Monday, February 9, 2009

New Jersey Cowbell

A clever recurring reference among some Oiler and Flames blogs is to the "cowbell" of a particular fanbase or coach. You can read the history of this expression at the blog that originated it, Five Hole Fanatics, but in essence it refers to a specific skill or attribute that gets a disproportionate weighting in player evaluation (the name came from the famous SNL skit where Christopher Walken was obsessively focused on the sound of the cowbell).

The absence of Martin Brodeur is making it pretty obvious that New Jersey fans have a couple of cowbells in terms of evaluating goalie play. They are rebound control and puckhandling. Take a look at this HFBoards thread where Devils fan after Devils fan piles on Scott Clemmensen for his poor rebound control and his lack of Brodeur-esque puckhandling skills.

Apparently whether or not Clemmensen actually stops the puck is pretty unimportant, it is the style with which he does it and the way he handles it that is critical. I don't mean to downplay rebound control too much, since obviously poor rebounds create goals against, but the majority of rebounds are cleared away by the defence, and the majority of rebounds the other team does actually get to are stopped again by the goalie.

If Devils fans are correct in their assessment, then that strongly suggests Clemmensen has been better at making the first save than Brodeur. I don't have the exact comparison, but let's say for example that Brodeur gives up one rebound shot per game while Clemmensen gives up three, and that goals are scored on 25% of rebounds. That allows us to estimate the first shot save percentage for each of them:

Clemmensen: 25.6 1st SA/60, 3 reb/60, 2.36 GAA, .937 est. 1st shot sv%
Brodeur: 24.7 1st SA/60, 1 reb/60, 2.16 GAA, .923 est. 1st shot sv%

I think the other teams this season have been throwing a lot more perimeter shots on net against Clemmensen than they have against Brodeur, which might explain both the shot differential and why we could expect Clemmensen to have a strong first-shot save percentage.

This leads us to a hypothesis of goaltending play: Given how few rebound shots are actually taken in a typical hockey game (about 3 per game for both teams combined, according to one study), the effect of a goalie who is great at controlling rebounds might not just be allowing fewer second chance shots, but also facing fewer total shots by deterring the opposition from shooting in the first place. One of the main reasons to shoot from a bad angle or a long distance is to try to create a rebound, and if the shooters don't think they will get one then they are more likely to pass instead. The game score evidence that I have been working with lately shows that shots are to some degree discretionary (teams shoot substantially more often when trailing, for example), so I am starting to think that perhaps the main cause of the observed shot differentials between goalies may not be the direct effect of them giving up a lot of extra rebound shots or not being able to clear the zone, but because for whatever reason the other team thinks it is good strategy to put more or fewer pucks on net. Even if this is true is still could be argued that excellence in those types of goalie skills do effectively "prevent" or "create" shots against, but because the effect is indirect it would not be as easy to isolate it as it would be to, for example, count rebound shots.

Speaking of counting rebound shots, I am interested in seeing The Forechecker's rebound numbers for this season to see if perception matches reality in New Jersey. If we could isolate Brodeur's rebounds per game vs. Clemmensen's, that should be able to give us a sense of the direct impact of a goalie in terms of rebound control given the discrepancy between the two of them in that particular skill. Whether great rebound control can impact goal prevention indirectly through shot prevention is a topic that requires more study.

143 comments:

The Forechecker said...

I'll see if I can update my rebound info later this week, stay tuned!

Kent W. said...

Heh. When I came up with "cowbell", I thought people might laugh at it at the time and then forget about it. Funny that it's stuck around.

BTW - this post was the original one, although both contain the explanatory blurb.

Anonymous said...

On last weeks programming on the NHL Network they did an in locker room showing of the Nashville pregame against New Jersey. To quote Barry Trotz,

"Clemmenson has been good for them, although they do a good job of protecting him. Put the puck on net with him, from anywhere on on the ice. He isn't Brodeur, you don't need a perfect shot to beat him, and if you get the puck in deep, keep it low. He gives up lots of rebounds."

Maybe Trotz just overrates Brodeur like everyone else, and obviously he was exaggerating when he suggested only a perfect shot will will beat Brodeur, but his insight seemed true to a T, in regards to Clemmenson.

Anonymous said...

This post was total garbage. "Whether great rebound control can impact goal prevention indirectly through shot prevention is a topic that requires more study. "

Are you serious? You play goal, right? How can you POSSIBLY type that statement then? Less rebounds equals less scoring chances. Period.

Anonymous said...

there are a lot of bogus and biased posts by the author of this "lets bash brodeur" blog.

Anonymous said...

GARBAGE! Laughable post as usual, keep the comedy coming, its great.

Marty is the greatest of all time, everyone knows.

The Contrarian Goaltender said...

Are you serious? You play goal, right? How can you POSSIBLY type that statement then? Less rebounds equals less scoring chances. Period.

Obviously fewer rebounds equals fewer scoring chances, that is the direct impact of rebound prevention. I was discussing the indirect impact in the sentence you quoted, as well as in much of the above post.

That is illustrated by the quote in one of the comments above, where Trotz calls for his team to put the puck on net from everywhere against Clemmensen, whereas he wouldn't have the same strategy against someone like Brodeur. That is the indirect impact of rebound control - deterring the other team from taking low percentage shots since they think is less likely that something good (a rebound, deflection, etc.) will happen. I said it needed more study because we don't know the size of the effect, although Brodeur's injury is giving us a pretty good test case for it.

The Contrarian Goaltender said...

there are a lot of bogus and biased posts by the author of this "lets bash brodeur" blog.

What's bogus? Seriously, I want to know. Obviously I am not a huge fan of Brodeur, so if I am going too far in my criticism I'd appreciate it if somebody points it out. Here is a summary of the points made in the above post, so please feel free to specifically address the ones that are inaccurate and/or biased:

1. New Jersey fans place a heavy weighting on rebound control and puckhandling when evaluating goalies.
2. According to most observers, Scott Clemmensen gives up more rebounds than Brodeur.
3. If this is true, then according to the numbers Clemmensen has a better first shot save percentage than Brodeur.
4. If Clemmensen is facing more shots per game than Brodeur and stopping a higher percentage of the first shots he faces, that implies he is facing a lot more perimeter shots than Brodeur has.
5. Teams are more likely to take shots on goalies that give up rebounds, but rebound shots are not very frequent and most of the time do not result in goals against.
6. This implies that goalies with great rebound control face fewer shots not only because they allow fewer rebounds, but because the other team takes fewer perimeter shots.

Anonymous said...

All those things listed in the way you put them are pretty accurate. I would not necessarily limit the people who emphasize rebound control and puckhandling to those from New Jersey however. I'm pretty sure the entire NHL hierarchy felt puckhandling had some sort of negative impact on offense, otherwise we wouldn't have seen the Brodeur rule implemented after the lockout.

However it would make sense to see Clemmenson stopping more pucks than Brodeur because teams know they are more likely to score on him with a bad shot, or as a result of a rebound from a bad shot, than they are on Brodeur. This is just another instance in which the stats only say so much, and in this case in particular, are misleading. It may also be insight into the enigma regarding goalies who set the league on fire for a few years and then fall off into mediocrity. Teams throw everything they can at a guy, and after a while they start to figure out what works and what does not.

Anonymous said...

1. Everyone puts weight into rebound control and puck handling skills when evaluating a goalie. They are both highly beneficial skills a goalie can have in addition to stopping a puck, as both can lead to, not only preventing further shots, but also to clearing the defensive zone or even scoring chances going the other way. Without one or both of those skills, the defense has to work all the harder. One doesn't judge a forward purely on his ability to score goals, but also takes into consideration their passing ability and defensive play. Why should a goalie be looked at any differently?

2. Correction: Clemmensen gives up more uncontrolled rebounds than Brodeur. The defense has had to adjust to protecting him and the net, or else he's taken to freezing the puck at every chance and relying on NJ's abilities in the face off circle to clear the zone.

3. According to what numbers?

4. A distinct possibility, and a direct function of the defensive adjustments. Recall the Devils defensive play the first 5 or 6 games after Brodeur was injured. They were awful and Weekes was hung out to dry. Then they adjusted.

5. Teams are more likely to take shots on goalies that give up goals. They're more likely to crash the net on goalies that give up rebounds.

6. This implies an assumption that hockey is a team vs. a goalie and not another team. Teams adjust to many factors including, but not limited to: the current score; the goalie in net; the importance of the game; whether or not someone is playing without a stick; the current line of players on the ice; etc.

You allow your bias to cloud your judgment.

The Contrarian Goaltender said...

I still don't get your bias argument here. This post is not about Brodeur, it is about the effect of rebound control on shot and goal prevention. And secondly, my main point is that a goalie with great rebound control probably contributes to shot prevention not only directly but not allowing rebounds, but indirectly by deterring shots against. Proving and quantifying this link would be a benefit to Martin Brodeur because we would be able to assign him a portion of the responsibility for New Jersey's excellent record of shot prevention.

If you think I'm wrong that's fine, but the accusation of bias in this instance is curious.

With respect to the individual points:

1. Yes, everybody puts a weighting on puckhandling and rebound control, but no other fanbase weights them as heavily as New Jersey's does. Did you read that message board thread? I have noticed that rebound control rarely comes up in discussions of other goalies on this blog, even goalies who have either good or weak rebound control.

Your point about other things mattering for forwards is well-taken, but when people talk about Ovechkin, what do they bring up first, his goalscoring or his defensive play? Same thing with goaltending - all skills are to some degree important, but because of the nature of the position some are much more important than others. Someone who puts the same weighting on rebound control as they do on stopping the first shot is making a poor assessment and needs to review the relevant data.

3. The reason is simple math. Clemmensen has a higher save percentage than Brodeur, so if he gives up more rebounds then logically it follows that he stops a higher percentage of the first shots he faces. The only way this wouldn't be true is if Clemmensen has a much better save percentage against rebounds. That is unlikely since he gives up more bad rebounds than Brodeur.

This implies an assumption that hockey is a team vs. a goalie and not another team.

6: No, it doesn't imply that. There are many variables in hockey, I agree, but it is a basic reality that opposing teams tailor their strategy to the opposing team and goaltender. See the Barry Trotz quote in a previous comment for supporting evidence of my theory.

Anonymous said...

The mention of bias refers to your site as a whole, not specifically to this particular argument, though its existence leads to this discussion, and in turn affects your points.

I'm not sure what "excellent record of shot prevention" you refer to. In 07-08 they ranked 8th in the league in shots against, 10th in 06-07, and 11th in 05-06. Currently, they rank 10th. Brodeur in net or not, they're pretty consistently not "excellent."

1. Firstly, a small sample of people posting to one message board in response to someone deliberately trying to discount their favorite goalie is not representative of an entire fan base. Secondly, I read through the thread and fail to see more than a few posts that even mention rebound control or puck handling, never mind an entire fan base that "weighs them as heavily as New Jersey's does." And lastly, it probably doesn't come up as much because it's not an issue. When you watch one of the best at both of those things for a time and then watch someone who isn't as good at it, it is noticeable. When you watch a tandem of goalies who are close in skill at either or both, it's a non-issue. There's no discussion. The mere fact that it comes up at all is a testament to Brodeur's skill at both.

Well, let's compare apples to apples here. Who talks about *any* forwards' defensive play outside of a discussion about Norris Trophy candidates? People will talk about a highlight reel goal over a highlight reel stick check (if there even is such a thing) because it's a sexier topic, not just because it's Ovechkin scoring it.

Someone who puts the same weighting on rebound control as they do on stopping the first shot is making a poor assessment and needs to review the relevant data.

Well, without stopping the first shot, there is no rebound to control, so it's a moot point. Talking about controlling rebounds is, obviously, with the assumption that the shot has been stopped.

3. Check your stats, Brodeur has a higher save percentage. Brodeur has a career save % of .914, Clemmensen a career .908. And removing his year in Toronto and only including his time with NJ gives him a career % of .888. In any event, save percentage is irrelevant in a discussion about rebound control.

6. As the numbers above just prior to point #1 show, the Devils are giving up about the same number of shots on goal this year with Clemmensen and Weekes in net as in the previous three with Brodeur and Weekes/Clemmensen. Your statement completely dismisses defensive play and credits shot totals entirely to a goalie's rebound control, which is absolute bunk.

Anonymous said...

Crud, of course, I meant the Selke Trophy, not the Norris.

The Contrarian Goaltender said...

Anonymous: I appreciate your feedback, but you seem to reading a lot into everything I write. It is not my goal to bash Brodeur with every single post, but that seems to always be your interpretation. As much as you may not believe it, my goal with this site is to study goalie play. I'm planning to do a lot more posts on this topic and related ones, because Brodeur's injury has given as a good opportunity to try to estimate the effects of the more intangible goaltending skills (puckhandling, rebound control, etc.).

I'm not sure what "excellent record of shot prevention" you refer to.

How long have you been a fan of the NHL? Do you not remember hockey before the lockout or are you just trying to score some cheap argument points? Brodeur is famous for playing behind a stingy defence in New Jersey. In his career, he has faced 25.4 shots per game. The league average in that time has been 28.4.

Regarding the issue of New Jersey fans and their "cowbells", I could have a wrong impression, but as you can probably understand I have become somewhat familiar with the New Jersey fan base through my operation of this site. Maybe I'm wrong, I don't know, but that was just an observation I threw out there but really has little to do with the point of the post. Note that I agree with their evaluation that Brodeur is good at those things and Clemmensen is not, I just believe those skills are generally overvalued by most people in hockey, a position I have consistently held since I started this blog. If I see good evidence that I'm wrong about that then I will change that position. That is the whole point of posts like this one, to collect and evaluate that evidence.

Well, without stopping the first shot, there is no rebound to control, so it's a moot point.

It's not a moot point, that is entirely the point that I am making. There are people who overrate rebound control because they fail to take into account the number of shots that go directly in. A rebound is not the desired outcome, but a direct goal against is much worse.

To offer an extreme example, I'd rather have a goalie that stops 95% of his first shots and gives up 5 rebound shots against per game than a goalie who stops 90% of his first shots and never gives up a single rebound.

In any event, save percentage is irrelevant in a discussion about rebound control.

No it isn't. In fact, save percentage is useful in illustrating just how important rebound control is. As the previous example I gave shows, if you are stopping a very high percentage of the first shots you can have terrible rebound control and still have pretty good results.

I was using Clemmensen's numbers from this season for my estimates, but if you want to use his career numbers from NJ that is fine as well. Turns out that if Clemmensen allows anything more than 0.5 rebound shots against per game than Brodeur, then he has a higher first-shot save percentage (assuming Clemmensen's save percentage on rebound shots is equal to or worse than Brodeur's, which is probably a very reasonable assumption).

What do you think, does Clemmensen allow one extra rebound shot against every two games compared to Brodeur? If he does, then I'm correct by saying Clemmensen is more likely to make the first save. If he doesn't, then rebound control apparently doesn't have much of an impact, given that Brodeur is one of the best and Clemmensen is below average.

Your statement completely dismisses defensive play and credits shot totals entirely to a goalie's rebound control, which is absolute bunk.

This is misconstruing what I wrote. I wrote that a goalie likely has some effect on shots against, but more study is required. I have spent a lot of time arguing that shots against totals are predominantly dependent on the rest of the team, and I still hold that position.

Anonymous said...

GARBAGE!

... go get a job or a hobby you moron!

Anonymous said...

Its all bullshit... Marty never really won 544 games, its all an illisuion! I have the adjusted save percentage and quality shots stats to prove it!

See in 7 wins you asshat loser. lol.

Anonymous said...

Anonymous: I appreciate your feedback, but you seem to reading a lot into everything I write. It is not my goal to bash Brodeur with every single post, but that seems to always be your interpretation.

You'll have to forgive my skepticism as I post a response to a blog entitled "New Jersey Cowbell" on a website named "Brodeur is a Fraud."

How long have you been a fan of the NHL?

Hmm, well, the first game I ever went to featured Chico Resch in net, so I guess it's been a few years.

Do you not remember hockey before the lockout or are you just trying to score some cheap argument points?

Not at all. The reference pages I used only went back three years.

Brodeur is famous for playing behind a stingy defence in New Jersey. In his career, he has faced 25.4 shots per game. The league average in that time has been 28.4.

I come up with 25.1, but I won't split hairs over a couple tenths of a percentile.

To the first point: And? You say it like he's to be condemned for having a good defense in front of him.

To the second: And? Three less shots on goal over a 16 year career than the league average. By definition, average means that's the middle number... that there were teams who gave up less and teams that gave up more.

Neither of your points explain why you maintain that the Devils have an "excellent record of shot prevention." In fact, they show that they only have a "slightly above average" record of shot prevention.

Regarding the issue of New Jersey fans and their "cowbells", I could have a wrong impression, but as you can probably understand I have become somewhat familiar with the New Jersey fan base through my operation of this site.

You've become familiar with a small portion of the fan base who only make their presence known to defend their favorite team and goalie.

If I based my opinion of, say, all Rangers' fans based solely on my experiences with a few of them at Devils games it would be unfair and illogical.

Maybe I'm wrong, I don't know, but that was just an observation I threw out there but really has little to do with the point of the post.

Then why throw it out there?

Note that I agree with their evaluation that Brodeur is good at those things and Clemmensen is not, I just believe those skills are generally overvalued by most people in hockey, a position I have consistently held since I started this blog. If I see good evidence that I'm wrong about that then I will change that position. That is the whole point of posts like this one, to collect and evaluate that evidence.

The problem is you're trying to quantify something that is, ultimately, unquantifiable. There is no stat that will undeniably prove it true, just as there is none that will irrefutably prove it false.

If you gave a new coach the ability to build a team from scratch; pulling from a pool of all the players that play today, do you think he'd build his team around a strong goalie or a strong goalie who had good rebound control and puck movement? Those skills are not overvalued... they're just immeasurable.

