Tuesday, February 3, 2009

Alone in the Universe, Apparently

From Yahoo Sports:

"5. New Jersey Devils (32-15-3, Previous: 7) – The Devils were hoping to play well enough to maintain a playoff spot when top goalie Martin Brodeur went down Nov. 1. But no one in the universe imagined they’d go 26-13-1 without the future Hall of Famer."

I didn't go on record with a prediction when Brodeur went down, and I wouldn't have gone as far as to forecast a .662 winning percentage in his absence. However, you can count me as someone who certainly imagined the possibility of the New Jersey Devils getting on just fine without their franchise goaltender.

Actually, from the results of a poll I put up on this blog at the time, I know there are a number of others who had a similar opinion. I think it is abundantly clear at this point that the New Jersey Devils are one of the top contenders in the Eastern Conference, and look to be a very strong team heading into this year's playoffs whether they have a healthy Brodeur or Scott Clemmensen in net.

Yahoo Sports also has another article up on the Devils' season so far. The key quote is probably from coach Brent Sutter: “I think the misconception was always outside, that this was a one-man team. We always felt we had a good hockey team, and we just carried on. Yes, Marty is an elite level player, an elite level goaltender, but he is just one part of it.”

39 comments:

Anonymous said...

Yeah, imagine that... a team that has frequently been in the top 10 or top 5 for the past many years... even with a goalie who in some seasons has had an average ability to make saves (aka an average save pct.)... is once again doing well but this time with a no-name goalie.

True, this replacement goalie is above-average in save pct... but insert a goalie with an average save pct, play him 65-70 games, & you'd still have a top 5-10 team & goalie with 35-40 wins.

It's just simple math.

Anonymous said...

... & play the average save-pct goalie 65-70 games for 15 yrs for the strong team... 15 x 35-40 wins... equals the most goalie wins of all time, therefore one of the all time great goalies!!!! :)

Anonymous said...

... & play the average save-pct goalie 65-70 games for 15 yrs for the strong team... 15 x 35-40 wins... equals the most goalie wins of all time, therefore one of the all time great goalies!!!! :)

Anonymous said...

You're not alone in the Universe, I thought I was until I found this website.

Clemmensen has put up numbers this year that won Brodeur a Vezina last year.

Its sad that this guy is gonna go into the Hall of Fame some day.

Anonymous said...

Even though it is only a very small number of people, I love how there are people who assume one man's success means another guy isn't good lol. Kind of silly. There is also two different groups of people, the box score checkers, i.e. the ones who couldn't tell you anything about a guy other than what they read in the stat column, and then the people who watch the games. Give Brodeur the goal support Clementine has been getting and he'd have multiple 50 win seasons in his career. And as good as Clementine's numbers are, .918 save percentage, 2.35 gaa, anyone who's watched even a few games can tell you his numbers are misleading as his puckhandling has directly lead to goals against, and his 5 hole is terrible. There are good goalies with good numbers, and bad goalies with good numbers. How many times in the past decade alone have we seen guys have a good year or two statistically, and then disappear?

Also, this silly argument that a slight difference in save percentage means anything is garbage because the save percentage is a ceiling statistic meaning that you hardly ever see an NHL goalie with a save % lower than .900 because those guys get sent down. You also almost never see a goalie with a save percentage higher than .930 because in order to do so a goalie would either have to face a very minimal number a scoring chances a game, and still an average number of shots (15-30) only letting in 1-2 MAX, or he would have to face a ton of shots, and still only let in a max of 3, which is unrealistic because nobody consistently sees 50+ shots a game. In you average game, there are between 27-32 shots per team. No goalie in the history of the game has come close to a season where he can only allow 1 goal when facing just that average number shots per game. In fact, very, very few have been able to consistently only allow 2 goals when facing those "average" number of shots. Thus assuming a goalie fluctuated between being amazing and only allowing 1-2, and then on occasion, still having a decent game but allowing 3 goals when facing an "average" number of shots, his save percentage would still fall not that far off from goalies that play mediocre. AND THIS IS NOT ACCOUNTING FOR NUMBER OF SCORING CHANCES FACED EITHER.

Again, I know how crazy you conspiracy theory Brodeur haters are about save percentage numbers, but save percentage is just as irrelevant a stat as gaa, or wins because it is dependent on many different variables.

Basically if you can not tell the difference between Scott Clemmenson and Martin Brodeur, you are either one of those losers who just looks at box scores not having a clue how those numbers became the way they did, or you have never seen either of them play. Either way, you have no business evaluating either. It is also equally as stupid to make claims about Brodeur if you are not equally scrupulous of other goalies, particularly the "real" lol Hall of Famers, you say he doesn't belong with. With the exception of AN article here or there about some other goalie being overrated, there is very little evidence here used to suggest that most goalies do not benefit from things(either the same, or different ones) the way Brodeur does. Basically all this site is a place for people who hate Brodeur to bash on him, hidden under the guise of "research".

