time MLS goalkeepers (1996-2003) by GK rating, formatted this time:


RANK GOALKEEPER MIN SA S CP GR
1 Friedel, Brad 3419 234 177 141 0.848
2 Cannon, Joe 9261 580 433 368 0.845
3 Busch, Jon 3430 224 167 143 0.845
4 Hartman, Kevin 13769 805 602 416 0.834
5 Hahnemann, M 5847 439 317 293 0.833
6 Thornton, Zach 15201 891 653 511 0.830
7 Howard, Tim 7681 607 442 359 0.829
8 Dodd, Mark 7977 691 495 429 0.825
9 Rimando, Nick 9208 619 449 338 0.822
10 Campos, Jorge 4327 273 197 153 0.822
11 Meola, Tony 17517 1212 886 595 0.820
12 Garlick, Scott 14901 1071 763 594 0.815
13 Jordan, Matt 10358 660 458 421 0.813
14 Sommer, Juergen 4868 360 249 225 0.810
15 Zenga, Walter 4051 254 168 192 0.807
16 Brown, Adin 6971 519 357 287 0.799
17 Feuer, Ian 4032 279 183 198 0.799
18 Reis, Matt 3239 179 125 88 0.798
19 Simpson, Mark 4531 265 181 139 0.792
20 Presthus, Tom 8766 614 425 275 0.787
21 Cassar, Jeff 4161 293 191 180 0.784
22 Causey, Jeff 6885 487 307 345 0.784
23 Lagerwey, Garth 4242 316 212 149 0.776
24 Dougherty, Mark 10766 759 505 354 0.772
25 Kramer, David 8996 627 406 340 0.771
26 Ammann, Mike 11155 743 478 393 0.767
27 Oshoniyi, Bo 3294 259 169 125 0.766
28 Salzwedel, Dave 2847 198 125 105 0.759

[tachyon1]

Hi Guys,
some good stuff especially from NS,

try this alternative method to evaluate save percentages.I've used it to before to interpret win/loss records.

First take an historical sample of representative goalies for the league that you're interested in.

Record their numbers of saves and non saves,so that you can get an idea of the distribution of save percentages that you are likely to see in the league.For example you might find that around 20% of the keepers in the league save 80% of the shots that they face.

A fairly typical distribution could be;

4% of keepers save 50% of their shots,
16% of keepers save 60% of their shots,
56% of keepers save 70% of their shots,
23% of keepers save 80% of their shots,
1% of keepers save 90% of their shots.

Once you have a probability distribution for the league take an actual set of saves/non saves for a particular keeper.

Lets take Tim Howard.

From NS's data he saved 74 shots and let in 18(an 80.4% success rate).

The next step is to calculate via the binomial,given that Howard has 74 successes & 18 failures what the probability is that he is really one of the 4% of keepers who save 50% of their shots.

Then you calculate the probability that,given his record,he is really a 60% keeper...and so on for each probability band upto the 1% of keepers who really do save 90% of their shots.

If my maths is upto scratch it's around a 2% chance that TH is really a 50% shotstopper,
a 12% chance that he's a 60% stopper,
a 58% chance that he's a 70% stopper,
a 27% chance that he's a 80% stopper,
& a 1% chance that he's a 90% stopper.

There's no real set in stone way to go next.If you multiply the probability that the keeper is a 50% keeper by the 50%,do this for each band & total them you get TH,on the basis of his 74/18 record & in the context of the league he is playing,coming out as a 78.5% shotstopper compared to the 80.4% that his 74/18 comes out at.

The method to some degree allows for good/bad fortune and the number of shots a keeper has faced.

Stick a record of 4 shots saved from 5 (an 80% percentage) into this method and you get a likely longterm average for the keeper of only 71%.He hasn't faced enough shots to prove that the 80% is reliable.

You could run a similar trial for catch/punch percentage.

Sorry to be dumb,but what does GAA stand for?

T.