It's not a moot point, that is entirely the point that I am making. There are people who overrate rebound control because they fail to take into account the number of shots that go directly in. A rebound is not the desired outcome, but a direct goal against is much worse.

It is a moot point because if it goes directly in, there is no rebound. Why would shots that go in be taken into account in a discussion of rebound control?

To offer an extreme example, I'd rather have a goalie that stops 95% of his first shots and gives up 5 rebound shots against per game than a goalie who stops 90% of his first shots and never gives up a single rebound.

And I'd rather have a goalie that stops 90-95% of first shots and kicks the rebounds to the corner or to a defenseman, or waffleboards it into the mesh than a goalie with poor rebound control who gives up rebound shots at all.

No it isn't. In fact, save percentage is useful in illustrating just how important rebound control is. As the previous example I gave shows, if you are stopping a very high percentage of the first shots you can have terrible rebound control and still have pretty good results.

Save percentage takes into account all shots. If you've got a reliable source that only counts first shots, and quantifies a rebound shot better than "any shot taken within 5 seconds," I'd like to see it. Otherwise, save percentage is far too broad a category to use in a discussion of something as precise as first shots vs. secondary shots, etc.

I was using Clemmensen's numbers from this season for my estimates, but if you want to use his career numbers from NJ that is fine as well. Turns out that if Clemmensen allows anything more than 0.5 rebound shots against per game than Brodeur, then he has a higher first-shot save percentage (assuming Clemmensen's save percentage on rebound shots is equal to or worse than Brodeur's, which is probably a very reasonable assumption). What do you think, does Clemmensen allow one extra rebound shot against every two games compared to Brodeur? If he does, then I'm correct by saying Clemmensen is more likely to make the first save.

See above.

If he doesn't, then rebound control apparently doesn't have much of an impact, given that Brodeur is one of the best and Clemmensen is below average.

Right. So if you manage to scrape up some statistical evidence that Clemmensen makes a few tenths of a percentage point more saves on first shots, then it obviously means rebound control has little impact. Because there's really no difference in kicking a rebound to a defenseman or an opponent.

This is misconstruing what I wrote. I wrote that a goalie likely has some effect on shots against, but more study is required. I have spent a lot of time arguing that shots against totals are predominantly dependent on the rest of the team, and I still hold that position.

That is what you say now, and with that I agree. But earlier you stated:

"5. Teams are more likely to take shots on goalies that give up rebounds, but rebound shots are not very frequent and most of the time do not result in goals against.
6. This implies that goalies with great rebound control face fewer shots not only because they allow fewer rebounds, but because the other team takes fewer perimeter shots."

which implies that it is rebound control and not team play that affects shot totals, which is bunk.

-e

The Contrarian Goaltender said...

which implies that it is rebound control and not team play that affects shot totals, which is bunk.

It doesn't imply that at all. Of course team play affects shot totals, there are many variables involved in goaltending. That is obvious to anyone with any sense. Saying that A is a cause of B does not mean that A is the sole cause of B, it means that if you hold everying else constant a change in A will have an affect on B. In the real world, B is also likely to be affected by variables C, D, and E. When I write things like "this implies that goalies with great rebound control face fewer shots" I am speaking in relative terms, and more specifically dealing with the comparison between goalies on the same team like Brodeur and Clemmensen.

I'd like to be more accurate and say "goalies with great rebound control face 0.5 fewer shots per game than average goalies", but I haven't been able to estimate that effect yet.

I was actually hoping you would answer the question I posted in my last comment: Do you think Clemmensen gives up more or less than 0.5 extra rebound shots per game compared to Brodeur? How big do you think the effect is?

Three less shots on goal over a 16 year career than the league average. By definition, average means that's the middle number... that there were teams who gave up less and teams that gave up more.

There weren't any teams that gave up less shots than New Jersey. Since 1993-94, the Devils have allowed fewer shots against per game than any other team in the league. If being the best in the league over a 15 year span is not excellent, then what is exactly?

Also, in a recent study on playoff performance, I looked at how each goalie did based on era-adjusted shots against per game, and Brodeur ranked #1 of all the top goalies since 1968 in fewest shots against per playoff game, ranking ahead of even goalies who played on great teams like Dryden, Cheevers, Roy, etc.

You've become familiar with a small portion of the fan base who only make their presence known to defend their favorite team and goalie.

Not at all. I was referring not only to comments and emails directed to me, but to discussion of my posts and related subjects on team message boards and blogs. Devils fans were criticizing Roberto Luongo's rebound control long before I started writing. I have not polled every single New Jersey fan out there, so there could be some sample size bias, as I conceded. I included it in the post because it was my observation and I believe it to be correct.

The problem is you're trying to quantify something that is, ultimately, unquantifiable. There is no stat that will undeniably prove it true, just as there is none that will irrefutably prove it false.

It doesn't work to say that something is important, but at the same time it has absolutely no observable impact on any measure of goal prevention. If something doesn't affect goals for or against, then it has little value in terms of winning or losing hockey games. I agree that we can't measure the effect exactly to five decimal places, but that's not what I'm trying to do. I'm trying to estimate the effect to get a rule of thumb that we can use to assess the benefit of one goalie's rebound control compared to another.

There are lots of different ways we can analyze the data. We can compare the total shots faced and goals allowed by each of the goalies. We can use NHL play-by-play logs to count rebound shots and rebound goals allowed, as well as to look at the type and distance of each shot faced to estimate the shot quality faced by each goalie. If there is an effect, it will show up somewhere. With Brodeur and Clemmensen, we have two goalies playing on the same team, one who is great at the skill and one who is below average. It is in many ways an ideal test case.

Those skills are not overvalued... they're just immeasurable.

If they are immeasurable, how do you know they aren't overvalued? Even if teams can't quantify it, they still have to evaluate it somehow. If you are a GM looking to draft a goalie, and Goalie A is better at making the first stop but worse at rebound control than Goalie B, then you need to make an estimate of the effect of Goalie B's rebound control and whether that makes him better.

If everything else is equal, of course you would take the goalie with better peripheral skills. But the choice is always far more complicated in the real world. Do you take the goalie with better rebound control and puckhandling or the guy who stops the first shot more often? Do you take the checking centre who kills penalties well but is bad at faceoffs or a checking centre who is great at draws but weaker on the PK? Scouts face questions like that all the time, and whether they can quantify them or not they have to estimate the value of each skill and come up an overall grade for each player or goalie.

If you've got a reliable source that only counts first shots, and quantifies a rebound shot better than "any shot taken within 5 seconds," I'd like to see it.

The numbers I use for rebounds are all based on the work done by The Forechecker, whose definition of a rebound shot is "a shot taken within two seconds of another, from less than 30 feet from the endboards, and without any intervening event." Do you have a problem with that definition? If we can count rebound shots, then we also can count first shots, because of course First shots = Total shots - Rebounds shots.

Bruce said...

6. This implies that goalies with great rebound control face fewer shots not only because they allow fewer rebounds, but because the other team takes fewer perimeter shots.

I agree with this.

1. Everyone puts weight into rebound control and puck handling skills when evaluating a goalie. They are both highly beneficial skills a goalie can have in addition to stopping a puck, as both can lead to, not only preventing further shots, but also to clearing the defensive zone or even scoring chances going the other way. Without one or both of those skills, the defense has to work all the harder. One doesn't judge a forward purely on his ability to score goals, but also takes into consideration their passing ability and defensive play. Why should a goalie be looked at any differently?

I agree with this too.

my main point is that a goalie with great rebound control probably contributes to shot prevention not only directly but not allowing rebounds, but indirectly by deterring shots against. Proving and quantifying this link would be a benefit to Martin Brodeur because we would be able to assign him a portion of the responsibility for New Jersey's excellent record of shot prevention.

Maybe the light is finally coming on. A goalie who excels at puck-handling and rebound control just happens to play on a team that allows < 90% shots than the average team over a period of a decade and a half. Said goalie proceeds to win a Lot of games. Some people credit the goalie; others credit stingy team defence. Still others see the goalie in context as part of the stingy team defence.

I'm not sure what "excellent record of shot prevention" you refer to. In 07-08 they ranked 8th in the league in shots against, 10th in 06-07, and 11th in 05-06. Currently, they rank 10th. Brodeur in net or not, they're pretty consistently not "excellent."

Top third in the league every year is, if not excellent, very good. It is true that the post-lockout Devils are not the shot prevention machine that the pre-lockout Devils were, but whether that's due to the departures of Stevens, Niedermayer and Daneyko or the Brodeur Rule is open to interpretation, now isn't it? Room to advance both arguments.

Trouble is nothing happens in a vacuum. Other factors change besides whatever one you're trying to track. A test case of the Brodeur-less Devils is interesting, but the variables include:

1) Scott Clemmensen may be better than "replacement level" or he may be worse, we can't just assume he's average. To my eye he has been excellent.

2) The 2008-09 Devils may be better or they may be worse than the teams that preceded them. To my eye they have been excellent. They have the look of a serious contender, especially in the ideal case of having a healthy, fresh Brodeur in the spring. It may turn out that losing Broduer for awhile turns out to their benefit, as the team has risen to the challenge of finding something of a new identity. One measurement of this is the Devils Goals For/Goals Against compared to league average:

2006-07: 89% / 83%
2007-08: 90% / 86%
2008-09: 110% / 87%

That they still -- as always -- have an excellent defensive record speaks to a more balanced squad; well better than average on both sides of the puck. To me that's a team that's positioned to make real noise in the playoffs, unlike the popgun Devils of recent vintage.

which implies that it is rebound control and not team play that affects shot totals, which is bunk.

Is there not room in this discussion that BOTH have an effect? Rebound control, and puckhandling, actually contribute to team play, as well as to the strategies of the opposing team (as CG points out above, and I did in a quote of Ron Wilson's Brodeur-specific game plan early in the season). It's all part of the flow of play, and some goalies are more active and efficient than others. The data points that trigger the various statistics -- shots against, save percentage -- capture that flow imperfectly to say the least.

Anonymous said...

It doesn't imply that at all. Of course team play affects shot totals, there are many variables involved in goaltending. That is obvious to anyone with any sense. Saying that A is a cause of B does not mean that A is the sole cause of B, it means that if you hold everying else constant a change in A will have an affect on B. In the real world, B is also likely to be affected by variables C, D, and E. When I write things like "this implies that goalies with great rebound control face fewer shots" I am speaking in relative terms, and more specifically dealing with the comparison between goalies on the same team like Brodeur and Clemmensen.

So, relatively speaking, you're saying Brodeur faces less shots than Clemmensen?

I'd like to be more accurate and say "goalies with great rebound control face 0.5 fewer shots per game than average goalies", but I haven't been able to estimate that effect yet.

So, you figure out what you'd like to say first, and then search for evidence to estimate it?

I was actually hoping you would answer the question I posted in my last comment: Do you think Clemmensen gives up more or less than 0.5 extra rebound shots per game compared to Brodeur? How big do you think the effect is?

Based purely on observational and anecdotal evidence, I'd say he gives up more.

The effect is that the rest of the team has had to adjust to that difference in style and ability.

There weren't any teams that gave up less shots than New Jersey.

If 25.4 shots against per game is the league's best, and the league average in that time is 28.4, then the entire range of shots against per game, from top of the line team #1 to bottom of the barrel team #30, is 6. Hardly a stat to rest your argument on.

Since 1993-94, the Devils have allowed fewer shots against per game than any other team in the league. If being the best in the league over a 15 year span is not excellent, then what is exactly?

The above aside, I'm not sure where you're getting your stats from. A quick perusal of just last season shows that the Red Wings were the best with only giving up 1930 shots, for a per game average of 23.5, whereas NJ gave up 2257, for a per game average of 27.5. For contextual purposes, the worst team, Atlanta, gave up 2782, for a per game average of 33.9.

I'm assuming you're averaging the entire 15 year period, and I'd be curious to see your source, since I can't locate anything that goes that far back to confirm your numbers.

Also, in a recent study on playoff performance, I looked at how each goalie did based on era-adjusted shots against per game, and Brodeur ranked #1 of all the top goalies since 1968 in fewest shots against per playoff game, ranking ahead of even goalies who played on great teams like Dryden, Cheevers, Roy, etc.

A recent study? Era-adjusted shots? Mm-k. Even with the large assumption that this study is on the up-and-up, it doesn't back up your argument.

Not at all. I was referring not only to comments and emails directed to me, but to discussion of my posts and related subjects on team message boards and blogs. Devils fans were criticizing Roberto Luongo's rebound control long before I started writing. I have not polled every single New Jersey fan out there, so there could be some sample size bias, as I conceded. I included it in the post because it was my observation and I believe it to be correct.

It's a generalization, which, by your own admission, "...was just an observation I threw out there but really has little to do with the point of the post."

It doesn't work to say that something is important, but at the same time it has absolutely no observable impact on any measure of goal prevention.

Who said it has no observable impact? All I said was that it wasn't quantifiable.

If something doesn't affect goals for or against, then it has little value in terms of winning or losing hockey games.

Of course it has an effect. It affects how the entire team plays and how coaches strategize a game, but it isn't something like a blocked shot or a steal that can be statistically tracked.

I agree that we can't measure the effect exactly to five decimal places, but that's not what I'm trying to do. I'm trying to estimate the effect to get a rule of thumb that we can use to assess the benefit of one goalie's rebound control compared to another.

But you're trying to estimate the effect numerically.

There are lots of different ways we can analyze the data. We can compare the total shots faced and goals allowed by each of the goalies. We can use NHL play-by-play logs to count rebound shots and rebound goals allowed, as well as to look at the type and distance of each shot faced to estimate the shot quality faced by each goalie. If there is an effect, it will show up somewhere. With Brodeur and Clemmensen, we have two goalies playing on the same team, one who is great at the skill and one who is below average. It is in many ways an ideal test case.

Go ahead and sort those numbers then and please let us know what you come up with.

If they are immeasurable, how do you know they aren't overvalued? Even if teams can't quantify it, they still have to evaluate it somehow. If you are a GM looking to draft a goalie, and Goalie A is better at making the first stop but worse at rebound control than Goalie B, then you need to make an estimate of the effect of Goalie B's rebound control and whether that makes him better.

See above.

If everything else is equal, of course you would take the goalie with better peripheral skills. But the choice is always far more complicated in the real world.

I was responding to your hypothetical "extreme example," to illustrate there was more than the two choices you presented.

Do you take the goalie with better rebound control and puckhandling or the guy who stops the first shot more often? Do you take the checking centre who kills penalties well but is bad at faceoffs or a checking centre who is great at draws but weaker on the PK?

Depends on how I've built my team and what my needs are.

Scouts face questions like that all the time, and whether they can quantify them or not they have to estimate the value of each skill and come up an overall grade for each player or goalie.

Which is why they go to watch them play and don't rely purely on stats.

The numbers I use for rebounds are all based on the work done by The Forechecker, whose definition of a rebound shot is "a shot taken within two seconds of another, from less than 30 feet from the endboards, and without any intervening event." Do you have a problem with that definition?

Got a link?

If we can count rebound shots, then we also can count first shots, because of course First shots = Total shots - Rebounds shots.

Simplistically, yes; but that wouldn't take into account deflections, the effect of screens, point shots vs. wraparounds, power play shots, etc., so how would you go about comparing first shots? Assign each one a difficulty level and rate each shot and average them out over a game?

-e

Anonymous said...

damn, it looks like the brodeur bashers are in over their heads now that there are quite a few people who can intelligently make a case against the biased opinions and cherry picked statistics this blog is filled with.

just to add to the point, take an extreme. if goalie A is the best puckhandler in the league, and goalie B is so bad he doesnt even leave the net, goalie A will undoubtedly see fewer shots. going to another extreme, if goalie A is so quick he can effectively go out and clear EVERY dump in from the opposition, he will face few if any shots. meanwhile the fact that goalie B never leaves the nets means it is a foot race between defenseman and forwards which at the very least would create the opportunity for a shot, just about every time either as a result of the forechecker winning the race, or a the forechecker effectively putting pressure on the defenseman creating turnovers.

Anonymous said...

Anonymous - There have been previous discussions & studies on here regarding puckhandling goalies & the effect they may have on reducing shots against. Do a search.

Regardless of the degree to which shots against are reduced by straying/puckhandling goalies, shots against will still exist (won't be zero), & at that point the ability to stop the puck (aka save pct) determines goalie quality.

The Contrarian Goaltender said...

Anonymous: I'm going to post some links that you requested first before I respond to some of the other points.

I'm assuming you're averaging the entire 15 year period, and I'd be curious to see your source, since I can't locate anything that goes that far back to confirm your numbers.

I highly recommend Hockey Reference, probably the best site on the web for hockey stats.

Go ahead and sort those numbers then and please let us know what you come up with.

My first attempt is right here. More detailed analysis is on the way. Again, we have a situation where a goalie who is universally agreed upon to be an excellent puckhandler and great at rebound control is injured, and his backup has stepped in for most of the season. This is a great chance to try to estimate the effect of those skills, and being able to value them has implications for goaltending evaluation at all levels.

Rebound analysis can be found at The Forechecker

Simplistically, yes; but that wouldn't take into account deflections, the effect of screens, point shots vs. wraparounds, power play shots, etc., so how would you go about comparing first shots? Assign each one a difficulty level and rate each shot and average them out over a game?

Yes, that is the essence of shot quality measurements. See here for an introduction to the technique by Alan Ryder.

Anonymous said...

For ease of future reference and to reduce confusion, feel free to refer to me as E, rather than anonymous.

The first link doesn't work.

The second link refers to a source link that doesn't cover what the original article discusses, and does little but bury a reader in numbers that prove nothing... as admitted by the author.

The third link disclaims its information with: "About 10% of the data was missing. However, the data used would certainly qualify as a representative sample." and "There are data quality issues." and "One can only use the data available. Certain factors, which would seem to lead to shot quality variation, are not tracked (see below). As such, the shot quality model developed is cruder than I would prefer."