Anonymous said...

yea this site is garbage. first it is easy to say brodeur isn't a great goalie and then pick apart his numbers and aspects of his game to make your case. especially since, outside hasek, you make no claim as to who actually is great, except for your bullshit article about older guys like plante, dryden, esposito etc which is crap because you use goalies who have no recorded save percentages, and since your knock on brodeur revolves around save percentage, their is no way to evaluate those older goalies by the same standards. so essentially you just admit the older goalies also benefitted from their teams, but that it didnt matter as much as brodeur benefitting from new jersey did, which again is crap because just about every goalie anyone will ever bring up played for a great team at some point in there careers, if not for most of it. even with patrick roy, you could rip apart his game and dissect it to the point you do brodeurs, but you dont, and just write ONE lousy article about how roy never won anything singlehandedly hahha. really? no shit. no one has ever won anything singlehandedly in a team sport. what total garbage this blog is.

Anonymous said...

ok. lets play devil's advocate. lets say Clemmenson went down for the season too... would this team make it to the playoffs on Weekes back? No. And thats a big, hell no!

What if Marty was hurt in 2000 or 2003... would the devils win the cup without him? No. No way in hell.

What if Marty played for Detroit all these years instead... how many cups would he have by now, 5, maybe 6? Would he be on pace for 600+ wins or 700, you can bet your rear!

The devils have always been a good team, but thats because they pay as a team... they have never had any superstar forwards to carry the team like most other good teams do. just solid teamwork and strategic defence, backstopped by the greatest goalie of all time, Marty Brodeur.

You can't deny that if Marty played for Detroit or Colorado throughout his career, his numbers would be much better, and he would certainly have won more cups than he already has with a team stacked full of superstars.

As usual, u wrote a biased garbage article full of your typical "i hate Marty" rhetoric.

Anonymous said...

G

Anonymous said...

A

Anonymous said...

R

Anonymous said...

B

Anonymous said...

A

Anonymous said...

G

Anonymous said...

E

Anonymous said...

From the Original Anonymous: Yeah, you're right, there's no difference between a svpct of .900 & .930... applied to 2000 shots over 4000 minutes in a season, it would make no difference in the results of games. None at all. Yeesh.

So I guess you're arguing that Brodeur would've had a higher svpct, but he just didn't feel like it. All he wants to do is "win". I guess that means he loses focus when his team is easily winning, & gives up soft goals yet still allowing his team to win? Hmmm.

I don't "hate" Brodeur... I just realize that a goalie who has often had simply an avg svpct (prior to the lockout) is really not that "great" (prior to the lockout), regardless of how many shutouts & how many wins he has (prior to the lockout).

However, for whatever reason, since the lockout he has definitely been well above avg. That's what the svpct numbers show. Too bad he couldn't do that consistently prior to the lockout.

Is Clemmenson better than Weekes? I should think so, judging by his svpct. Could/can Clemmenson keep up his good svpct for several seasons, even after shooters learn his tendencies? Maybe not. (Of course, he's also learning the tendencies of the shooters, too.) But even at an avg svpct, playing in front of NJ for the past 15 yrs would've given him 65-70 games per yr, 35-40+ wins per yr, etc etc. In other words, very Brodeur-like.

It would be nice to see some actual analysis from the pro-Brodeur crowd, other than "I saw him on t.v. & he looked good" & "he has lots of wins".

Would Brodeur have great numbers playing for DET & COL? Duh. So would the avg goalie. Last yr, Osgood & Hasek had roughly avg svpct stats, yet combined they had lots of wins, low GAA, etc. And won the Cup. Why? Because their team was really good. Doesn't mean the goalies were "great", or even "really good".

Anonymous said...

There is the error. I don't know why all you stat freaks are so obsessed with save percentage, when it is also team influenced. Craig Anderson currently has a save percentage of .930 over a good number of games, yet he is far from being the starter. Why? I don't know, I guess everyone else is just dumb. I mean, I am sure you 3 or 4 statistics freaks who's sole measurement for a goalies is 1 team influenced stat (save percentage)know more than the scouts, coaches and gms who run the teams.

Anonymous said...

"There is the error. I don't know why all you stat freaks are so obsessed with save percentage, when it is also team influenced."

Yes, svpct is team influenced -- check Alan Ryder's shot quality adjusted svpct -- but it's the best stat to indicate goalie quality. Better than GAA, wins, shutouts or any other stat currently developed. The brodeurblog author has a bunch of articles relating to this.

Svpct would be even better if broken down by even-strength, powerplay, & shorthanded situation. Luckily for the NJ goalies, NJ has usually been one of the least penalized teams.