All of which immediately call into question any speculation stemming from its use.

-e

Anonymous said...

Looked up Hockey-Reference.com myself, but still cannot find a season comparison of SOG. Perhaps you could just link the page?

-e

The Contrarian Goaltender said...

There is no season comparison of SOG, you have to compile it yourself. I did it by using the Power Play function. Here is the link to New Jersey's combined goalie stats for the period in question: (http://www.hockey-reference.com/pp/psl_finder.cgi?request=1&sum=1&year_min=1994&year_max=2009&season_start=1&season_end=-1&age_min=0&age_max=99&franch_id=NJD&is_active=&is_hhof=&pos=G&handed=&c1stat=&c1comp=gt&c1val=&c2stat=&c2comp=gt&c2val=&c3stat=&c3comp=gt&c3val=&c4stat=&c4comp=gt&c4val=&order_by=games_goalie

The Contrarian Goaltender said...

The second link refers to a source link that doesn't cover what the original article discusses, and does little but bury a reader in numbers that prove nothing... as admitted by the author.

You were responding to my comment about the statistical tools we can use, so I posted a link that shows some of them applied to the New Jersey situation (shots against, goals against, shot quality against). I don't have data for rebound shots, as I admitted in the original post. If somebody posts that data, then we have that to include as well.

And no, it doesn't prove anything, gray areas are typical in goalie analysis. Everything is evidence of one thing or another, and the best thing to do is to accumulate all the evidence you can find and use that to inform your viewpoint.

All of which immediately call into question any speculation stemming from its use.

That was shot quality version 1. I linked it because it explains what shot quality is trying to do (which was very similar to what you suggested). Subsequent attempts have expanded the number of adjustments, have made sure to include the entire set of data, and have made adjustments for rink-to-rink scorer variations. Is the method foolproof? No, pretty much every stat in hockey is subject to some reporting error. But like anything else it is useful in providing evidence of the effectiveness of a goaltender or a team's defensive play.

The Contrarian Goaltender said...

So, relatively speaking, you're saying Brodeur faces less shots than Clemmensen?

Of course. That was clearly argued in the original post, and the numbers back it up.

So, you figure out what you'd like to say first, and then search for evidence to estimate it?

Of course not. The method is to analyze the evidence, come to a hypothesis, test the hypothesis, and then see whether a conclusion can be made. On the subject of rebound control, I'm still at the hypothesis stage. When I have some more good evidence, then I'll be able to add a little more precision.

Hardly a stat to rest your argument on.

I must honestly say I am stunned that you feel I even need to bring stats to the table to prove that New Jersey did not allow many shots against. This is hardly a contrarian viewpoint. Nevertheless, let me explain the significant of shot prevention. The average shooting percentage over the period looked at was about 10%. A difference of 6 shots per game means an expected difference of 0.6 goals against per game. That's about the difference in goals against between the New Jersey Devils and the St. Louis Blues this year, or between the Detroit Red Wings and Atlanta Thrashers.

Not to mention a 15-year period is a very large sample. If you look at any measure over that big of a sample, most teams will regress to something very close to the mean. To be 10% better than average over such a long time is very impressive.

If you want to dismiss New Jersey's excellent shot prevention and give Brodeur no credit for preventing shots, then you pretty much have to rate him based on his save percentage, which is good but not outstanding (something that has been repeatedly demonstrated in this space). I, on the other hand, being as biased I am, am making the effort to try to isolate the effect of Martin Brodeur so we can measure it and credit all goalies for their non-save contributions, if any.

Even with the large assumption that this study is on the up-and-up, it doesn't back up your argument.

Martin Brodeur faced fewer shots per game relative to the rest of the league in the playoffs than Ken Dryden did on the greatest team of all-time. Brodeur faced fewer adjusted shots per game than Gerry Cheevers did playing behind Bobby Orr, and fewer shots than Billy Smith did on the New York Islanders that won 19 consecutive playoff series. Somehow you interpret these facts as not showing that New Jersey was great at shot prevention?

By the way, thanks for the insulting insinuation that I am arbitrarily making up numbers for my own ideological views. Feel free to check the numbers yourself. Or if you want to see the spreadsheets and calculations I used, send me an email.

Who said it has no observable impact? All I said was that it wasn't quantifiable.

I meant that the impact would be observable in the numbers. Why isn't it quantifiable? Do rebounds not lead to goals against? If you take out a goalie who as great at controlling rebounds, and replace him with one who is terrible at it, don't you think goals against will go up?

You seem unwilling to throw out any numbers or guesses as to the effect. Do you have no idea what the effect of these things are? Still, you seem to trust scouts and the value of "watching them play", do they have the ability to accurately assess the value of these skills?

Evaluating a player or goalie based on a single skill is not difficult, and scouts are very good at breaking down skills. Assessing all of those skills in combination to reflect the player or goalie's overall contribution to winning hockey games is much more difficult. That is the goal of statistical analysis in hockey.

Anonymous said...

"It is true that the post-lockout Devils are not the shot prevention machine that the pre-lockout Devils were, but whether that's due to the departures of Stevens, Niedermayer and Daneyko or the Brodeur Rule is open to interpretation, now isn't it? Room to advance both arguments."

The Two-line Pass rule no longer being around has alot to do with it also.

Anonymous said...

I have noticed that since Luongo, (the guy you said qualifies as great) has come back, he has performed just as badly, if not more so than any of his backups. Fittingly nothing has been mentioned of this. Not that there is a problem with that, but I am just noting the contradictions because we all know if such a discrepancy occurs with Brodeur and his backups, it is immediately a headline on your blog, and more evidence used to support your opinions of Brodeur. This is the bias everybody talks about. However I assume you were just "waiting" for the Canucks to get on a role and maybe fix up those lousy numbers before you made your next post about how great Luongo is compared to his backups?

Bruce said...

The Two-line Pass rule no longer being around has alot to do with it also.

Good point, Anon999. That might have affected the Devils more than many teams cuz a big part of the Trap was to force the play before the opposition got to centre and could shoot the puck in. The fact the neutral zone was effectively compartmentalized into two (tiny) zones really played into their hands.

Anonymous said...

Obviously scouts, coaches & GM's evaluate each player accurately... that's why one of the greatest goalies of all-time -- Hasek -- wasn't even made a starter until the season he turned 29 (93-94) No need for stat analysis... just trust those who make the decisions... they're never wrong! :)

Anonymous said...

Re: Luongo & comparison to backups in 08-09... "I have noticed that since Luongo, (the guy you said qualifies as great) has come back, he has performed just as badly, if not more so than any of his backups."

Whhhaaa? In less than a minute I was able to find on the 'net these facts:

GP W L OTL GAA SV%
Roberto Luongo 27 14 7 5 2.61 .914
Curtis Sanford 19 7 8 0 2.59 .906
Cory Schneider 8 2 4 1 3.38 .877
Jason LaBarbera 6 2 1 2 3.07 .907

League average sv% = .908

http://sports.yahoo.com/nhl/teams/van;_ylt=AhiQB27y9KyA6_MHFYMn3y5ivLYF

Luongo doesn't appear to be playing amazingly well, but he certainly seems to be better than his fellow Canuck goalies as well as the league avg.

So how did you come to your conclusion? Saw a few highlights on t.v.? Probably.

Thanks for your continued articulate analysis... how about do some fact-checking before you spew, teen guy?

Anonymous said...

Of course not. The method is to analyze the evidence, come to a hypothesis, test the hypothesis, and then see whether a conclusion can be made. On the subject of rebound control, I'm still at the hypothesis stage. When I have some more good evidence, then I'll be able to add a little more precision.

The difference being that you are functioning under a preconceived notion and seek to prove it, rather than let the evidence show what it may.

I must honestly say I am stunned that you feel I even need to bring stats to the table to prove that New Jersey did not allow many shots against. This is hardly a contrarian viewpoint.

And I must honestly say that I am not really so stunned that you so flippantly utilize stats that are easily dismissed as, at the least, selectively chosen, and at worst, completely false. Not to mention present such stats as if they carry more or less weight depending on which you need to strengthen your argument.

Nevertheless, let me explain the significant of shot prevention. The average shooting percentage over the period looked at was about 10%. A difference of 6 shots per game means an expected difference of 0.6 goals against per game. That's about the difference in goals against between the New Jersey Devils and the St. Louis Blues this year, or between the Detroit Red Wings and Atlanta Thrashers.

Which, while true, is a misleading argument. It's a half-truth told as a whole-truth. Yes, NJ has given up .7 SOG less than St. Louis, and 35 less goals (which is, I do not argue, related), but the implication is that that is why NJ has a better record. There are more facets to that truth, like the fact that NJ has a goal differential of +36 to St. Louis' -14, or that Detroit is a +49 to Atlanta's -37.

If you want to dismiss New Jersey's excellent shot prevention...

As previously stated, you've not proven their shot prevention to be "excellent," but slightly above average.

...and give Brodeur no credit for preventing shots...

An as of yet unquantifiable statistic.

then you pretty much have to rate him based on his save percentage

Or, like most people do with all other goalies, could rate him based on things like wins, win percentages, Cups, various awards, save percentage, shutouts, playoff wins, shootout win percentages, etc. Why would someone rate him based solely on one stat?

which is good but not outstanding (something that has been repeatedly demonstrated in this space).

Which is why goalies, or any other player for that matter, aren't judged purely on one stat.

I, on the other hand, being as biased I am, am making the effort to try to isolate the effect of Martin Brodeur so we can measure it and credit all goalies for their non-save contributions, if any.

Then here's an interesting stat to throw into the mix: this season, the two best defensive teams in the league, Boston and Minnesota (who both play the much-maligned trap, btw), both allow .075 goals per shot given up (GAS%). Second best goes to Florida (believe it or not) with a .082 GAS%, followed by NJ and Chicago, both with a .086.

How much of that is system and how much of it is goalie? NJ doesn't utilize the trap anything like it used to, and certainly not as much as Boston or Minn, and without Brodeur's rebound control and puck handling, they still rank in the top three. Predictably, you'll argue that just proves Brodeur's ineffectiveness, but I counter that it further shows that hockey is a team sport and playing together and to your strengths leads to success and that things like rebound control and puck-handling do more to affect a team's strategy and overall style of play than singularly affect a team's success.

Martin Brodeur faced fewer shots per game relative to the rest of the league in the playoffs than Ken Dryden did on the greatest team of all-time. Brodeur faced fewer adjusted shots per game than Gerry Cheevers did playing behind Bobby Orr, and fewer shots than Billy Smith did on the New York Islanders that won 19 consecutive playoff series. Somehow you interpret these facts as not showing that New Jersey was great at shot prevention?

I repeat, "A recent study? Era-adjusted shots? Mm-k. Even with the large assumption that this study is on the up-and-up, it doesn't back up your argument."

I can quote "recent studies" all day too. How about a link? How about some proof of journalistic integrity?

I interpret facts when I see them, and I question BS when I see it.

By the way, thanks for the insulting insinuation that I am arbitrarily making up numbers for my own ideological views.

Forgive me. Allow me to openly say that you are arbitrarily making up, or else seeking out and selectively choosing, numbers to promote your preconceived notions.

Feel free to check the numbers yourself.

I have, as shown above in my last few posts.

I meant that the impact would be observable in the numbers. Why isn't it quantifiable? Do rebounds not lead to goals against? If you take out a goalie who as great at controlling rebounds, and replace him with one who is terrible at it, don't you think goals against will go up?

It depends. You once again present only two simple options, without considering any other realistic aspects of hockey that change when a goalie change is made.

You seem unwilling to throw out any numbers or guesses as to the effect.

Because, as I said, they're unquantifiable.

Do you have no idea what the effect of these things are?

I have a pretty fair idea of the effect, and I'm content with the fact that there isn't a numerical stat to define every last little bit of minutia of the sport.

Still, you seem to trust scouts and the value of "watching them play", do they have the ability to accurately assess the value of these skills?

Um, yes. That's their job. It's all in the job title. If they're no good at it, they probably don't keep their job for very long.

Evaluating a player or goalie based on a single skill is not difficult, and scouts are very good at breaking down skills. Assessing all of those skills in combination to reflect the player or goalie's overall contribution to winning hockey games is much more difficult.

Ok? So, what you're saying is, scouts have tough jobs?

That is the goal of statistical analysis in hockey.

Yes, but it isn't the be-all-end-all, defining answer. Numbers only tell a part of the story.

-e

The Contrarian Goaltender said...

I have noticed that since Luongo, (the guy you said qualifies as great) has come back, he has performed just as badly, if not more so than any of his backups. Fittingly nothing has been mentioned of this.

Luongo has been poor since he came back (.884, 3.50), I definitely agree. Then again, that is a 9 game sample size. In the 9 games before that he was .959, 1.22. I'm pretty sure Luongo isn't as good as he was playing then, and he isn't as bad as he is playing now. If he keeps playing the way he is now all season, however, then I will definitely downgrade him in my ranking.

When I refer to Luongo as great, I mean over the course of his career. Going through hot and cold stretches is part of goaltending. There have been long stretches and entire seasons where other goalies have outplayed Luongo. Several goalies, for example, would be ranked ahead of him in 2007-08, including Martin Brodeur.

The backup comparison in Vancouver is quite different than in New Jersey, however. Vancouver is 11-13-1 without Luongo while New Jersey is 29-15-1 without Brodeur, and the Devils' backups have much better stats. Yes there are a lot of team factors involved, there always are in goaltending, but those are very dissimilar records.

Anonymous said...

"Or, like most people do with all other goalies, could rate him based on things like wins, win percentages, Cups, various awards, save percentage, shutouts, playoff wins, shootout win percentages, etc. Why would someone rate him based solely on one stat?"

What weight do you give to each of these, & why?

What is a goalie's main function?

(Are you the Anonymous who has seen several NJ home games?)

By the way, spend some time reading the various posts/studies that CG has put on here over the years. A lot of your questions/accusations may be answered that way.

Anonymous said...

What weight do you give to each of these, & why?

I don't. Because when considering whether or not something is good, one doesn't look solely at the individual parts, but rather looks at the sum of its parts.

I don't buy a car only because it gets good fuel mileage. I don't order a steak because it cooks faster. Neither do I judge an athlete by one statistic.

What is a goalie's main function?

To win games, like any other player. Is this a trick question?

(Are you the Anonymous who has seen several NJ home games?)

Several? Yeah, I guess you could say I've seen several.

By the way, spend some time reading the various posts/studies that CG has put on here over the years. A lot of your questions/accusations may be answered that way.

You suggest this assuming I haven't already.

-e

Anonymous said...

"e" - How does a goalie win the game? By winning the game? Or by stopping the puck? If so, how is stopping the puck best measured? Can a goalie win 500+ games, or Cup(s), playing for mediocre (or worse) teams? Which is the least team-dependant stat? (Think about these things for awhile.... let me guess, you'll respond "you suggest this assuming I haven't already", haha)

You seem to have a very nebulous, unscientific way of assessment -- a typical diehard fan response when confronted with facts.

However, you seem satisifed with it, which is fine for you, so I suppose there is no need to try to persuade you (not that anyone invited you here, you inserted yourself)... or to respond to your vague, rather snotty prouncements (cars, steaks, goalies...? hilarious, man! And, "you suggest this assuming I haven't already" --Then why keep asking about things CG has already discussed in the past? Give me a break, you haven't read even 10% of the posts/studies on here.)

Go Devils Go!

Anonymous said...

"e" - Would the "several" (or more) NJ home games you've seen provide you with some unique insight into Brodeur's (or, any goalie's) abilities that aren't reflected in the stats?

The Contrarian Goaltender said...

Forgive me. Allow me to openly say that you are arbitrarily making up, or else seeking out and selectively choosing, numbers to promote your preconceived notions.

My site is read by and linked by a number of other hockey statistical analysts around the Internet, many of whom are better at math and statistics than I am. If I was arbitrarily pulling numbers out of nowhere or using horribly flawed statistics then I would have been called on it right away. When other sites have done similar work, their conclusions have generally been in agreement with mine. I wish I could claim that all my findings are highly original, but for the most part similar work has been done and and similar conclusions have been reached by many others before me, including my conclusions about Martin Brodeur.

I have done a variety of things on this blog, both attempts to evaluate goalies as well as investigations into other aspects of goaltending (like the post that led to this comment thread). In all of them I have explained my method and how I got my results. No doubt some of the techniques could be improved upon, and some of the numbers I have generated have been impacted by factors such as small sample size, reporting bias, lurking variables, etc., but the results are generally consistent in terms of their relative goalie rankings. Again, if you want to see the calculations, I can give them to you or tell you how to do them yourself.

Do I select what evidence I present and what examples I give? Yes, I freely admit that. So does every single other blogger, author, journalist and writer, because there is limited word space and we all have limited attention spans. Nobody wants to read a 5,000 word blog post on situational playoff save percentage. If people have additional questions regarding the results or methodology, that's what comments and email are for.

I also unashamedly admit to emphasizing certain stats and discounting others. That is because I believe certain stats are more valuable (stats like save percentage, ideally some form of shot quality adjusted save percentage, along with other team-adjusted stats). Giving goalies all the credit for wins and Cups is giving them credit for something that is mostly the work of their teammates.

Which, while true, is a misleading argument. It's a half-truth told as a whole-truth. Yes, NJ has given up .7 SOG less than St. Louis, and 35 less goals (which is, I do not argue, related), but the implication is that that is why NJ has a better record.

That implication is only there in your interpretation of what I clearly specified was "the difference in goals against" between those two teams. New Jersey is a far better defensive team than St. Louis. Do you disagree with that? As you point out, they are a much better offensive team as well but that is irrelevant to the point. I was trying to provide an example of the significance to a team's defensive record of allowing 0.6 fewer goals allowed per season, every season, for 15 seasons in a row. You are once again interpreting extra things from my writing. What was that you were saying about "preconceived notions"?

I can quote "recent studies" all day too. How about a link? How about some proof of journalistic integrity?

I'm sorry, it was the post right below this one, so I assumed you might have already seen it. Here is the link. Again, if you want a copy of my spreadsheet feel free to send me an email, and I will be happy to provide it.

I have, as shown above in my last few posts.

Great. When you find any errors or omissions, let me know and they will be corrected.

Because, as I said, they're unquantifiable.
I have a pretty fair idea of the effect

Do you not see the contradiction in these two statements? You feel you have a fair idea of the effect, but you can't assign a number to it, even an approximate estimate? How do you value goalie skills then, by using letter grades? Or do you assess goalies on the various component skills and then put it all together using some sort of magical evaluative process that somehow manages to take accurately into account the relative worth of each skill without ever actually knowing or defining the relative worth of each skill?

Again, I'm not saying I have numbers for all of things that happen in hockey, and I am not sure we will ever be able to quantify things like rebound control and puckhandling. I have repeated that many, many times. It certainly is not going to stop me from trying. If you want to wave your hands in the air and go around screaming, "Nothing is quantifiable!" then that's fine, and we have a difference of opinion.

I believe the problem we face in the real world is usually not that things can't be expressed in terms of numbers, but that it is difficult or costly to properly quantify things so we have to deal with imperfect stats. Goaltending is not a mystical black art that defies every form of numerical description, we merely have imperfect tracking and statkeeping. We should be counting things like rebounds allowed, scoring chances faced, completed passes and turnovers by goalies, pokechecks, intercepted passes, etc., just like is done with blocked shots, steals, hits, and other events.

Since those things are not being tracked, we have to content ourselves for now with trying to tease out evidence from the available data. Sometimes we will find nothing and other times we might find something of value. Generally to find something that is worthwhile you have to end up wading through a lot of stuff that is not so useful, but hitting a few dead ends does not invalidate the entire effort.

Predictably, you'll argue that just proves Brodeur's ineffectiveness, but I counter that it further shows that hockey is a team sport and playing together and to your strengths leads to success and that things like rebound control and puck-handling do more to affect a team's strategy and overall style of play than singularly affect a team's success.

Hockey is a team sport, the performance of individual players is affected by others on their team, rebound countrol and puckhandling are more an issue of style than impactful skills - that sounds like something I would write, because I don't disagree with any of that.

By the way, Brodeur is not "ineffective" and never has been. Some seasons he's been average, most of the time he's been above average, and there have been some stretches in his career where he has been undeniably great, in both the regular season and playoffs. He is clearly better than Clemmensen and Weekes, and New Jersey is better off with him in their net. However, just like every other goalie, Brodeur's results have been impacted by the rest of the team, and in his case most of the time the impact has been to his advantage. All goalie evaluation needs to take the rest of the team into account.

Anonymous said...

e - "I can quote "recent studies" all day too."

Ok, let's see some.

Anonymous said...

to the anonymous jumping in every other post riding CG's points as if they are your own:

your reading comprehension is obviously poor. that is probably why many of your remarks are ignored. if you would like to introduce some new point that hasn't already been used, do so. but please stop attempting to counter every point by crying save percentage, when using save percentage by itself to judge a goalie has already been shown as a flawed method.

also, the fact that you think that since returning from his injury, luongo has been better than his backups is hilarious. it just shows you do not really know what you are talking about, and that nothing you say should be taken seriously.

Anonymous said...

I often agree with CG, true.

My apologies re: Luongo. Obviously, the past 9 games since his injury his stats are not so hot.

Clearly, those 9 games override the few hundred previous games, where he exhibited qualilty at or near the top of the league.

It is shocking that CG has not yet spent the time examining these disastrous 9 games.

"...using save percentage by itself to judge a goalie has already been shown as a flawed method."

Uhhh... where exactly?

Nevertheless, how flawed? In your expert opinion, is it the best stat? The worst? What is your reasoning?

You offer only points that are just general "consensus"... Brodeur has Cups, won Gold medal, has 500+ wins... Brodeur great!!! Grow a brain, fanboy.

Anonymous said...

to put away this whole save percentage is the end all be all myth, look no further than dwayne rolosons numbers with minnesota, and then once he left. i also, using your logic must bring up niklas backstom as possibly the most underrated and by far THE GREATEST GOALIE EVER because of his amazing career save percentage of .925. wow, isnt that easy. i mean, why do we even keep track of stats such as wins, or goals against average? heck why do we have any goalie stats other than save percentage. we might as well wash away dryden, plante, and sawchuk from the record books because they are frauds since they dont have recorded save percentages, just meaningless wins.

so really what is obvious is you save percentage only junkies seem to prefer giving credit to guys who post high save percentages while only starting 40-50 games over the course of 4 or 5 years.

Anonymous said...

"Nevertheless, how flawed? In your expert opinion, is it the best stat? The worst? What is your reasoning?"

it is neither. it should be viewed as a supplemental statistic. it should be used in its correct context along with many other things that factor into a goalies value to a team. i think "e" put it best. you dont buy a car just because of its fuel efficiency. why would you possibly conclude that you can judge an athlete on one statistic. that is like rating a pitcher off his baa. i mean the pitchers job is to get the batter out. so wouldnt the pitcher with the lowest baa be "the best". the more this whole "only save percentage" thing is broken down the stupider it sounds, and the more bogus it seems that anyone who would want to be taken seriously would support using such a flawed method for evaluation.

Anonymous said...

My site is read by and linked by a number of other hockey statistical analysts around the Internet, many of whom are better at math and statistics than I am. If I was arbitrarily pulling numbers out of nowhere or using horribly flawed statistics then I would have been called on it right away. When other sites have done similar work, their conclusions have generally been in agreement with mine.

It's on the internet, it must be true!

Seriously, are you suggesting that there aren't groups who use one another's biased data to support their own bias?

I have done a variety of things on this blog, both attempts to evaluate goalies as well as investigations into other aspects of goaltending (like the post that led to this comment thread). In all of them I have explained my method and how I got my results. No doubt some of the techniques could be improved upon, and some of the numbers I have generated have been impacted by factors such as small sample size, reporting bias, lurking variables, etc., but the results are generally consistent in terms of their relative goalie rankings.

If flawed data matches other flawed data, they don't justify one another, and presenting either as factual is not only disingenuous, but downright unscrupulous.

Do I select what evidence I present and what examples I give? Yes, I freely admit that. So does every single other blogger, author, journalist and writer, because there is limited word space and we all have limited attention spans.

Yet you present such information as evidence to support your points.

Blaming space constraints and the reader is, well, for lack of a better word, disingenuous.

Nobody wants to read a 5,000 word blog post on situational playoff save percentage. If people have additional questions regarding the results or methodology, that's what comments and email are for.

I'm sure you underestimate the general population's interest in knowing the truth. I, personally, would rather have a lengthy truth than a shortened half-truth.

I also unashamedly admit to emphasizing certain stats and discounting others. That is because I believe certain stats are more valuable (stats like save percentage, ideally some form of shot quality adjusted save percentage, along with other team-adjusted stats).

I am quite amazed at how openly you admit to your deliberate tampering of information, but I appreciate your honesty.

Giving goalies all the credit for wins and Cups is giving them credit for something that is mostly the work of their teammates.

I think we're all in agreeance with that sentiment.

That implication is only there in your interpretation of what I clearly specified was "the difference in goals against" between those two teams.

The stat is presented as evidence to support your claim that goalies that can effectively puck-handle and control rebounds singularly results in fewer shots against.

The foundation of your argument is just not true.

The problem with many of your arguments is that you present extremes as the only alternatives, despite disclaiming an understanding that there are other factors involved. In this case, you state, "...the effect of a goalie who is great at controlling rebounds might not just be allowing fewer second chance shots, but also facing fewer total shots by deterring the opposition from shooting in the first place."

This speculation reduces the effects of every other aspect of the sport to a simple formula: If goalie A can do X, then team B does Y.

New Jersey is a far better defensive team than St. Louis. Do you disagree with that?

Nope.

As you point out, they are a much better offensive team as well but that is irrelevant to the point.

No, it isn't. It helps to illustrate that there are many other factors involved in why they're a better defensive team. An opponent can't score from their defensive zone. An opposing line is more likely to take a shot from the blue line or dump a puck into the offensive zone to get a line change if they've just spent a minute and a half fighting along the boards for possession.

I was trying to provide an example of the significance to a team's defensive record of allowing 0.6 fewer goals allowed per season, every season, for 15 seasons in a row.

No one is arguing it's not significant, merely that it isn't singularly attributable to one person.

I'm sorry, it was the post right below this one, so I assumed you might have already seen it. Here is the link.

A link which links to more of your own posts, all of which, I might add, contain posters also questioning your information and sources. How exactly is that supposed to convince me?

Great. When you find any errors or omissions, let me know and they will be corrected.

This word just keeps popping up, doesn't it? It's disingenuous of you to offer that I, in short order, sort through months, and perhaps years, of posts to find all of your inaccuracies.

How about just one? That way, we can all agree that if there is just one, there are probably more.

Above you state, "Since 1993-94, the Devils have allowed fewer shots against per game than any other team in the league."

Which I responded to in my post immediately following with, "just last season shows that the Red Wings were the best with only giving up 1930 shots, for a per game average of 23.5, whereas NJ gave up 2257, for a per game average of 27.5."

I also gave you the benefit of the doubt by assuming you meant on average, but when I requested your source, as I could not locate any site on the web that had the numbers that went back further than the lockout, you at first gave me an incorrect link, but when I found the site you referred to and still could not find the information, admitted that, "There is no season comparison of SOG," and provided a link to a page that lists Brodeur's career SOG.

Do you not see the contradiction in these two statements? You feel you have a fair idea of the effect, but you can't assign a number to it, even an approximate estimate?

Not at all. I "feel" I have a fair idea pretty much goes hand in hand with the concept of not being able to quantify something.

I feel love for my son, but I can't quantify it numerically. I have a fair idea of the effect telling my boss he's a jerk would have on my future employment, but I couldn't tell you what statistical chance I'd get fired would be.

How do you value goalie skills then, by using letter grades?

By their ability to play their role on a team successfully.

Or do you assess goalies on the various component skills and then put it all together using some sort of magical evaluative process that somehow manages to take accurately into account the relative worth of each skill without ever actually knowing or defining the relative worth of each skill?

Are those my only two options? Because if so, I don't think this test is very fair.

Again, I'm not saying I have numbers for all of things that happen in hockey, and I am not sure we will ever be able to quantify things like rebound control and puckhandling.

Precisely!

I have repeated that many, many times.

Hmm, I must have missed that in between all the arguing about how it's quantifiable.

It certainly is not going to stop me from trying.

I'm quite sure you'll eventually find some stats somewhere that, when presented in the right light, aid your argument.

If you want to wave your hands in the air and go around screaming, "Nothing is quantifiable!" then that's fine, and we have a difference of opinion.

I hardly think I've said that, now. Let's be fair. We do, however, have a difference in opinion.

I believe the problem we face in the real world is usually not that things can't be expressed in terms of numbers, but that it is difficult or costly to properly quantify things so we have to deal with imperfect stats.

I'm not sure if we're still talking about hockey here.

Goaltending is not a mystical black art that defies every form of numerical description, we merely have imperfect tracking and statkeeping. We should be counting things like rebounds allowed, scoring chances faced, completed passes and turnovers by goalies, pokechecks, intercepted passes, etc., just like is done with blocked shots, steals, hits, and other events.

I don't think anyone is arguing that goaltending is mystical. I, however, am arguing that even with all of that information, you can't quantify the effect on a game of something like puck-handling. The information required to even begin would be staggering, and there are far too many variables to consider for any of it to be taken as a final truth without a number of exceptions that would ultimately disqualify the information.

Since those things are not being tracked, we have to content ourselves for now with trying to tease out evidence from the available data. Sometimes we will find nothing and other times we might find something of value. Generally to find something that is worthwhile you have to end up wading through a lot of stuff that is not so useful, but hitting a few dead ends does not invalidate the entire effort.

Sometimes you just have to watch and go with your gut.

Hockey is a team sport, the performance of individual players is affected by others on their team, rebound countrol and puckhandling are more an issue of style than impactful skills - that sounds like something I would write, because I don't disagree with any of that.

And yet you spend inordinate amounts of time "...making the effort to try to isolate the effect of Martin Brodeur so we can measure it and credit all goalies for their non-save contributions, if any."

...However, just like every other goalie, Brodeur's results have been impacted by the rest of the team...

Agreed.

...and in his case most of the time the impact has been to his advantage.

And this is why he is a fraud?

All goalie evaluation needs to take the rest of the team into account.

Then why spend so much time trying to quantify their specific contributions?

-e

Anonymous said...

Anonymous said...

e - "I can quote "recent studies" all day too."

Ok, let's see some.


A recent study indicates that 84% of people refute evidence if it is contrary to what they want to believe.

A recent study shows that people are 67% more likely to be snarky online than in person.

-e

The Contrarian Goaltender said...

How about just one? That way, we can all agree that if there is just one, there are probably more.

Please try again. Please, for your own sake. Your repeated focus on shots against simply makes it look like you have no idea what you are talking about.

Nevertheless, here is my latest post on this blog that gives all the shot data you need and exactly how to find it.

By the way, the reason I couldn't give you a link to shot totals is that there is no magical link to all the information you need. Most of the time you have to do substantial work yourself to find out what you want, and it was even worse before Hockey-Reference showed up.

People like me spend hours sifting through data, compiling data, filling out spreadsheets, and running calculations. We don't do it because we want to irrationally slander people and evangelize our petty biased opinions to everyone, we do it because we want to drill down to the basic truths of the game of hockey and find better ways to evaluate players and goalies.

Anonymous said...

Please try again. Please, for your own sake. Your repeated focus on shots against simply makes it look like you have no idea what you are talking about.

Are you serious? This is an article about SOG! That was the topic at hand to which you quoted your fallacious information.

Nevertheless, here is my latest post on this blog that gives all the shot data you need and exactly how to find it.

You mean the "study" you've been pulling stats from that only just now exists? By your own hand?

So, in a 15 year period, the difference between top teams is .1 SOG per 60 minutes. Staggering evidence to NJ's superiority, but even still, as I stated before, it doesn't prove your point that Brodeur's abilities to handle the puck and control rebounds can be linked to those shot totals.

In point of fact, in 93-94, Brodeur faced more shots per game than Terreri did, and as I've pointed out before, for the last 3 years there have been at least 9 teams that give up fewer shots per game.

By the way, the reason I couldn't give you a link to shot totals is that there is no magical link to all the information you need. Most of the time you have to do substantial work yourself to find out what you want, and it was even worse before Hockey-Reference showed up.

By the way you were quoting information from "recent studies" and such, I guess I just assumed that they actually existed.

People like me spend hours sifting through data, compiling data, filling out spreadsheets, and running calculations. We don't do it because we want to irrationally slander people...

Odd name for a website, then.

...and evangelize our petty biased opinions to everyone...

Odd choice to publicly post them online, then.

...we do it because we want to drill down to the basic truths of the game of hockey...

Which, you admit previously, aren't all quantifiable.

...and find better ways to evaluate players and goalies.

Better than... what, exactly?

Altruistic BS aside, you're looking for statistical evidence to prove the points you want to make.

-e

The Contrarian Goaltender said...

Are you serious? This is an article about SOG! That was the topic at hand to which you quoted your fallacious information.

I'm sorry, I was referring to New Jersey's historical shot prevention. I thought that was clear by the quote I included right before my comments. Let me rephrase: Your comments in this thread about New Jersey and its history of shots against make it look like you have no idea what you are talking about.

Despite never having provided a single example of fallacious information, you continue to refer to everything I write as such, all while simultaneously maintaining that I have to meet a ridiculously high standard and write 5,000 words to fully explain every method I use.

You mean the "study" you've been pulling stats from that only just now exists? By your own hand?

Yes, those numbers just now came into existence and have never been seen before anywhere else. Over the last few hours, I have watched every NHL game since 1993-94, tracked all the shots, submitted it to the NHL and convinced them to make it part of their official record. My secret plan to disingenuously sabotage Martin Brodeur is complete!

Just because you are ignorant of something, E, does not mean it doesn't exist. I have told you twice now exactly where the numbers came from.

So, in a 15 year period, the difference between top teams is .1 SOG per 60 minutes.

Over the course of their careers, the difference between the two top per-game scorers ever is .04 points per game. This proves that Gretzky was not a superior player, and was in fact barely above average.

That would be a pretty silly claim to make, yet you are trying to use similar logic.

It doesn't prove your point that Brodeur's abilities to handle the puck and control rebounds can be linked to those shot totals.

Of course not. I have never claimed that it did. I merely referred to New Jersey's excellent history of shot prevention, and you jumped all over me for allegedly making up data.

If you want to read a post that deals directly with the issue of Brodeur and shots against, then read this one.

In point of fact, in 93-94, Brodeur faced more shots per game than Terreri did

While this is technically true, it just shows the world that you are completely unfamiliar with using hockey statistics. Just like goals against average is calculated as goals allowed per 60 minutes of play, the universal convention is to express shots against as shots against per 60 minutes of play. Based on the latter calculation in 1993-94, Brodeur was at 28.3 and Terreri 29.3.

By the way you were quoting information from "recent studies" and such, I guess I just assumed that they actually existed.

All the studies I referred to do exist and I provided you with links.

Odd name for a website, then.

See this is I think the basis of this disagreement - your failure to recognize the deliberately over-the-top nature of the name of this blog, and to interpret every single thing here in that light. Please read what I actually write, not what you assume I meant when I wrote it.

Which, you admit previously, aren't all quantifiable.

Another misquote. I said they may not be quantifiable, but the reasons for that are entirely because of the limitations of the data. With perfect tracking and statkeeping, everything in sports is quantifiable. Sports are activities with specific objectives under the constraint of specific rules and time limits. There are only a limited number of things goalies can do in the course of a hockey game. Watch the video and track the key events, and you will be able to quantify everything significant in goaltending.