There could be a host of reasons why C. Anderson is not the starter, perhaps none of them that good. Will he keep up a .930 over a full season? Over several seasons? Over a career? Perhaps not.

Of course, scouts, coaches & gm's always make the 100% correct choice about everything... yeah, right... if that were true, each team would be very similar in quality... there would be 30 teams all within a few pts of .500

I guess you are saying that we should just all give up on analysis, & conclude whichever goalie won the Cup is probably one of the very best if not the best that season.

Anonymous said...

So if thats is the case, and Brodeur isn't a "great" goalie, who is? I know you can name Hasek, but who else? Some guy who had 4 or 5 years with excellent save percentages and then disappeared? Please, and spare the Roy and Belfour nonsense too, because both had years were their save percentage wasn't that impressive. So, I guess you'll just mention the old timers right? The ones who don't have recorded save percentages? And if that is the case, why is it that every goalie anyone ever brings up played for some really good if not great teams? Just coincidence right? Cant fault those goalies, but you can fault Brodeur because you say he had "average" save percentage numbers, despite the fact that every year he played in the NHL he has had an ABOVE AVERAGE save percentage. You guys are a joke.

Anonymous said...

"...every year he played in the NHL he has had an ABOVE AVERAGE save percentage"

Actually, that is false. Compare his svpct each yr to the league avg svpct for that yr.

"...why is it that every goalie anyone ever brings up played for some really good if not great teams?"

I have no control over what "anyone" says.

Are good & great teams likely to have good & great goalies? Sure. That doesn't mean every good & great team had good & great goalies. There are lots of good & great teams that had middling goalies, if you just look at the goalie's ability to stop the puck, aka svpct.

"So, I guess you'll just mention the old timers right? The ones who don't have recorded save percentages?"

No, I won't mention the oldtimers who don't have recorded svpcts. Since I don't have access to those stats -- or the NHL has never compiled & released them -- I can't make a very good judgement on the quality of oldtimers, although I believe this blog author has made some attempts.

Why is it a joke to try to accurately assess goalie performance?

The Contrarian Goaltender said...

So if thats is the case, and Brodeur isn't a "great" goalie, who is?

Roy, Hasek, and Luongo (if he stays on the same career trajectory) are the only guys I would consider to be "great" in the last 20 years. Take a look at this post to see the gap between them and everyone else. I know that is based on save percentage, which you don't seem to like very much, but I have done other studies that agree with the results.

Nobody is saying that Scott Clemmensen is better than Brodeur, or that Brodeur is a terrible goalie. My view of Brodeur is that he is like Curtis Joseph with better teammates, or Ed Belfour with more games played, i.e. a good starting goalie for a long time but just not in that top tier of all-time greats.

Please, and spare the Roy and Belfour nonsense too, because both had years were their save percentage wasn't that impressive.

Patrick Roy had one year with an unimpressive save percentage, his rookie year where he posted an .875 when league average was .874. He was .005 better than league average in the shortened lockout season of 1994-95. Every single other season of his career he outperformed league average by .009 or more. For comparison's sake, Brodeur was .009 better than league average in only 7 out of his 15 seasons, less than half the time. Brodeur was also below average twice, in 1998-99 and 2001-02.

Yes Roy played for great defensive teams. But those are also great save percentages.

Still, I agree with you Roy is overrated (see "The Overrateds" on the top right of the page), but I just don't think he is overrated by nearly as much as Brodeur.

Anonymous said...

It's just incredibly silly, and just as much so flawed to place as much emphasis on save percentage as you number junkies seem to do. Save percentage is very much so team influenced, and is completely varying from team to team based not only on the system they play, but the players, and the opponents that team faces on a regular basis. Is it coincidence Luongo, Vokoun, Anderson all have great save percentage's with Florida? I dont know, but it definitely draws questions about team effect when you consider the Flyers have not had a goalie with a good save percentage in a while. But I doubt their style of play has anything to do with it.

Again, here is the difference between those of us who watch the games, and those that just like to check box scores for their fantasy teams. Last nights Rangers Thrashers game, and tonight's Toronto Buffalo game. rank the goalies in accordance with how well they played. Obviously Lehtonen played the best. the stats are correct. However you numbers junkies are going to probably say Lundqvist or Miller were 2nd and 3rd, with Pogge a distant 4th. Yet the truth is, and anyone who saw the games will agree, Pogge despite giving up more goals, had no defensive help, still stood on his head and easily outperformed both Miller, who faced a low number of shots, and had next to zero quality chances to contend with, and Lundqvist who let in one of the weakest goals possible and otherwise faced nothing but shots from sharp angles and shots from 65 feet out with no traffic. Yet by the numbers Lundqvist and Miller were not just good, but "great".

The Contrarian Goaltender said...

Again, here is the difference between those of us who watch the games, and those that just like to check box scores for their fantasy teams.

Oh yes, the old false dichotomy between the stat-loving geeks and the real hockey fans. All hockey fans watch the games. If they don't, they aren't hockey fans. In my experience, the primary motivating factor for statistical analysis among people who love the numbers is try to convince other hockey fans of the views and opinions they have developed from watching the games.

Where do you think my opinion on Brodeur came from? I developed it long before I had heard of shot-quality neutral save percentage, or I compared his save percentage to league average, or even looked at the numbers his backup goalies had put together. It was from watching him play.

Having said that, you are right when you say the numbers can be misleading, and that save percentage is team-dependent. I point that out all the time. The thesis of this website is that goalie play is dependent on the team, and that includes all traditional goalie stats.

Save percentage is the least flawed stat out there, but it is still flawed. That is why I try to use many different tools to analyze goalie play. I have closely followed several attempts to develop shot-quality adjusted numbers, and if we ever refine those shot quality adjustments we will be much better off in terms of analyzing which goalies are the best. Until then we have to make do with what we have. Yes there will be caveats to every analysis, but I'll still take it over the old-fashioned and outdated wins, shutouts and GAA model.

Anonymous said...

Granted it may be all we have at the moment, but to no degree would I say it is enough to draw a conclusion that anyone, let alone a goalie who has excelled largely as an outlier in ALL the traditional categories by which goalies are measured. Just because it may be the least flawed, doesnt mean it should be the ONLY measurement used, as is often the case here.

Especially when for many if not all of the years you contend Brodeur was not superb with save percentage, the division he played in was one of the highest scoring in hockey. I admit I have no clue as to the results of any work done on this, and I dont think anyone has done any work on this, however I would assume during the first half of Brodeur's career, facing the Flyers with the LOD and their powerhouse offense, the Capitals streaky offense, the Rangers and their hired guns whether it be Gretzky, Messier, Leetch, Bure, Graves, etc, or even the Islanders who could put up goals quick, facing those teams more times a year than any non divisional goalie, would have a negative impact on anyones save percentage. If you are willing to piece together the numbers of goalies in New jersey's division and their save percentage numbers against each division, and then compared them to how goalies in other divisions did when they faced those teams, then perhaps we will get somewhere. At the very least head to head division numbers may prove that Hasek benefitted tremendously from facing garbage Ottawa, Montreal,Carolina, and Boston teams throughout the late 90's. Either way, it is silly to assume that save percentage is any more meaningful than any other stat unless there is a consistent method used, i.e. games against specific teams, yet even then it is subject to error because the goalie still have different players on their respective teams influencing the outcome.

Anonymous said...

I obviously do not have the programs necessary to strip data like some of the stat nerds do, however after simply looking at a small piece of the head to head common opponent equation, as I guessed, there seems to be a correlation.

All results come strictly from the playoffs between 95-00. First common opponent was Philadelphia. Simply put, Hasek was mediocre against Philly, while Brodeur seemed to have his usual solid numbers against them. Even when Buffalo finally knocked off Philly, Haseks numbers were not spectacular.

Next common opponent is Ottawa. Hasek performed at a very high level here, and although New Jersey actually lost to Ottawa, Brodeur played relatively strong.

Montreal was also a team both faced, and again the difference wasn't much however the advantage goes to Brodeur.

Boston as a common opponent again shows Hasek doing well, but Brodeur doing much better.

The final common opponent was Toronto, upon who Brodeur dominated, while Hasek was below average.

So if anything what does this relatively small sample show? At least a hint of evidence that suggests that Brodeur performed at just as high a level as Hasek against teams from the division Hasek faced more often than any other team, at least in the playoffs. I would not say Philly alone is enough to draw a conclusion about the division Brodeur faced most frequently, however against Philly alone, Hasek didnt look so good, which at least suggests he may not have faired so well against other common opponents, but rather just beat up on the same teams consistently and more often than goalies who played in other divisons.

The Contrarian Goaltender said...

Either way, it is silly to assume that save percentage is any more meaningful than any other stat unless there is a consistent method used, i.e. games against specific teams

It is silly to assume any other goalie stat is more meaningful than save percentage because of simple logic. Save percentage is flawed because it does not incorporate shot quality against. GAA also does not incorporate shot quality against, and GAA also includes the number of shots against, which is mostly determined by the rest of the team. Wins are flawed because of shot quality, number of shots against, number of goals for, etc. The point is that save percentage has one flaw, and every other goalie stat still has that same flaw plus others. Therefore save percentage is the best stat we have.

If you want to factor in strength of opposition, that is again something that affects all goalie stats, not just save percentage.