Better than... what, exactly?

Better than your gut or my gut, for one. Better than the groupthink that is so prevalent in hockey. Better than the standard practice of awarding the Vezina Trophy to the goalie that plays the most games and records the most wins.

Altruistic BS aside, you're looking for statistical evidence to prove the points you want to make.

I'm looking for statistical evidence, period. When I find compelling evidence, I change my viewpoint. Most of the stuff I do is investigative, and there are many examples of me changing my mind. Just off the top of my head, I changed my view on Tony Esposito's playoff record, I have downgraded Patrick Roy and upgraded Dominik Hasek in my personal all-time ranking, I went from believing that goalies have zero impact on shots against to thinking they have some impact, I found out that Evgeni Nabokov is overrated and identified historically overlooked goalies like Dan Bouchard and Al Rollins, and over the last three years I have moved Martin Brodeur from outside of my top 15 all-time to inside my top 10 because over the past 2 and a half seasons he has actually played like a top goaltender.

Anonymous said...

"over the last three years I have moved Martin Brodeur from outside of my top 15 all-time to inside my top 10 because over the past 2 and a half seasons he has actually played like a top goaltender."

More bullshit. The only difference between Brodeurs performance the past 2 years and before then was a miniscule improvement in save percentage, so I would be interested in seeing what exactly changed your mind on Brodeur. If I had to guess I would say that the more Brodeur plays, the stupider your argument becomes, therefor forcing you to change it over time.
In fact, your save percentage argument is completely reliant on Brodeur's save percentage numbers from 98-02. Other than those 4 years, your argument, even in regards to save percentage, is pathetic. So, god forbid a guy has 4 average or even slightly below average years over the course of a 15+ year career. Shit, it must mean he is a fraud.

The Contrarian Goaltender said...

The only difference between Brodeurs performance the past 2 years and before then was a miniscule improvement in save percentage,

There are two key differences you are forgetting. First of all, pre-lockout there was the league-wide tight defensively that characterized the so-called "Dead Puck Era" and inflated save percentages. Secondly, Brodeur had Scott Stevens and Scott Niedermayer on his team.

That is why the difference between Brodeur's play in 2006-07 and 2007-08 is substantially more impressive than earlier. His team defence got worse, and the league average save percentage went down, and yet his save percentages went up.

To express it in relative terms, I saw one shot quality save percentage measurement that had Brodeur very close to leading the league in 2007-08. In 2003-04, a similar shot quality measurement had him as an average goalie.

In fact, your save percentage argument is completely reliant on Brodeur's save percentage numbers from 98-02.

No, again you are failing to take into account that the Devils were a great defensive team. Based on raw save percentage numbers you would be correct, but we have shot quality estimates beginning in 2002-03 that show New Jersey allowed the easiest shot quality against. I'd say Brodeur's "average" period extends from 1998-99 to 2005-06.

These are his age 26 to age 33 seasons, which are typically a goalie's prime years, which whatever you think of Brodeur's play is certainly not a typical career progression. That's why I have referred before to Brodeur's "upside-down career curve" - good early, average during his expected prime years, and then good late.

Anonymous said...

I'm sorry, I was referring to New Jersey's historical shot prevention. I thought that was clear by the quote I included right before my comments. Let me rephrase: Your comments in this thread about New Jersey and its history of shots against make it look like you have no idea what you are talking about.

Ah, you mean the information you referred to several times that you only put up after I made my post? Yes, I can see how that, coupled with 3 years of stats that show your statement to be incorrect would suggest I don't know what I'm talking about.

Despite never having provided a single example of fallacious information, you continue to refer to everything I write as such, all while simultaneously maintaining that I have to meet a ridiculously high standard and write 5,000 words to fully explain every method I use.

Proving the authenticity of your sources and full disclosure of statistics you base your arguments on is hardly a high standard.

Yes, those numbers just now came into existence and have never been seen before anywhere else.

How does one cite a source that doesn't exist? Or if it does exist, how is one unable to provide a link to it?

Just because you are ignorant of something, E, does not mean it doesn't exist. I have told you twice now exactly where the numbers came from.

You'll have to forgive me for not taking someone of admitted bias' word for it and actually asking for the information.

Over the course of their careers, the difference between the two top per-game scorers ever is .04 points per game. This proves that Gretzky was not a superior player, and was in fact barely above average.

That would be a pretty silly claim to make, yet you are trying to use similar logic.


Actually, my point is that numbers alone are an unreliable method to judge a players performance, no matter how many of them you throw out there. That example just provides further evidence that your quest to single out an individual's contribution in a team sport is futile.

Of course not. I have never claimed that it did.

Really? Because above you wrote: "the effect of a goalie who is great at controlling rebounds might not just be allowing fewer second chance shots, but also facing fewer total shots by deterring the opposition from shooting in the first place."

I merely referred to New Jersey's excellent history of shot prevention, and you jumped all over me for allegedly making up data.

As you can see from the above quote, you did more than that, and I jumped on your illogical conclusion. You attempted to back up your assertion with more statistics, which is what I questioned.

While this is technically true, it just shows the world that you are completely unfamiliar with using hockey statistics. Just like goals against average is calculated as goals allowed per 60 minutes of play, the universal convention is to express shots against as shots against per 60 minutes of play.

Oh no, I'm found out! I didn't actually read the NHL Guide to Understanding Useless Statistics to Make Dumb Arguments pamphlet and thought I could slip one by you by logically calculating an actual shots per game total, rather than the totally obvious concept of dividing by 60 minutes.

I'm found out! My credibility is destroyed!

All the studies I referred to do exist and I provided you with links.

After the fact, and not when I actually requested the link for it.

See this is I think the basis of this disagreement - your failure to recognize the deliberately over-the-top nature of the name of this blog, and to interpret every single thing here in that light. Please read what I actually write, not what you assume I meant when I wrote it.

Everything I've read here concerning goalies, and in particular, Brodeur, all share in common that you, whether outright or subtly, attempt to use statistical figures to further your agenda. Even if the name of your site were something more neutrally sounding, it would be pretty apparent to an objective reader just what your agenda is.

Another misquote. I said they may not be quantifiable, but the reasons for that are entirely because of the limitations of the data. With perfect tracking and statkeeping, everything in sports is quantifiable. Sports are activities with specific objectives under the constraint of specific rules and time limits. There are only a limited number of things goalies can do in the course of a hockey game. Watch the video and track the key events, and you will be able to quantify everything significant in goaltending.

Wrong. You cannot truly quantify the effect of an individual's total significant impact on a team sport like hockey.

You can track save percentages and shots on goal and every other number until you're blue in the face, but some things just can't be put into numbers.

For example: How does one quantify what effect illness or injury to one player has in regards to how the team performs? How does one quantify the effect of a high scoring forward driving the net while a teammate has the puck on a defense or goalie?

Hockey isn't baseball. Baseball, though a team activity, is, in reality, a pitcher vs. a hitter. Statistics there are much more reliable because other individual's performances that day don't impact either the pitcher or the batter.

Hockey is, on the other hand, far more than just a shooter vs. a goalie. The closest thing hockey has is the shootout, which seems to be an oft overlooked statistic in conversations about goalies.

Better than your gut or my gut, for one.

You undervalue instinct and intangible factors. Sometimes coaches mix up lines. Do you think they consult stat sheets to figure out who should be where? No. They go with their gut feeling that putting this player and that on a line together might shake out some offense, or that putting goalie B in net that night is a better idea than goalie A.

Better than the groupthink that is so prevalent in hockey. Better than the standard practice of awarding the Vezina Trophy to the goalie that plays the most games and records the most wins.

The Vezina goes to the goaltender judged to be the best at his position and most valuable to his team during the regular season.

Having the durability to play more games than any other goalie, and win more games than any other goalie, defines that criteria.

Besides, the GM's vote on it, not the fans or the media. Or are you saying they're all just wrapped up in the hype also?

I'm looking for statistical evidence, period. When I find compelling evidence, I change my viewpoint.

Or, admittedly, don't use it or otherwise find ways to manipulate it.

Most of the stuff I do is investigative, and there are many examples of me changing my mind. Just off the top of my head, I changed my view on Tony Esposito's playoff record, I have downgraded Patrick Roy and upgraded Dominik Hasek in my personal all-time ranking, I went from believing that goalies have zero impact on shots against to thinking they have some impact, I found out that Evgeni Nabokov is overrated and identified historically overlooked goalies like Dan Bouchard and Al Rollins, and over the last three years I have moved Martin Brodeur from outside of my top 15 all-time to inside my top 10 because over the past 2 and a half seasons he has actually played like a top goaltender.

Anonymous has already called you on this.

-e

Anonymous said...

Speaking of following a logical line of thinking, is it fair to assume Nicklas Backstrom is the greatest goalie to ever play due to his .925 career save percentage? You talk about how in order for someone to rank Brodeur as the best, that same person would be expected to rank Glen Hall number 2 based on consistency and durability; well if using the save percentage argument, it would only be logical to call Nick Backstrom the best. But of course the trend I have noticed is that if there is a result you dont agree with, then you "adjust" the numbers whether using this bogus "shot quality" stat or some other method of making the numbers say what you want. Have you read the book "How to Lie with Numbers"? Perhaps you should. or maybe, you already did.

Anonymous said...

Ah yes... evaluting by using "the gut"... especially satisfying when a favourite goalie plays for an excellent team & therefore wins 40+ games & Cups even during seasons that his actual ability to stop to the puck (measured by.... ?) is nothing special.

He has 40 win seasons... people say he's great... therefore, he IS great!

Easy!

Anonymous said...

ALL HAIL NICKLAS BACKSTROM!! THE SAVE PERCENTAGE KING AND THE GREATEST GOAIE TO EVER PLAY HOCKEY

Anonymous said...

Backstrom?

What is his shot-quality-neutral SV%? What is his SV% at EV? While on PK? How do these compare to the league avg, as a %?

Anonymous said...

Ah yes... evaluting by using "the gut"... especially satisfying when a favourite goalie plays for an excellent team & therefore wins 40+ games & Cups even during seasons that his actual ability to stop to the puck (measured by.... ?) is nothing special.

He has 40 win seasons... people say he's great... therefore, he IS great!

Easy!


Try and keep up with the argument here before butting in.

-e

Anonymous said...

But that IS your argument, "e".

As for "butting in", I've been reading this site & posting here far far longer than you. Since I've read most of the articles & studies here, I don't need to send off 2-page diatribes bemoaning the supposed inadequacies of the observations & studies... without having read any of the background material or being familiar with basic stats.

Anonymous said...

But that IS your argument, "e".

Let's see if we can count the argumentative fallacies.

For starters, you assume knowledge you don't have (who my favorite goalie is).

You assume knowledge of my opinion (that I assert Brodeur is "great").

You extract a single point I've made (that sometimes a hunch is as good as anything) and twist it to attempt to represent something I never said, that I argue it is the only method of judging a player.

And then you attempt to discredit me by implying that I have no individual opinion and follow along with what "people say."

That about sum it up?

As for "butting in", I've been reading this site & posting here far far longer than you.

Ah, then your opinion is obviously far more important than mine. Feel free to jump into conversations you clearly haven't read or comprehended and put your argumentative shortcomings on display.

Since I've read most of the articles & studies here, I don't need to send off 2-page diatribes bemoaning the supposed inadequacies of the observations & studies... without having read any of the background material or being familiar with basic stats.

Another failed assumption.

Care to continue?

-e

Anonymous said...

"E", you clearly love to argue for the sake of arguing.

My comments are not entirely directed towards you, "e". They are also directed towards the other Anonymous... & towards the general fanbase that doesn't do much thinking beyond "40+ wins, Cup(s), Gold medal, therefore Great."

"You assume knowledge of my opinion (that I assert Brodeur is "great")."

So you *don't* think Brodeur is "great"? Hmmm, you sure seem offended by even the title of this Blog.

"Another failed assumption". Err... it is pretty obvious that you haven't read most of the articles/blogs/studies on this site, or other hockey stats sites. Numerous times you've indicated that you're uncertain where CG has found numbers & studies, to the point where you infer that he's making things up. Then he holds your hand & provides the links for you, & you move onto some other argument.

You just like to argue, plain & simple. Have fun with that.

Anonymous said...

"Backstrom?

What is his shot-quality-neutral SV%? What is his SV% at EV? While on PK? How do these compare to the league avg, as a %?"

Once again butting in without reading, otherwise you wouldnt have been silly enough to do exactly would I called you out on being likely to do before you even did it. Using that same flawed shot quality neutral crap and the adjusted numbers.

Bottom line, no ones cares about this "shot quality neutral" stat because it is so flawed there is little to no use for it. NHL.com has a ton of statistics, as does hockeydb, but shot quality neutral is not one of them. Why might this be? Heck scouts don't even use it. But I guess you know everything.

Anonymous said...

thats funny, the only place that i can find this so called gold measuring stick of goalie greatness called shot quality neutral save percentage is on a few blogs. i mean it is nowhere else. i was fairly certain that if it was at the very least bit accurate or useful that is would be more common to find. i guess not. however i have decided to make up my own stat called goal scoring neutral shooting percentage in which i have found dino ciccarelli to be the best forward ever! now i just need to start up my own blog and find a bunch of gretzky and lemieux haters to support it.

Anonymous said...

"E", you clearly love to argue for the sake of arguing.

I enjoy a good debate, tis true. Which is why it pains me to see someone try so hard and fail.

My comments are not entirely directed towards you, "e". They are also directed towards the other Anonymous...

Then you should not begin your comments by specifically naming me.

...& towards the general fanbase that doesn't do much thinking beyond "40+ wins, Cup(s), Gold medal, therefore Great."

Well, I can't speak for them, but I can certainly see how looking at an accomplished and decorated athlete might lead some to believe that he was great at his job.

So you *don't* think Brodeur is "great"?

My personal feelings are irrelevant to this discussion.

Hmmm, you sure seem offended by even the title of this Blog.

Ignorance and those that prey upon it with self-aggrandizing propaganda are enemies of mine.

Err... it is pretty obvious that you haven't read most of the articles/blogs/studies on this site, or other hockey stats sites.

Numerous times you've indicated that you're uncertain where CG has found numbers & studies, to the point where you infer that he's making things up. Then he holds your hand & provides the links for you, & you move onto some other argument.


When someone makes a claim, regardless of what it is, the onus is on them to prove it. Some people may take them at their word and some people may be convinced with a few half-truths and manipulated evidence, but some people actually require proof, and if that proof stems from other sources, it is on the claimant to show that those sources are legitimate.

This is how it works in the real world. Go to a website populated by intelligent people who actually discuss things instead of just reaffirming one another's bias, and you'll find the most commonly used phrase is, "Cite?" Meaning they're requesting a citation to back up a claim.

So, when CG cites a source that I am unable to find, I'm going to request a link to it (and, no, a link to another of your own articles does not count). When no link can be provided, the entire conclusion is called into question. See how that works?

You just like to argue, plain & simple. Have fun with that.

I like the truth, and don't have any qualms about digging a little to get there.

You may think it arguing for it's own sake, but that suits me just fine. All I asked from the beginning was that you not butt in unless you had something to actually add to the conversation.

-e

Anonymous said...
This comment has been removed by the author.
Anonymous said...

1. So you *don't* think Brodeur is "great"?

My personal feelings are irrelevant to this discussion.


Oh, I see... your endless "challenges" & assertions have nothing to do with your personal feelings... funny, because you've generally said that you assess a goalie by amalgamating all data into that supercomputer brain of yours & therefore glean the quality/greatness of the athlete. I would call that using your instinct, or "gut", or personal feelings.

2. You really feel that this site is "self-aggrandizing propaganda"? Again, go back & read the blogs from the beginning... go back 2-3 yrs & put things into context.

If anything, your passive-aggressive, vague statements, haughty lecturing yet lack of evidence provided (where the onus is always on the other person aka hold your hand & guide you through multiple websites & past studies) show you to be rather self-aggrandizing.

3. All I asked from the beginning was that you not butt in unless you had something to actually add to the conversation.

If you really were in a "conversation" with CG, & desired no comments from others, you'd email him directly rather than posting on a blog. That's the nature of a blog. But it appears that you enjoy seeing your words posted online.

Anonymous said...

Oh, I see... your endless "challenges" & assertions have nothing to do with your personal feelings...

My personal feelings are irrelevant because "great" is a subjective term, and not the issue at hand.

The issue at hand concerns shots on goal in relation to a goalie's ability to puck-handle and control rebounds.

My, or anyone else's, opinion on one, or all, goalies is a distracting, unnecessary tangent.

...funny, because you've generally said that you assess a goalie by amalgamating all data into that supercomputer brain of yours & therefore glean the quality/greatness of the athlete.

No. I've specifically said that some things can't be quantified and that numbers don't tell the whole story.

I would call that using your instinct, or "gut", or personal feelings.

Then you would be wrong.

I would call that assessing a player purely on provided data. In fact, that would be the exact opposite of going with one's gut.

Besides, instincts and personal feelings are different things.

2. You really feel that this site is "self-aggrandizing propaganda"? Again, go back & read the blogs from the beginning... go back 2-3 yrs & put things into context.

Faulty logic is being used to argue the above conclusion. I fail to see how additional context is going to change that.

If anything, your passive-aggressive, vague statements, haughty lecturing...

My statements have been neither passive-aggressive nor vague. From the beginning I've stated my argument in as clear a fashion as I could. If you think that haughty, then that's your perception.

My handling of your argument, perhaps, has been haughty, but that is because it is difficult to keep some level of condescension from creeping in when dealing with someone who makes all sorts of ridiculous assumptions and argumentative fallacies in a weak attempt to discredit the person, rather than their argument.

...yet lack of evidence provided (where the onus is always on the other person aka hold your hand & guide you through multiple websites & past studies) show you to be rather self-aggrandizing.

The onus is on the person making the claim.

If I told the scientific community that I had arrived at a newer, better theory of relativity, it would be on me to prove that my assertion were true... not on them to disprove it.

As to "holding hands," I'm not sure what you mean.

If you really were in a "conversation" with CG, & desired no comments from others, you'd email him directly rather than posting on a blog. That's the nature of a blog. But it appears that you enjoy seeing your words posted online.

I have no issues with others joining the conversation. I'm quite aware of the nature of public posting.

I welcome your input, if you have any to offer. As I said, "All I asked from the beginning was that you not butt in unless you had something to actually add to the conversation."

-e

Anonymous said...

NICKLAS BACKSTROM IS THE BEST EVER!! .925 SAVE PERCENTAGE

The Contrarian Goaltender said...

Yes, I can see how that, coupled with 3 years of stats that show your statement to be incorrect would suggest I don't know what I'm talking about.

Ask a random sample of hockey fans which team has given up the fewest shots per game since 1993-94. I bet the majority of them will pick New Jersey, because it is pretty well known that they had a great defence for a long time. Yet you refuse to agree that they were anything better than slightly above average, even when I post the numbers. Ignorance of an obvious fact and stubbornness to recant when presented with proof is, in my book, evidence that somebody doesn't know much about the subject matter at hand.

"The fewest shots against since 1993-94" does not mean that New Jersey led the league every season since 1993-94. It means just what it says, that over the period from 1993-94 to present they rank #1. You haven't proven anything to be incorrect, in this case or anywhere else. I am of course quite familiar with New Jersey's defensive record over the last 3 seasons. That is why I give Brodeur more credit for his recent performance.

Proving the authenticity of your sources and full disclosure of statistics you base your arguments on is hardly a high standard.

As far as I know I have cited every source that you asked me about. If there is anything that remains unclear as to the source, just ask. Yet I again I invite you to email for a copy of a spreadsheet or more information if you would like to peruse my data or calculations.

Every time I post some data for some sort of detailed study, I always mention factors that may be influencing the data or any caveats that the reader should know about. Take my work on adjusted playoff save percentage that you completely dismissed before you even read. I described the objective of my study, the reasons for undertaking it, my methods, the justification for my methods, how I calculated my adjustments, what period I was using and why I was limited to using that period. I also pointed out several factors that could be impacting the data, including goalie utilization and that save percentage does not take into account the strength of the team in front of the goalie. I then posted all of the results for goalies that met my specified minimum games criteria, and then discussed some of the results I found particularly interesting.

So please fill me in, what did I leave out that would turn that post from a series of lies, flawed data and half-truths to something that would be able to pass your apparently extensive list of qualifying criteria? Do you disagree with the method, or merely with the results?

How does one cite a source that doesn't exist? Or if it does exist, how is one unable to provide a link to it?

The source is Hockey-Reference.com, as I outlined twice before, a site that posts data from the official records of the National Hockey League. That is the source for the vast majority of my posts. When you want to look at aggregate data for multiple teams over multiple seasons, links to the specific data are not usually available. However, just because data relevant to that specific query are not already nicely compiled doesn't mean the underlying data do not exist. One merely has to go through and collect and compile the numbers, as I did at your request.

You'll have to forgive me for not taking someone of admitted bias' word for it and actually asking for the information.

I have no problem with you asking for the information. It shows you are curious. I wish you would show even more interest in my numbers, information and methods because then you would realize that they are correct. Unfortunately, you seem to have some conviction that having an opinion on something that is different than your own = falsifying data. All while claiming that your personal feelings are "irrelevant".

Actually, my point is that numbers alone are an unreliable method to judge a players performance, no matter how many of them you throw out there. That example just provides further evidence that your quest to single out an individual's contribution in a team sport is futile.

How does it show that? It only shows that Lemieux and Gretzky both scored points at a very high rate. Ask scouts who have never looked at a stat sheet if Lemieux and Gretzky were two of the greatest players ever, and they will tell you yes. Ask a stats geek who has never seen a game of hockey if they were two of the greatest players ever, and he will look at their fantastic numbers and also say yes.

Really? Because above you wrote: "the effect of a goalie who is great at controlling rebounds might not just be allowing fewer second chance shots, but also facing fewer total shots by deterring the opposition from shooting in the first place."

Yes, that is a correct quote giving my general hypothesis on the topic. I never claimed that just because New Jersey gave up fewer shots than normal means that it was necessarily caused by Brodeur. Maybe they would have given up even fewer with some other goalie in net. Again, it is a simple logical fallacy: If there is some A that causes B, and B is observed, that does not mean that A is present. A could be present, but it is not guaranteed, because the causal link goes only in one direction. I have found some evidence that Brodeur was assisting in preventing shots through comparative analysis, but the mere fact that New Jersey was a consistently great defensive team doesn't prove anything regarding their goalies.

Everything I've read here concerning goalies, and in particular, Brodeur, all share in common that you, whether outright or subtly, attempt to use statistical figures to further your agenda. Even if the name of your site were something more neutrally sounding, it would be pretty apparent to an objective reader just what your agenda is.

The feedback I have received from people with a statistical background I respect a lot more than yours seems to be pretty positive and agreeable. I guess they are all just biased Brodeur haters?

For example: How does one quantify what effect illness or injury to one player has in regards to how the team performs?

One looks at how the team performs with them in the lineup compared to how they perform without them. Let's see: Change one thing, hold everything constant, observe result. The scientific method in a nutshell. Oh, but I forgot, for some reason that works for everything in the world except for team sports.

You undervalue instinct and intangible factors. Sometimes coaches mix up lines. Do you think they consult stat sheets to figure out who should be where? No. They go with their gut feeling that putting this player and that on a line together might shake out some offense, or that putting goalie B in net that night is a better idea than goalie A.

Some coaches do consult stat sheets, while others often follow their instincts. That doesn't mean I'm undervaluing instinct, because their gut feel is not always correct. Does goaltending B play better than goaltender A in the games where his coach had a feeling he would play well? Not always. Sometimes they work and sometimes they don't, and it would be difficult to assess the overall impact on the team's success.

By the way, you consider the actions of NHL coachces to be acceptable evidence? Terrific. Let me repost the Barry Trotz quote from earlier in this thread: "Clemmenson has been good for them, although they do a good job of protecting him. Put the puck on net with him, from anywhere on the ice. He isn't Brodeur, you don't need a perfect shot to beat him, and if you get the puck in deep, keep it low. He gives up lots of rebounds."

Or, admittedly, don't use it or otherwise find ways to manipulate it.

Apparently you think I admitted to a lot of things that I never actually admitted to. Quick recap: Worthless stats are worthless. I don't put any weighting in worthless stats. Therefore I don't use them. That is not manipulating data, it is interpreting data based on its significance.

I was talking about compelling evidence. I have never found compelling evidence that Brodeur is even close to one of the all-time greatest goalies. Nothing I have ever done, or seen anyone else ever do, has made me question my overall viewpoint. I have modified my viewpoint, as previously mentioned, based on Brodeur's performance. But a variety of different studies looking at a number of different traditional and non-traditional stats pretty much agree that Brodeur is a good goalie who played most of his career behind a great defence. The only area that remains to study is non-save goalie skills, and to try to assess Brodeur's impact there. Conceivably it could be found that there is a significant effect from those skills that upgrades his all-time status, but a lot more study is required.

The Contrarian Goaltender said...

NICKLAS BACKSTROM IS THE BEST EVER!! .925 SAVE PERCENTAGE

Let me quote my overall description of this blog and its intent:

In hockey, goalies get all the credit. But recent analysis shows that more often than not it is the team that determines the 'tender's success, not the other way around. This recreational goalie identifies goalies who receive the credit that should be going to the teams in front of them.

The rest of the team has a big impact on goalie stats. That is especially true in the case of Niklas Backstrom. His coach is Jacques Lemaire, his team is an outstanding defensive team, and his backups have posted a .915 save percentage while league average has been .906. He also has only played two and a half seasons, and there have been many goalies who have started their careers well and then fallen off the pace. Take all those things into accout, and there is no reason to rank Backstrom among the best to ever play.

Nevertheless, even if you were to look at unadjusted save percentages only and fail to consider career length, Niklas Backstrom is still not the best goalie ever because he has played at a time when the league average save percentage is fairly high. Dominik Hasek has him beat in save percentage compared to league average.

Anonymous said...

Ask a random sample of hockey fans which team has given up the fewest shots per game since 1993-94. I bet the majority of them will pick New Jersey, because it is pretty well known that they had a great defence for a long time. Yet you refuse to agree that they were anything better than slightly above average, even when I post the numbers.

My disagreement isn't about their defensive play, but, still, that Brodeur's ability to handle the puck and control rebounds quantifiably affected those totals.

Ignorance of an obvious fact and stubbornness to recant when presented with proof is, in my book, evidence that somebody doesn't know much about the subject matter at hand.

Showing that the team faced less shots over a period of time than other teams doesn't prove that one person had anything to do with it at all.

"The fewest shots against since 1993-94" does not mean that New Jersey led the league every season since 1993-94. It means just what it says, that over the period from 1993-94 to present they rank #1.

Which is a perfect example of your deliberately presenting evidence in a manner that reinforces your argument, and while not technically being false, isn't entirely true.

Originally, you put it, "Since 1993-94, the Devils have allowed fewer shots against per game than any other team in the league. If being the best in the league over a 15 year span is not excellent, then what is exactly?"

Which suggests not that they, over 15 years averaged .1 less than the next best team, but that they were consistently "the best in the league over a 15 year span" and were "excellent."

Your choices of words and presentation of facts is deliberately misleading.

As far as I know I have cited every source that you asked me about. If there is anything that remains unclear as to the source, just ask. Yet I again I invite you to email for a copy of a spreadsheet or more information if you would like to peruse my data or calculations.

You've linked to your own articles, which in turn link to your other articles, or else linked to a site that one could potentially garner such information from, but not any actual neutrally compiled stats.

I don't have to "prove" your information incorrect; you've already admitted, "Do I select what evidence I present and what examples I give? Yes, I freely admit that. So does every single other blogger, author, journalist and writer, because there is limited word space and we all have limited attention spans."

Every time I post some data for some sort of detailed study, I always mention factors that may be influencing the data or any caveats that the reader should know about. Take my work on adjusted playoff save percentage that you completely dismissed before you even read.

I dismissed it because A) it was irrelevant to the conversation, B) you freely adjust stats to get the info you desire C) you admit the base information is flawed but very scientifically speculate that "I don't think it would have much of an impact over the course of a career" and D) follow up with a disclaimer that it doesn't take into account team defense, which may or may not entirely skew the chart.


I described the objective of my study, the reasons for undertaking it, my methods, the justification for my methods, how I calculated my adjustments, what period I was using and why I was limited to using that period. I also pointed out several factors that could be impacting the data, including goalie utilization and that save percentage does not take into account the strength of the team in front of the goalie. I then posted all of the results for goalies that met my specified minimum games criteria, and then discussed some of the results I found particularly interesting.

So please fill me in, what did I leave out that would turn that post from a series of lies, flawed data and half-truths to something that would be able to pass your apparently extensive list of qualifying criteria? Do you disagree with the method, or merely with the results?


See above.

The source is Hockey-Reference.com, as I outlined twice before, a site that posts data from the official records of the National Hockey League. That is the source for the vast majority of my posts. When you want to look at aggregate data for multiple teams over multiple seasons, links to the specific data are not usually available. However, just because data relevant to that specific query are not already nicely compiled doesn't mean the underlying data do not exist. One merely has to go through and collect and compile the numbers, as I did at your request.

Which doesn't answer my original question, how were you able to quote these numbers if they weren't compiled yet?

Unfortunately, you seem to have some conviction that having an opinion on something that is different than your own = falsifying data.

And I quote, "Do I select what evidence I present and what examples I give? Yes, I freely admit that.

All while claiming that your personal feelings are "irrelevant".

Are you arguing that personal feelings are somehow relevant to a factual discussion?

How does it show that? It only shows that Lemieux and Gretzky both scored points at a very high rate. Ask scouts who have never looked at a stat sheet if Lemieux and Gretzky were two of the greatest players ever, and they will tell you yes. Ask a stats geek who has never seen a game of hockey if they were two of the greatest players ever, and he will look at their fantastic numbers and also say yes.

Citing two players that people might (emphasis on might, since this is pure speculation on your part) subjectively call "great" based purely on either stats or having scouted them doesn't prove your point that you can accurately judge a player purely on statistics.

This is another example, of which your site is full, of writing which on the surface sounds accurate, but is in reality little more than misleading nonsense.

Yes, that is a correct quote giving my general hypothesis on the topic. I never claimed that just because New Jersey gave up fewer shots than normal means that it was necessarily caused by Brodeur.

Interesting, because later on you wrote that it "...still could be argued that excellence in those types of goalie skills do effectively "prevent" or "create" shots against..."

Maybe they would have given up even fewer with some other goalie in net.

Like, oh, say, Clemmensen?

Again, it is a simple logical fallacy: If there is some A that causes B, and B is observed, that does not mean that A is present. A could be present, but it is not guaranteed, because the causal link goes only in one direction. I have found some evidence that Brodeur was assisting in preventing shots through comparative analysis, but the mere fact that New Jersey was a consistently great defensive team doesn't prove anything regarding their goalies.

Precisely!

The feedback I have received from people with a statistical background I respect a lot more than yours seems to be pretty positive and agreeable. I guess they are all just biased Brodeur haters?

Or perhaps they've only read a few posts that jibed with their sentiments.

One looks at how the team performs with them in the lineup compared to how they perform without them. Let's see: Change one thing, hold everything constant, observe result.

You can't hold everything else constant. That is the problem with your analysis.

The scientific method in a nutshell. Oh, but I forgot, for some reason that works for everything in the world except for team sports.

The scientific method requires repeatable tests to first, support predicted results, and then be able to be predictably shown through repeated testing.

If you give a defenseman the flu and throw him out on the ice to play a game, are you suggesting that the results will be predictable and repeatable?

By the way, you consider the actions of NHL coachces to be acceptable evidence? Terrific. Let me repost the Barry Trotz quote from earlier in this thread: "Clemmenson has been good for them, although they do a good job of protecting him. Put the puck on net with him, from anywhere on the ice. He isn't Brodeur, you don't need a perfect shot to beat him, and if you get the puck in deep, keep it low. He gives up lots of rebounds."

No, I maintain that you place too much importance on statistics and don't give enough credit to going with one's gut.

Apparently you think I admitted to a lot of things that I never actually admitted to. Quick recap: Worthless stats are worthless. I don't put any weighting in worthless stats.

And I quote, "Do I select what evidence I present and what examples I give? Yes, I freely admit that." and "I also unashamedly admit to emphasizing certain stats and discounting others." and "No doubt some of the techniques could be improved upon, and some of the numbers I have generated have been impacted by factors such as small sample size, reporting bias, lurking variables, etc."

That, is manipulation.

Therefore I don't use them. That is not manipulating data, it is interpreting data based on its significance.

Semantics. And disingenuous.

I was talking about compelling evidence. I have never found compelling evidence that Brodeur is even close to one of the all-time greatest goalies. Nothing I have ever done, or seen anyone else ever do, has made me question my overall viewpoint. I have modified my viewpoint, as previously mentioned, based on Brodeur's performance. But a variety of different studies looking at a number of different traditional and non-traditional stats pretty much agree that Brodeur is a good goalie who played most of his career behind a great defence. The only area that remains to study is non-save goalie skills, and to try to assess Brodeur's impact there. Conceivably it could be found that there is a significant effect from those skills that upgrades his all-time status, but a lot more study is required.


What, precisely, makes a goalie the greatest of all time? It's an entirely subjective topic, and you can throw as many numbers out there as you want, it won't prove or disprove an opinion.

-e

Statman said...

I've been off this site for awhile... holyyyy sh*t, what on earth is wrong with this Anonymous/'e' person.... ??

"disingenuous" "manipulation"

"What, precisely, makes a goalie the greatest of all time? It's an entirely subjective topic, and you can throw as many numbers out there as you want, it won't prove or disprove an opinion."

Uh, I guess.... it's possible, with respect to any subject (not just hockey/sports) that someone can have a certain opinion that won't be swayed by any facts or numbers. Although at a certain point you'd have to question the sanity of this person.

Anonymous said...

Uh, I guess.... it's possible, with respect to any subject (not just hockey/sports)that someone can have a certain opinion that won't be swayed by any facts or numbers.

Not quite. Stats alone could definitively isolate the car with the best fuel mileage or a lawyer with the all-time best record of convicting criminals, regardless of previously held opinions.

Although at a certain point you'd have to question the sanity of this person.

As I would question the motives of one who utilizes ad hominem arguments.

-e

Anonymous said...

"The fewest shots against since 1993-94" does not mean that New Jersey led the league every season since 1993-94. It means just what it says, that over the period from 1993-94 to present they rank #1."

So are you talking about shots against, or Martin Brodeur's all time rank. I got confused when earlier you gave no weight to being among the best over a long period of time, and now it seems you are giving it some significance.

Statman said...

Fuel mileage - But have you considered different driving conditions, effect of tire type, effect of weather differences, effect driver personality/tendences, effect of relative quality/consistency of fuel, effect of car aging during period of study, effect of...

Anonymous said...

Fuel mileage - But have you considered different driving conditions, effect of tire type, effect of weather differences, effect driver personality/tendences, effect of relative quality/consistency of fuel, effect of car aging during period of study, effect of...

Fuel mileage is determined through a universal standard of measurement.

Such a thing is not possible in athletics, thus, the terms "best" or "greatest" are, ultimately, nothing but opinions, no matter how much "evidence" is thrown about.

Regardless, labeling one athlete or another "the best" isn't the discussion at hand.

-e

Statman said...

If there can be no "best" or "greatest", there can be no ranking.... & therefore it's entirely plausible to say that Red Light Racicot was a better goalie than Brodeur.