Yahoo Sports has some nice split stat breakdowns which made looking up the numbers pretty easy, so I checked the numbers for both Brodeur and Hasek from 1991-92 to 2000-01. Brodeur wasn't affected much by his division (.911 vs. division, .911 vs. rest of East, .916 vs. West), but you might be onto something with Hasek (.934 vs. division, .925 vs. rest of East, .925 vs. West). It looks like Hasek did benefit from playing in that division, and the effect is primarily from his results against Ottawa (.950 sv%).

This is interesting, I'm going to look into it a bit deeper and do a post on it. Thanks for the comment.

Anonymous said...

Another example of the complete inaccuracy of save percentage stats.

Jonathan Quick vs Washington

41 saves 4 goals against

save percentage= .911

stat boy conclusion? eh about average performance, his save percentage wasnt anything special

The Truth= watch the game and try telling me he was anything short of brilliant

Anonymous said...

Anonymous - If you pick out individual, one-game performances, of course you are going to get some stats that might not truly encapsulate what really occurred.

(If some 4th line checker happens to score 2 goals one night, is he awesome & probably as good as someone who scores 60 goals in a season?)

However, over a full season, where a goalie might play 3000-4000 minutes, & face 1500-2000 shots... law of large numbers etc... (no use explaining math to you, I suppose).

By the way, .911 is above the league avg, I believe. Add in the shot quality -- I'll take your expert opinion that he was "brilliant" -- & you probably have a quite above avg performance.

Go do some research on shot quality & save pct. before your next whine-fest, ok?

Anonymous said...

Or, how about you do some research before you conclude that what ONE admittedly flawed statistic says, allows you to conclude which goalies are overrated, and which ones aren't.

And by the way, is it not you same stat junkies who knock Brodeur for having save percentages around .911? Or is it only "average" and "fraudulent" when he posts those numbers?

And by the way, the only thing associated with this shot quality and save percentage is that as shot quality goes up, save percentage goes down. Thus further disproving your whole save percentage argument because a goalie who faces nothing but shots from the blue line all game will have an inflated save percentage while another goalie can face 15 shots, 4 of which being 2 on 1's and allow 2 goals, and all of a sudden he is a fraud while the guy facing blue line wrist shots is an all star. Doesn't really make sense does it.

Your law of averages again is irrelevant if the variable that changes for each goalie, which is team effect, stays the same throughout all 3500 minutes.

If this was really about finding a true method for evaluation, rather than just bashing a guy for doing his job at a very high level, then you wouldn't place so much importance on this one flawed stat. Obviously it can be used effectively to gain insight, however using anything that is flawed, as a means to forming a concrete conclusion, (such as Brodeur is overrated) makes no logical sense whatsoever. This is no different than trying to solve a math related problem, using an equation with the wrong variables, and then acting like because the answer is close enough to what you want it to be, that it is acceptable. This really doesn't fly anywhere, and is hardly credible. Oh yea, and once again subjective. It appears as though only because the "save percentage only" measuring stick supports what you want it to, that there hasnt been any effort made to find additional data.

Anonymous said...

Ok, I finally signed up for a Google Account.

“And by the way, is it not you same stat junkies who knock Brodeur for having save percentages around .911? Or is it only "average" and "fraudulent" when he posts those numbers?”

I’m not sure what the league avg savepct is this year, but I’m guessing that .911 is a bit above it. It’s not bad compared to the league avg, that’s for sure. Aside from shot quality aspects, if Brodeur (or any goalie) has had a svpct above the league avg, chances are he is above avg. And vice versa.

While savepct is flawed to some extent, it is the best stat there is for evaluating goalies. I won’t bother repeating the many analyses that are on this webpage – feel free to spend a week or 2 or 3 examining them before you make your accusations.

Here is the “brilliant” (your word) play by J. Quick of LA vs. Washington:

http://tsn.ca/nhl/scores/boxscore/?id=9941

GA – 4
SA – 45
Svs – 41
Svpct - .911
Mins – 59.817
GAA – 4.01
W/L/T – W

Let’s see… he won the game, so that’s good I guess… but, sometimes goalies win games & they didn’t play all that great (they might’ve been worse than the losing goalie & only won due to their teammates &/or the poor play of the opposition)…. J. Quick had a GAA of 4.01, which is pretty bad when the league avg is (I believe) less than 3.00, so obviously GAA can be very inaccurate in terms of assessing goalie quality… BUT, he did have a svpct of .911, which is not bad… actually, probably above avg this yr.

As for those who apparently watched the game live, & weren’t “stat nerds”/slaves to the numbers, they chose the 3 Stars as such:

1: Anze Kopitar, 2: Alexander Frolov, 3: Dustin Brown.

Hmm, not one of the 3 stars, despite being “brilliant”. So, the only indication (besides your expert view) of his play being above-avg is…. YES…. SAVEPCT.

Yes, adjusting for shot quality would be a very important next step to apply to savepct. Yes, Ovechkin had 9 SOG, & he is an exciting player & high scorer, although he actually only has a shooting pct of 11.11, which is nothing special.

Are you able to watch every minute of the hundreds & hundreds of games played this year, by all teams, & then come to accurate conclusions about each & every player? Or would you have to rely on some stats of some sort?

Even if you were able to watch every minute of each player on every team, what if someone else disagrees with your very subjective (non-stats based) view? (A person could watch a couple of Leaf games & see Jason Blake skating like mad all over the place, taking a ton of shots, & conclude this guy is a dynamite 40-plus goal scorer….)

“And by the way, the only thing associated with this shot quality and save percentage is that as shot quality goes up, save percentage goes down.”

Yes, it would seem that for any goalie, as the shots faced become of higher quality, that the save pct would tend to go down.

“Thus further disproving your whole save percentage argument because a goalie who faces nothing but shots from the blue line all game will have an inflated save percentage while another goalie can face 15 shots, 4 of which being 2 on 1's and allow 2 goals, and all of a sudden he is a fraud while the guy facing blue line wrist shots is an all star. Doesn't really make sense does it."

Well, that’s actually a point against Brodeur, as he has typically throughout his career faced shots that are easier to save.

Not sure if you understand what we mean about adjusting svpct for shot quality: A goalie that saves 90% of shots that are, on avg, of lower shot quality, will have an adjusted save pct that is lower than 90%. A goalie that saves 90% of shots that are, on avg, of higher shot quality, will have an adjusted save pct that is higher than 90%. Alan Ryder of hockeyanalytics.com has a big study about this.

“Your law of averages again is irrelevant if the variable that changes for each goalie, which is team effect, stays the same throughout all 3500 minutes.”

The law of averages is not irrelevant, because not every single game is the same for a certain goalie (or any player); you can’t point to one game, & then conclude that his whole season was very similar if not identical. You don’t simply look at one game & then extrapolate to a whole season. The team effect for each particular goalie does not stay the same through the whole season. If it did, we could predict very accurately which team would win each game.

“Obviously it [Save Pct.] can be used effectively to gain insight, however using anything that is flawed, as a means to forming a concrete conclusion, (such as Brodeur is overrated) makes no logical sense whatsoever.”

But how flawed do you think savepct is? 10%? 5%? 50%? What if shot quality is combined with it? Less flawed? Is it the best stat? Is there a better stat? Why would it make no logical sense whatsoever to apply a stat that is the best stat there is, & is likely to correlate highly – over hundreds & hundreds of minutes played & hundreds of shots faced over a season – with goalie skill? What’s the alternative? As we've seen with J.Quick, a goalie's GAA & whether he won or lost the game can be very misleading... are you going to add GAA & Wins into the mix? Just watch a few games & decide who is good or bad or great or lousy? Or just listen to what the sports commentators say & take their word for it? (Most sports commentators couldn’t pass a high school math class & are subject to groupthink.)

“It appears as though only because the "save percentage only" measuring stick supports what you want it to, that there hasnt been any effort made to find additional data.”

Actually, additional data such as shot quality has been investigated by some people. This is an important facet to evaluating goalies by combining it with the best stat there is – save pct.

The facts are, that for much of the pre-lockout era, Brodeur was not that “great”. He played a lot of games for a team that tended to allow few shots again, & of lower quality, which resulted in a low GAA (relative to league avg) & not bad but unspectacular savepct (relative to league avg), & a lot of wins. If his team had allowed more shots &/or higher quality shots, with his save pct. he would’ve likely lost a lot more games… his whole career could’ve been altered heavily… perhaps other quality goalies would’ve been brought in by NJ, Brodeur gets traded, etc.

Since the lockout, Brodeur has been much better. I have no idea why, especially at this stage of his career & age.

[Why do so many critics of this site became so outraged, & think that somehow people who are merely applying analysis to hockey are “jealous” of Brodeur, “hate” him etc? Very weird, unless these critics are die-hard Brodeur/NJ fans &/or are teens/pre-teens. ]

The Contrarian Goaltender said...

It appears as though only because the "save percentage only" measuring stick supports what you want it to, that there hasnt been any effort made to find additional data.

I'm not sure if this was directed at me or the other Anonymous, but I must dispute the claim that there hasn't been any effort made to find additional data.

In addition to all the various adjustments I have made for save percentages (shot quality adjusted, era adjusted, opponent adjusted), I have done a lot of other different things, like comparing to backup goalies, looking at shot prevention and missed shots against, adjusting wins, adjusting shutouts, looking at stolen games, trying to assess team factors, looking at special teams performance, measuring giveaways, looking at performance vs. playoff teams, looking at Hart Trophy and award voting, and several in-depth analyses of situational save percentages.