Anonymous said...

The best is a subjective term in regards to analyzing a goalie. If I had as much time on my hands as some who frequent this site do, I could probably sit down and compare Dan Cloutier to Dominic Hasek, and using the tactics displayed by the author of this site, I could probably sift through all the stats, find an area where Cloutier was better, then skew the information by finding a lot of adjusted information and numbers that support the stat I am trying to prove makes Cloutier better than Hasek. It is not hard to make outlandish claims much like the ones this site is founded upon. All you need is a ton of time. The people who have successfully called out all the bs on this site are the ones with enough time to sort through all the nonsense, and put it back into proper terms. Most however do not care enough to do that, or do not have the time to do that, and thus just laugh at the notion of even getting involved.

Statman said...

"I could probably sit down and compare Dan Cloutier to Dominic Hasek, and using the tactics displayed by the author of this site, I could probably sift through all the stats, find an area where Cloutier was better, then skew the information by finding a lot of adjusted information and numbers that support the stat I am trying to prove makes Cloutier better than Hasek."

I find that hard to believe. [Maybe Dan is better then Hasek when playing Tuesday games after a long weekend? How could we skew that into showing him better than Hasek.... ?] Nevertheless, I'm open to seeing evidence that ranks one goalie over another... & whether the preponderance of evidence enables a solid judgement as to who is better than who.

"The people who have successfully called out all the bs on this site are the ones with enough time to sort through all the nonsense, and put it back into proper terms."

I've read through some of the recent, very long, very picky/semantic arguments, & there doesn't seem to be much if any evidence provided that actually disproves what I've read on this site. E.g. Quoting the blog author as saying that he lists the evidence which proves his point, doesn't mean his point is wrong. (If I say Gretzky has the highest single season pts record, but then don't list every season by every player to ever play in the NHL & compare their points to his, doesn't mean I'm being misleading or disingenuous.)

Another example is the 1993-94 to present NJ shots allowed data... I can't believe how some have jumped all over that. "Popular opinion" is that NJ allows the least amount of shots, generally, over nearly 20 yrs... & of course much more importantly the facts show this (no, they didn't lead 15 yrs straight, which would be nearly impossible when compared to other teams' level of dominance in any one facet over a long period of time). If NJ has allowed more than 10% fewer shots per 60 min's than the league avg, over the past 15 yrs, & that is the best mark, then it's safe to say they have an excellent record in this regard. Just look at the percentiles.

Every study on here would be dozens of pages long if someone had to do that.

Whatever... I usually don't comment on here but I've been struck by the recent persistent criticisms (with little evidence provided along with the criticims), esp. that "Luongo is garbage!" & other juvenile "fan of player x" stuff. I doubt I'll reply further.

Anonymous said...

If there can be no "best" or "greatest", there can be no ranking.... & therefore it's entirely plausible to say that Red Light Racicot was a better goalie than Brodeur.

Exactly my point!

Quoting the blog author as saying that he lists the evidence which proves his point, doesn't mean his point is wrong.

It means it isn't right. Who's arguing semantics now?

Admitting to selectively choosing pertinent information to make a point is tantamount to admitting fraud.

Another example is the 1993-94 to present NJ shots allowed data... I can't believe how some have jumped all over that. "Popular opinion" is that NJ allows the least amount of shots, generally, over nearly 20 yrs... & of course much more importantly the facts show this (no, they didn't lead 15 yrs straight, which would be nearly impossible when compared to other teams' level of dominance in any one facet over a long period of time). If NJ has allowed more than 10% fewer shots per 60 min's than the league avg, over the past 15 yrs, & that is the best mark, then it's safe to say they have an excellent record in this regard. Just look at the percentiles.

That they allow fewer shots on goal per 60 minutes than the rest of the league over a period of time isn't the concern, but rather that it is in any way related to Brodeur's puck-handling and rebound control.

The fact that they've been far less efficient since the lockout suggests, in fact, that it is due more to their system and countless other variables, as the constant of goaltender (who, CG has conceded is playing better in recent years than he was during his "prime") remains unchanged.

Whatever... I usually don't comment on here but I've been struck by the recent persistent criticisms (with little evidence provided along with the criticims), esp. that "Luongo is garbage!" & other juvenile "fan of player x" stuff.

Surely you refer to another thread.

-e

Anonymous said...

"Another example is the 1993-94 to present NJ shots allowed data... I can't believe how some have jumped all over that. "Popular opinion" is that NJ allows the least amount of shots, generally, over nearly 20 yrs"

Popular opinion is also that Martin Brodeur and Patrick roy are the 2 greatest goalies ever. Funny how we credit popular opinion when it help us, and discredit it when it has no use in the point we are trying to make.

In regards to Luongo, he is terribly overrated, if anything. For all the "Luongo is great, and so underrated" garbage, why is it that his numbers, which were good on a bad Florida team for a year or two, have actually gone down since he's played in what is easily regarded as a solid defensive team, in Vancouver. How is it, that a case can be made for Luongo, when he has declined statistically while playing for a better team?

Statman said...

--> If there can be no "best" or "greatest", there can be no ranking.... & therefore it's entirely plausible to say that Red Light Racicot was a better goalie than Brodeur.

"Exactly my point! "

I guess there is no need to make any attempt to rank goalies, then. It's just too difficult!

--> Quoting the blog author as saying that he lists the evidence which proves his point, doesn't mean his point is wrong.

"It means it isn't right. Who's arguing semantics now?

Admitting to selectively choosing pertinent information to make a point is tantamount to admitting fraud."

That's certainly not true. Any study or article of any kind simply cannot list every single fact, statistic etc. If you read a basic news article, & they have omitted some information (whether completely relevant or not to the point at hand), it is not correct to label the article as false/fraud whatever. That's way too overblown. Odd, really.


As for Brodeur & shots against, I am skeptical as to whether he has a very big impact on shots against.

Statman said...

"In regards to Luongo, he is terribly overrated, if anything. For all the "Luongo is great, and so underrated" garbage, why is it that his numbers, which were good on a bad Florida team for a year or two, have actually gone down since he's played in what is easily regarded as a solid defensive team, in Vancouver. How is it, that a case can be made for Luongo, when he has declined statistically while playing for a better team?"

I don't know if your description is that accurate:

Luongo was definitely underrated, prior to coming to VCR.

Luongo's last 3 yrs with FLA:

.918, .931, .914

Luongo's 3 yrs with VAN:

.921, .917, .914 (this yr)

Luongo's performance dropped bigtime at the end of the yr last yr, perhaps due to his wife's condition in FLA.

As for this yr, injuries & whatnot. The season's not over.

http://www.hockey-reference.com/players/l/luongro01.html

Anonymous said...

Statman, if you are going to use excuses to justify why not so great performances and numbers should not be viewed as such, then we will be here all day.

Otherwise, the example used by one of the anonymous is perfect. I've followed this blog a while, and never really been able to put my finger on why all these numbers seemed off base with reality. Using the Dan Cloutier example puts things into better context, especially when you realize the difference between Brodeur and Hasek was minimal. Only after, as someone pointed out, the author takes all the numbers that favor Brodeur and dismisses them, says they have no value, then, picks the only stats that support Hasek as being the better goalie, then assesses that stat as having all value, and abra kadabra, there is the argument.

I am glad to see many of the readers start to pick up on the lies and half truths this site is filled with. I am enjoying seeing the small amount of credibility the author once had being taken away.

Statman said...

Well, maybe Brodeur's performance increased the past 3 yrs because his rel'ship with his brother's wife (or whatever it was) was resolved.

"...never really been able to put my finger on why all these numbers seemed off base with reality."

Well, it's so obvious there is a problem, you should be able to point it out. That's the basis of scientific examination.

I don't know what you're talking about, re: Cloutier, Brodeur & Hasek.

"Only after, as someone pointed out, the author takes all the numbers that favor Brodeur and dismisses them, says they have no value, then, picks the only stats that support Hasek as being the better goalie, then assesses that stat as having all value, and abra kadabra, there is the argument."

Really? Is that really what the author does? Provide an example.

I think this site is open to any & all reasonable, verifiable assertions. All you have to do is prove it. Let's see it. Where is the argument, other than Goalie X has lots of wins, has won some Cups, must be great?

Anonymous said...

"I think this site is open to any & all reasonable, verifiable assertions. All you have to do is prove it. Let's see it. Where is the argument, other than Goalie X has lots of wins, has won some Cups, must be great?"

No it isn't. After reading through a bunch of comment threads, there are a ton of points that are just completely ignored, and/or not answered.

Also, I have no idea where you get this idea that people just assume because a guy has a bunch of wins, and Stanley Cups, that he is great. However perhaps you could back up your assertions by providing a list of goalies with over 300 wins, a Vezina, and 2 or more Stanley Cups, and tell me how many of them do not qualify as great. I would think it is safe to say the majority are great. Many people however, especially with Brodeur, consider having many wins, the lowest GAA in NHL history, an above average save percentage, 3 Stanley Cups, 4 Vezinas, 5 Jennings, 1 Olympic Gold, multiple world cup medals, and 98 shutouts to signify greatness. You however, seem to think that everything there is the result of luck. Talk about not having an argument.

Anonymous said...

HAHAHAHAHA LOOKS LIKE THE STAT GEEKS AND SAVE PERCENTAGE IS THE ONLY STAT THAT COUNTS JOKERS CANT BACK UP THEIR BOGUS CLAIMS

Anonymous said...

THE ONLY FRAUD HERE IS THE CONTRARIAN GOALTENDER AND ALL OF HIS "STUDIES"

Statman said...

Anonymous of 8:06 pm... are you same Anonymous that posted at 8:23 & 8:32?

Statman said...

Lowest GAA in NHL history?

Is that raw GAA, or GAA adjusted for era? Because the goal avg has varied dramatically over the decades. E.g. 4.0125 per game in 1981-82; down to 2.5-2.7ish by the late 90's & 2000's.

That's a huge swing.

Anonymous said...

listen stat boy, enough with your imaginary "adjusted" numbers hahaha. we re going by the record books, not geek land manipulated numbers. and its pretty funny how out of all the credentials mentioned for brodeur all you can do is "adjust" one of them.

either way, way to provide that list of not so great goalies who have a lot of wins, cups and vezinas. cant put your money were your mouth is? didnt think so

Statman said...

Sure, "by the record books" aka basic counting numbers, Brodeur's numbers are at or near the top.

Sorry, I haven't taken a look at the 300+ win list, but if these guys are on it, they are overrated (beneficiaries of playing on strong teams) & not "great":

Fuhr, Vernon,

Anonymous said...

vernon won a vezina? and many people consider fuhr great, its funny however you happen to rip off 2 names from the contrarian frauds overrated list.

also, can you tell me who is the "adjusted" home run king? or the "adjusted" td leader in football is? god your argument is weak

Anonymous said...

actually i should "adjust" what i just said. contrarian frauds argument is weak. you have yet to present an argument of your own.

Statman said...

So if a goalie wins 300+ games, has 2+ Cups, but doesn't win a Vezina, then he isn't considered "great" by some people? (I could care less, actually) As I said, I didn't go consult the list of goalies that are "great" according to you/popular opinion (apparently must meet 300+ wins etc.).

Yes, many people consider Fuhr "great". I know that. I really don't care what the popular opinion is. The Oilers of the 80's were my favourite team. But I don't think Fuhr is truly "great". He was lucky to play an acrobatic, crowd-pleasing flamboyant style in front of a very high-scoring team that bailed him out.

[By the way, in all these discussions of "great"... should this be the top 1%? Top 5%? 10%? E.g. with regard to forwards, are just Gretzky, Lemieux and maybe Howe "great"? Or do we include "lesser" greats like Messier, Lafleur, Beliveau...?]

Actually, there are some interesting, extremely detailed studies on baseball stats... including "adjusted" (for era, for park dimension, among other things) home runs? Do a web search, there are whole books on the subject. I've got a few downstairs, written by a PhD in mathematics. But it sounds like you wouldn't like it/too tedious for you. That's fine, enjoy sports however you want to.

There doesn't appear to be much use in responding to your child-like freak-outs. Even a kid in Grade 1 could go consult a record book & therefore pronounce someone as "great" because he has X Cups, X Trophies, X Wins etc. If that's what you believe, fine. I'm not sure why you continue to post, though. All you do is quickly resort to angry name-calling. I suspect you're the same Anonymous that has been posting on here for months, leaving short, angry rants (with zero analysis). I guess you're lucky you're just an "anonymous" online somewhere, & not having to back anything up in person.

Statman said...

Feel free to spread your wisdome... Here are some other links... you can go there & rant about how their use of stats is "garbage!" etc etc etc:

http://hockeyanalytics.com/Links.htm

I recommend starting with MC Hockey.

Anonymous said...

so essentially you could not provide evidence to support your claim so you took the "sentimental" approach. its funny how the argument chances when you run out of real data

and again, you resort to pretending to be offended by my namecalling (apparently fraud is too much for you to handle), yet you openly support the author of the site who's site is founded on name calling in the context of "brodeur is a fraud"? again, its interesting to see how the points of focus change dependent upon how much evidence you have on your side

Statman said...

"so you took the "sentimental" approach"

I'm not following you. I said I liked the Oilers, but that even so, I did not & do not consider Fuhr to be "great".

As for namecalling, if you're the same Anonymous who has been posting for a few months (same writing style, grammar), you frequently use more than just "fraud" as a derogatory device.

If you've been following this blog for a few yrs, you'd know that the blog author uses "fraud" tongue-in-cheek (he actually considers him to be above avg over his career, & certainly one of the best over the past 3 yrs - I agree with him)... & it helps to be a bit sensationalist that way to attract attention.

Although getting angry mini-rants is probably not the kind of attention he is seeking.

Are you by chance related to Brodeur? Is he one of your faves? Are the NJD one of your faves? Because it sure seems like you are personally offended, which is odd. It's just sports.

Anonymous said...

i have no relation to brodeur nor am i a devils fan. i have an issue with misleading information, something this blog is filled with, not only in regards to brodeur, but even more subtly with roy. its like anybody who has had success is torn down in order to promote those that didnt or those that have but to a much lesser degree.

the other issue i have a problem with is assuming that just because another player, like a backup, has success, it must mean that the starter isnt good. god forbid credit is given to both guys.

Statman said...

I've yet to see any misleading information posted on here by the author. As far as I can recollect, he prefaces everything with the appropriate amt of caution.

Sure, usually the backup isn't as good as the starter. But, a backup can be good... perhaps even better than the starter. At one pt Brodeur was a backup, as was Hasek.

Raw stats like wins, GAA, & Cups are heavily dependant on the overall team strength, in my view. Some people might disagree with that. Those that don't disagree are trying to isolate what it is that a goalie does (independant of team strength/weakness) to show his skill & quality.

A goalie could play 60 games for a weak team & end up with a GAA of 3.30 & 25 wins this year.... or, if he's on a stronger team, could (in the same year -- because year to year the avg GAA & SV% changes) have a GAA of 2.40 & 35-40 wins. It's the same goalie, just different conditions. If you just look at very basic stats, the goalie on the weaker team appears much worse. Over a career, under 1 condition the goalie looks very avg or bad, & under the other condition he's very good/great.

[Baseball - check out Michael Schell's books for adjusted home runs, adjusted avg etc.]

The Contrarian Goaltender said...

I am glad to see many of the readers start to pick up on the lies and half truths this site is filled with. I am enjoying seeing the small amount of credibility the author once had being taken away.

Scott Clemmensen is still 25-12-1, 2.35, .918, is he not? And unless I'm mistaken, Martin Brodeur is still third on his own team in save percentage. Please provide a link to any mainstream media article or post that predicted the New Jersey Devils would be just fine without Martin Brodeur. Got nothing?

I find it to be an interesting interpretation of observed results to imply that I am somehow losing credibility.

The Contrarian Goaltender said...

Your choices of words and presentation of facts is deliberately misleading.

Your lack of reading comprehension allows you to be easily misled by your dogged persistence to find some illogical fault with what I write. Over the last 15 years, the New Jersey Devils have allowed fewer shots per game on average than any other team in the league, and have outperformed league average by over 10%. That is what I wrote, and that is what happened, as I proved to you with hard evidence. Do you disagree they were the best in the league over that time span? Do you still disagree that they were excellent? I really don't see why you still apparently see a problem here. And I simply can't imagine that any reasonable person who happened upon this site would read your posts in this thread and think, "You know, that E has a point there about New Jersey's shots against."

You've linked to your own articles, which in turn link to your other articles, or else linked to a site that one could potentially garner such information from, but not any actual neutrally compiled stats.

You seem to think that there are people out there who do nothing all day but compile stats, on the off-chance that you might want to see those numbers. The world doesn't work that way, you have to do your own work. You asked earlier why another commenter mentioned me "holding your hand" - well, asking me to link to compiled statistical data that you could easily find and generate yourself certainly qualifies as that.

I don't have to "prove" your information incorrect; you've already admitted, "Do I select what evidence I present and what examples I give? Yes, I freely admit that. So does every single other blogger, author, journalist and writer, because there is limited word space and we all have limited attention spans."

That has nothing to do with the information. That has to do with the blogging medium. Go ask a journalist if they choose which evidence they include in their articles. When they say yes, tell them that proves all of their information is incorrect. See if they are pleased or displeased with that insinuation.

I dismissed it because A) it was irrelevant to the conversation, B) you freely adjust stats to get the info you desire C) you admit the base information is flawed but very scientifically speculate that "I don't think it would have much of an impact over the course of a career" and D) follow up with a disclaimer that it doesn't take into account team defense, which may or may not entirely skew the chart.

I included a disclaimer? I thought I was nothing but misleading and disingenuous. And you keep talking about how I "freely adjust" stats, without ever directly addressing any of the adjustments. Do you disagree with era adjustments? Opponent adjustments? Team adjustments? Do those things not affect goaltending results? And do you really think I was somehow scheming to give John Vanbiesbrouck the attention he so obviously deserves, and rigged it so he would finish first?