And you know what I have found doing all these things? That save percentage is a very good measuring stick.

Let's take, for example, shutouts adjusted for the number of shots per game, like I did in this post. Run the numbers, and Hasek and Luongo show up at the top, while Brodeur is above average but not outstanding, finishing behind Roy and Belfour. That's similar to the results you would get if you ran the save percentages.

GAA compared to backups, same thing - Hasek is great, Brodeur is midpack. Same thing for wins compared to backup goalies, and a bunch of other studies.

Almost every study I have ever done comes up with Hasek at or near the very top, and with Brodeur above average but well back of the leaders. The only regular season stat that Hasek beats Brodeur in is save percentage. So what should that tell us about which stat is the most important number for goalies?

There is enough agreement between methods that I would say the burden of proof is definitely on those people who claim save percentage is useless. And by "proof", I certainly don't mean single games by journeyman Los Angeles Kings goalies.

I also don't know why you feel compelled to explain why 41/45 is a great performance, while something like Miller against Toronto (16 save SO) was not. I would call a 16 save shutout the "Brodeur special", and last I checked I wasn't exactly an advocate for Mr. Brodeur.

A shutout on anything less than 20 shots against, and you can pretty much guarantee that the goalie's team dominated the opposition, given what we know about playing to the score effects and the way the shot rate usually jumps in the third period for the trailing team. So no, I'm not likely to make that error in the case of an individual game.

Still, I don't like using save percentages in single games or other small samples. I prefer to look at the number over the course of the season, and then hopefully develop some context for the number with additional data on shot quality or various other team factors. Is there a potential for error? Sure. Is it large? No. Will that error persist over the course of an entire career? I strongly doubt it.

Anonymous said...

So again, in regards to anonymous, the "additional" stats used to back up save percentage is quality save percentage? Is that a joke? Most importantly, that stat is labeled as highly flawed, by THE GUY WHO CREATED IT. And secondarily, it is again largely based off of save percentage. So basically you are supporting a Pinto, with a hybrid engine, while acting like because you toyed with it a bit, that it still wont blow up the second something hits it.

CG.. you have made some heavy efforts to use other stats, however, it seems when it comes down to labeling Brodeur, the only one you give any weight to is save percentage, or some derivative of it. Why, I am assuming because he excels at just about every other stat, which would make your case a lot harder. Additionally, there isn't much that separates Roy from Brodeur, the only difference is your lack of enthusiasm for meticulously picking apart Roy's game. Roy if anything is regarded as the "Greatest of all time" by the masses. Yet you say he is overrated? Wouldn't that then mean he is more overrated than Brodeur? These are the things that don't make any sense. People speculate the Brodeur will be the greatest of all time, however I do not think I've ever heard a case saying he currently is. More often than not it is people wondering were he will finish at. With Roy, calling him one of the greatest ever, yet saying he is overrated would be taking yourself way too seriously, as to assume you have some inside track on Roy that separates him from the common perception of being the greatest, and only puts him and a distinct 2nd or 3rd best? that is silly.

Anonymous said...

"...the "additional" stats used to back up save percentage is quality save percentage? Is that a joke?"

I'm sure you haven't read all the studies relating to various goalie evaluation methods by now, so how else am I suppose to convince you? Go read them. I even used your J.Quick example to show that even based on one single game, the game that you chose, that save pct is the best indicator of goalie skill.

"Most importantly, that stat is labeled as highly flawed, by THE GUY WHO CREATED IT."

That's news to me... please give us the link.

"So basically you are supporting a Pinto, with a hybrid engine, while acting like because you toyed with it a bit, that it still wont blow up the second something hits it."

Ummm... well, if the analogy is to get somewhere, then I guess save pct is the Pinto & your other measures (like what? GAA? Wins? General sportscaster/public opinion?) are the... hubcaps? The seatbelts? The windshield wipers?

Please do some reading -- not for a couple of hours, but an adequate time -- & provide some good points.

Anonymous said...

I should add that if J.Quick was actually "brilliant" in that game, & that his svpct tended to reflect this (while his GAA, on the face of it, certainly did not), that this is just one game.... the savepct shows us that in that 1 game, he likely played well. As CG points out, this is just one game, & over a season & career his skill may be not so hot.

The Contrarian Goaltender said...

Why, I am assuming because he excels at just about every other stat, which would make your case a lot harder.

Brodeur excels at every other unadjusted, non-rate stat, that is correct. But remove some of the team variables, and his advantage of playing more games than everyone else, and you'll see a whole lot of air let out of the numbers.

Did you read the post I linked? Look at the stats, and you would think Brodeur must be great at shutting out the other team, since he is soon to become the all-time career leader. However, shutouts are easier the fewer shots you face. Take into account the number of shots, and suddenly Brodeur is far from the best.

Do the same thing with wins vs. backup goalies, or GAA vs. backup goalies, or compare how the backup goalies performed on other teams vs. how they did on Brodeur's team and adjust based on that, there are many ways to do it and they all tell you the same thing: Brodeur's performance is good but not elite.

Let me post again these statistics for Brodeur's backup goalies just to show what I'm talking about (and just for you, I'll completely avoid using save percentage here):

Backups, NJ: .522 win%, 2.56 GAA
Backups, other: .419 win%, 3.08 GAA
Brodeur: .630 win%, 2.20 GAA

Based on the way the backups played in New Jersey compared to on other teams, we would expect an average goalie in New Jersey to have a .620 win % and a 2.40 GAA. Compare that to Brodeur's numbers, and you can see why I call him a good, but not great, goalie.

For comparison's sake here is Luongo compared to an average goalie using the same method:

Luongo: .477 win %, 2.60
Avg goalie: .359 win %, 3.30

This isn't some big save percentage conspiracy, I could come to the exact conclusions I have without ever once invoking save percentage. I don't use save percentage because it gives me the conclusions I want, I use it because it is the best goalie stat. Brodeur has played on some great teams, and to not take that into account would be to give him credit that should be going to his teammates.

By the way, shot quality is not "highly flawed", it is merely affected by scorer bias (it varies somewhat from rink to rink). Since we are aware of it we can either adjust for it or use road shot quality only to minimize the effect. Simple fix.

The Contrarian Goaltender said...

With Roy, calling him one of the greatest ever, yet saying he is overrated would be taking yourself way too seriously, as to assume you have some inside track on Roy that separates him from the common perception of being the greatest, and only puts him and a distinct 2nd or 3rd best? that is silly.

Well, call me silly then, because Roy isn't the best goalie ever. I would rank Roy 4th all time, which, as you say, is about 3 spots lower than most people would.

I would probably rank Brodeur somewhere around the 9th or 10th best ever. Most people would probably put him in the top 2 or 3. That's a bigger gap than with Roy, so that is why I say Brodeur is more overrated.

You need to look beyond career wins and Stanley Cup victories if you think there is little that separates Roy from Brodeur. I already demonstrated how Roy's save percentage record is far better. If you adjust Roy's career wins based on era, Brodeur is still a couple of seasons away from passing him. And of course Roy's career playoff stats are way out of reach for Brodeur.

Anonymous said...

Loungo is over-rated and complete garbage. The guy has done nothing his entire career, and I sincerely doubt he will ever even win a cup.

Brodeur leads nearly every conceivable stat and will shatter most records by the time he hangs his skates. He also has won olympic gold, and several international tourneys. How is he anything less that the greatest of ALL TIME?

If anything Brodeur is underrated and deserves far more credit than he gets. No one is so consistent. Trade him to any, repeat, any team, and he would still win, flourish, and grab another cup or two.

The devils have never had any superstar players and usually were filled with a roster of undrafted college players and minor leaguers nobody else wanted. They won on Brodeur's back. The team sucked before he arrived, and its nothing short of a miracle that they are even in the playoff race this year.

I just have one thing to say, If Brodeur played for Roy's teams, he would far outplay and post better numbers than Roy & would win more cups.

Brodeur is the greatest, saying anything short of that is just a display of pure ignorance, or jealous hatred.

Anonymous said...

Colorado has an average goalie playing for them now, looks like they'll miss the playoffs. Wouldn't be the case if they had Brodeur.

The Contrarian Goaltender said...

Brodeur is the greatest, saying anything short of that is just a display of pure ignorance, or jealous hatred.

Sorry Anonymous, but the only one here showing pure ignorance is you. There are a couple of lines that I have come to view a sign that someone simply doesn't understand how to evaluate goaltending play independent of the team context. Your post pretty much includes all of them.

(Luongo) has done nothing his entire career

Brodeur leads nearly every conceivable stat and will shatter most records by the time he hangs his skates

No one is so consistent (as Brodeur)

The devils have never had any superstar players

Brodeur is the greatest


All of those statements are demonstrably false, and have been shown to be false by this blog and others.

Anonymous said...

Anonymous (the dumb one) -- seriously dude... how old are you? 11? 14?

Personally I have nothing against Brodeur... in fact since the lockout he's been one of the very best goalies. But, most of his career stats have been accumulated during the pre-lockout period, when he was often nothing spectacular.

If Brodeur was an absolutely amazing goalie -- clearly the best of all time, e.g. sv% of .950 or something amazing like that -- then I'd be glad, because that would heavily weigh in favour of Canada winning Olympic Gold etc.

It's you people who make absurd claims like "Luongo is garbage" -- when all the facts say otherwise -- that are the ones that are being incredibly biased to the point of delusion.