Which doesn't answer my original question, how were you able to quote these numbers if they weren't compiled yet?

Because they were compiled already. I have lots of different spreadsheets with lots of numbers in them. I did not consider the shots against information to be particularly interesting, as the rankings were fairly obvious to well-informed hockey fans. That's why I didn't post it until you asked for it.

And I quote, "Do I select what evidence I present and what examples I give? Yes, I freely admit that.

Sure. Because I'm writing blog posts, not novels. Do you go on right-wing political sites and berate them after every single post for not addressing every possible liberal counterargument to every one of their arguments?

I don't falsify data. I gather data, process it, analyze it, and present the results. Some results are more interesting and relevant than others, and so they get written about while others don't. Just like people spend hours in the library researching and then include only a fraction of the material in their final book or paper. I can't understand your problem with a very simple, universal truth about the writing process.

Citing two players that people might (emphasis on might, since this is pure speculation on your part) subjectively call "great" based purely on either stats or having scouted them doesn't prove your point that you can accurately judge a player purely on statistics.

If statistics were completely irrelevant, it seems highly unlikely that an experienced scout and an amateur number-cruncher would come to the same conclusion, does it not? And are you seriously implying that Gretzky and Lemieux aren't great? There is no subjectivity involved, that is a simple fact. Only someone with a complete lack of knowledge and/or understanding of hockey would even consider the possibility that those two weren't great players.

That, is manipulation.

No, it is filtering out the chaff and focusing on the wheat. Again: Worthless stats are worthless. I don't put any weighting in worthless stats.

Like, oh, say, Clemmensen?

No, the Devils would not give up fewer shots with Clemmensen in net than with Brodeur. We have hard data that proves the direct opposite. The Devils have given up 28.8 shots per 60 minutes with Clemmensen in the net, compared to 25.8 with Brodeur.

What exactly is your position on Brodeur's shot prevention impact? Is it that Brodeur has zero effect on the shots taken by the opposing team? If so, I would like to hear how you would explain the 3 shot per game discrepancy between Brodeur and Clemmensen, and the fact that Brodeur has faced 0.7 shots per 60 minutes fewer than his backups over the course of his career.

What, precisely, makes a goalie the greatest of all time? It's an entirely subjective topic, and you can throw as many numbers out there as you want, it won't prove or disprove an opinion.

Sure it will, or at least it should. Opinions can be and often are wrong. If you think Andre "Red Light" Racicot was better than Martin Brodeur, or even possibly could have been better, then you are either unintelligent or ignorant about hockey.

The Contrarian Goaltender said...

However perhaps you could back up your assertions by providing a list of goalies with over 300 wins, a Vezina, and 2 or more Stanley Cups, and tell me how many of them do not qualify as great.

Tom Barrasso says hi. Rogie Vachon still isn't in the Hall of Fame, last I checked. Grant Fuhr is a pretty clear case of somebody who happened to be in the right place at the right time. Billy Smith as well, unless you rate goalies completely based on playoff results. Gump Worsley was a fine goalie, but doesn't usually pop up on any all-time greatest lists. Roy, Sawchuk, Plante, Hasek are pretty much considered great by the majority of observers, and I'd include Turk Broda as well. Based on that accounting, the odds that somebody who meets your criteria is great is 50/50.

Many people however, especially with Brodeur, consider having many wins, the lowest GAA in NHL history, an above average save percentage, 3 Stanley Cups, 4 Vezinas, 5 Jennings, 1 Olympic Gold, multiple world cup medals, and 98 shutouts to signify greatness. You however, seem to think that everything there is the result of luck. Talk about not having an argument.

Many other people consider wins, GAA, Cups and international medals to be indicators of team success rather than goalie skill, and prefer to analyze goalie play within the context of the team rather than merely looking at unadjusted numbers. There is a lot of work done on this blog that addresses all of these points, look it up.

Anonymous said...

Your lack of reading comprehension allows you to be easily misled by your dogged persistence to find some illogical fault with what I write. Over the last 15 years, the New Jersey Devils have allowed fewer shots per game on average than any other team in the league, and have outperformed league average by over 10%. That is what I wrote, and that is what happened, as I proved to you with hard evidence. Do you disagree they were the best in the league over that time span? Do you still disagree that they were excellent? I really don't see why you still apparently see a problem here. And I simply can't imagine that any reasonable person who happened upon this site would read your posts in this thread and think, "You know, that E has a point there about New Jersey's shots against."

Speaking of reading comprehension issues, this whole paragraph is in response to, "Your choices of words and presentation of facts is deliberately misleading."

Does that sound like I'm questioning the data itself?

You seem to think that there are people out there who do nothing all day but compile stats, on the off-chance that you might want to see those numbers. The world doesn't work that way, you have to do your own work. You asked earlier why another commenter mentioned me "holding your hand" - well, asking me to link to compiled statistical data that you could easily find and generate yourself certainly qualifies as that.

You quoted it, I asked for the reference. If you'd have said it was your study and you hadn't published it, that would have been that, but you started passing out bad links and didn't admit that the "recent study" was actually your own statistical compilation.

Again, that you quoted it leads one to believe it exists. Faulting me for somehow being wrong in requesting your sources or for not doing my own legwork is where the disingenuity comes in.

That has nothing to do with the information.

That has everything to do with the information when you arbitrarily decide which parts of it you want to use.

That has to do with the blogging medium.

That's just a lame excuse.

Go ask a journalist if they choose which evidence they include in their articles. When they say yes, tell them that proves all of their information is incorrect. See if they are pleased or displeased with that insinuation.

Slight difference: a journalist has objectivity and fact checkers.

Bloggers (who are not journalists by any stretch, but nice try) have neither.

I included a disclaimer? I thought I was nothing but misleading and disingenuous.

Which I stand by when you cite that article as evidence of a further claim, knowing that the information is probably not entirely correct to begin with. That's the disingenuous part.

And you keep talking about how I "freely adjust" stats, without ever directly addressing any of the adjustments. Do you disagree with era adjustments? Opponent adjustments? Team adjustments? Do those things not affect goaltending results?

I disagree with manipulation of figures to get desired results.

I don't falsify data. I gather data, process it, analyze it, and present the results. Some results are more interesting and relevant than others, and so they get written about while others don't. Just like people spend hours in the library researching and then include only a fraction of the material in their final book or paper.

See above concerning actual journalism vs. blogging.

And, by the way, the book would have an entire section dedicated to citing their sources.

I can't understand your problem with a very simple, universal truth about the writing process.

As I've said, ignorance and those that prey upon it are my enemies.

If statistics were completely irrelevant, it seems highly unlikely that an experienced scout and an amateur number-cruncher would come to the same conclusion, does it not?

How do you continue to get out of my argument that statistics don't tell the whole story to mean that I'm arguing they are completely irrelevant?

And are you seriously implying that Gretzky and Lemieux aren't great?

I'm arguing that "great" is subjective.

There is no subjectivity involved, that is a simple fact.

That is simple opinion. Which is the exact opposite of fact. That there are many who hold that opinion doesn't make it a fact, it just makes it common perception (see also: religion).

Only someone with a complete lack of knowledge and/or understanding of hockey would even consider the possibility that those two weren't great players.

Of course, as per usual, you neglect to list option C, which would be, "Only someone who understood the difference between opinion and fact would consider the discussion of greatness pointless."

No, it is filtering out the chaff and focusing on the wheat. Again: Worthless stats are worthless. I don't put any weighting in worthless stats.

The fact that you, an admittedly biased person, decide on the worth of data and its being discarded or included nullifies any and all conclusions you reach based upon that information.

What exactly is your position on Brodeur's shot prevention impact? Is it that Brodeur has zero effect on the shots taken by the opposing team?

Isn't this where we started? It's unquantifiable. Man, talk about hand holding.

If so, I would like to hear how you would explain the 3 shot per game discrepancy between Brodeur and Clemmensen, and the fact that Brodeur has faced 0.7 shots per 60 minutes fewer than his backups over the course of his career.

Simple, the entire team plays differently when he's in there. I thought we covered this much earlier also...

Sure it will, or at least it should. Opinions can be and often are wrong.

Belief is funny like that.

If you think Andre "Red Light" Racicot was better than Martin Brodeur, or even possibly could have been better, then you are either unintelligent or ignorant about hockey.

Or, again, option C, "think the argument a waste of time until everyone agrees on what defines 'great'". Until it is, someone somewhere could dig for the right stats and argue anything they wanted, quite possibly somewhat convincingly.

I'd offer a link as evidence, but I think you already know the url of your own site.

-e

Statman said...

"I think the argument a waste of time until everyone agrees on what defines 'great'"

Really? Everyone must agree on the definition?

That will never happen. So why even bother posting on this blog? Aren't your comments themselves a form of argument? But you say that arguing is a waste of time?

Statman said...

"In regards to Luongo, he is terribly overrated, if anything. For all the "Luongo is great, and so underrated" garbage, why is it that his numbers, which were good on a bad Florida team for a year or two, have actually gone down since he's played in what is easily regarded as a solid defensive team, in Vancouver. How is it, that a case can be made for Luongo, when he has declined statistically while playing for a better team?"

His numbers have gone down in VAN?

http://www.hockey-reference.com/players/l/luongro01.html

GAA (02/03, 03/04... 08/09)
Last 3 yrs in FLA - 2.71, 2.43, 2.97
Last 3 yrs in VAN - 2.29, 2.38, 2.59 (this yr)

Wins (02/03, 03/04... 08/09)
Last 3 yrs in FLA - 20, 25, 35
Last 3 yrs in VAN - 47, 35, 18(this yr)

PTS% (02/03, 03/04... 08/09)
Last 3 yrs in FLA - .385, .444, .534
Last 3 yrs in VAN - .667, .541, .683 (this yr)

If you look at the basic stats (GAA, Wins, Pts%), which you seem to rely upon in order to rank goalies, it appears that Luongo is much better with VAN than with FLA, which is contrary to your assertion.

Of course, I don't think that he's improved with VAN, & actually his last yr with FLA was probably his best. But, I'd have to use another stat to show this, & it's a stat you absolutely hate.

Statman said...

Correction, I think Luongo's second-last season with FLA was probably his best.

Anonymous said...

Really? Everyone must agree on the definition?

I would call that the very definition of the word definition.

That will never happen.

You're arguing my point for me.

So why even bother posting on this blog?

It certainly wasn't to argue which player or players are or aren't "great." I thought I made that pretty clear.

Here, for the third time, I'll state my reason for posting: because ignorance and those that prey upon it are my enemies.

Aren't your comments themselves a form of argument? But you say that arguing is a waste of time?

I do? If I did, I was certainly mistaken or misunderstood, but please direct me to where I said that arguing was a waste of time.

-e

Statman said...
This comment has been removed by the author.
Statman said...

"please direct me to where I said that arguing was a waste of time."

2nd last para. of your previous post.


"Here, for the third time, I'll state my reason for posting: because ignorance and those that prey upon it are my enemies."

Your valiant efforts are probably wasted here on this little blog... think of the millions of other blogs, news sites, publications, governments, town-hall meetings, private conversations, etc etc etc you can save. :)

Anonymous said...

2nd last para. of your previous post.

A misquote on your part. I said: "Or, again, option C, "think the argument a waste of time until everyone agrees on what defines 'great'"."

Pertaining to that one specific argument, not all arguing.

But, I'm guessing you already knew that.

Your valiant efforts are probably wasted here on this little blog... think of the millions of other blogs, news sites, publications, governments, town-hall meetings, private conversations, etc etc etc you can save. :)

Patronizing snark. That'll discredit me.

I think it fairly obvious that you have little to offer but weak attempts at defamation and probably should have followed your original instincts and not replied again.

-e

Bruce said...

What exactly is your position on Brodeur's shot prevention impact? Is it that Brodeur has zero effect on the shots taken by the opposing team? If so, I would like to hear how you would explain the 3 shot per game discrepancy between Brodeur and Clemmensen,

Glad to see you're coming around, CG. It was worth reading the hundred screens above just for that.

Statman said...

Ha ha.... whatever, "e".

Statman said...

Bruce - nothing could be worth reading through that! haha :)

Anonymous said...

Your persistence to be misleading is incredible. You refuse to acknowledge Luongo's poor performance in a sample of 10 games, however, you misleadingly state earlier "Martin Brodeur is 3rd on his team in save percentage" as if it is some sort of indication of performance. Again, one guy playing 10 game (Luongo) means nothing, you have yet to say his backups have outplayed him since his return, however, with Brodeur, 10 games is enough to make statements about his performance?

Further the save percentage nonsense continues, as a goalies job is not to stop shots, his job is to prevent goals. Straight up, if there are 2 goalies, one who is guaranteed to finish the year with a gaa of 1.50, and have a save percentage of .90, I will take that every day of the week over a guy with a save percentage of .95 and a higher gaa. Further discrediting your claim that a goalies job is to "stop shots", I guess you could say a goalie has no right to intercept a pass across the crease then because his job is not to prevent shots, hell, instead he should wait for the shooter to get the puck, so he can attempt to stop the shot and effectively do what you claim to be is his job?

And further you fail to acknowledge any point that proves the hypocrisy of your arguments, and I mentioned earlier. You detest "E"'s claim that New Jersey was not "the best" at not allowing shots over that 15 year period. Instead you use the cumulative results, of which I am assuming NJ was routinely top 10 in fewest shots, and say that doing this, makes them "the best" or "great", as the consistency over those times is what separates them from everyones else. Yet when presented with the EXACT same claim for Martin Brodeur, you for some reason change your mind. Hypocrite? I think so.

Anonymous said...

Further discrediting your argument regarding save percentage only, IF there is ANY correlation between rebound control, and/or puckhandling, and the number of shots allowed per game, something almost anyone will acknowledge, then you must now take gaa into consideration as it is the only statistic which may be reflective of a goalie who excels at those things.

You seem to be blind to the effects of a goalie who can play the puck as well as the likes of Brodeur, Turco, or Dipietro, as I remember pretty often how people would jokingly refer to New Jersey's trap system as a 1-2-3. the 3 obviously accounting for the 2 defenseman, and then the 3rd, Brodeur.

Statman said...

"Straight up, if there are 2 goalies, one who is guaranteed to finish the year with a gaa of 1.50, and have a save percentage of .90, I will take that every day of the week over a guy with a save percentage of .95 and a higher gaa."

I guess the goalies, in your example, have the ability to influence shots against. Is that your implication?

I suggest that goalies -- well, good ones, certainly those who are paid to play -- make every attempt to save every shot, no matter what the score. I doubt that goalies' main thoughts are how many goals they've allowed (as if they feel they can let in shots because they don't need to make the save). I suspect they are much more concerned about saving each shot.

If you trade the goalies one-for-one (switch their teams), what happens to the .95 goalie's GAA, & what happens to the .90 goalie's GAA?

If you had a chance to trade Goalie A, who has a .90 sv% & a 1.50 GAA, for Goalie B, who has a .95 sv% & a 1.90 GAA -- would you, assuming their games played, salary etc aren't a concern? The math obviously shows that Goalie B faced more shots than Goalie A.

I suspect even the most traditional, non-stat-oriented GM would make that trade.

Anonymous said...

Statman, in the example I used, I an obviously speaking about using extremes, with goalie A being great at preventing shots, and goalie B just stopping them.

In regards to the "playing to the score" example you used, there is no distinct associated either way, as obviously no goalie purposely gives up goals, however teams, and goalies do tend to lose focus in blowouts, which would naturally lead to more goals scored.

"If you trade the goalies one-for-one (switch their teams), what happens to the .95 goalie's GAA, & what happens to the .90 goalie's GAA?"

Again, answering this question requires making assumptions, however you could conclude that goalie A's puckhandling and rebound control was a major component of the system in which his team relied upon, and with goalie B now taking his spot, and not possessing those skills, the system breaks down and thus both his save percentage of .95 and gaa would deteriorate.

Meanwhile, being a major piece of his teams system, goalie A now brings those skills to the team that allows a lot of shots, and naturally their shots allowed decreases. Perhaps they decide to adopt the same system goalie A's team used, and the results are similar.

Again, all of these scenarios involve making assumptions, and nothing is concrete, although it is fairly obvious the preventing goals is far more important than stopping shots, because essentially preventing shots is stopping the shot before it happens, something in which there is not stat for.

Statman said...

"You refuse to acknowledge Luongo's poor performance in a sample of 10 games, however, you misleadingly state earlier "Martin Brodeur is 3rd on his team in save percentage" as if it is some sort of indication of performance. Again, one guy playing 10 game (Luongo) means nothing, you have yet to say his backups have outplayed him since his return, however, with Brodeur, 10 games is enough to make statements about his performance?"

Are you talking to me? I really can't tell anymore. You didn't address any of my most recent points, but I think you are talking to me. (Hint - get a "name" on here, it's easy.)

Yes, I realize that over the first 9 or 10 games since he came back from injury, Luongo's stats apparently are no better or may be worse than his backups. I really don't care... I'm not a Luongo or VAN fan. I'm only really aware of his season-long totals & his career totals. Same with Brodeur. I had no idea that showing that Brodeur's stats were worse than his backups would send you into a rage.

Above I listed Luongo's stats for each yr, for the past 6 yrs. GAA, Wins, Pts %. I didn't even include SV%, which is actually the stat that corresponds with YOUR statements about how he has played over the past 6 yrs. How on earth does that equate to "Your persistence to be misleading is incredible." ?? Bizarre!

You're going on & on about "10 games"... ok, whatever, 10 games isn't much. If Brodeur has only played 10 games, it's not a large enough sample to make a conclusion about his play, really. Luongo has actually played 30 games so far this yr, & those are the 30 games worth of 2008-09 stats I referred to above.

"Yet when presented with the EXACT same claim for Martin Brodeur, you for some reason change your mind. Hypocrite? I think so."

Yes, cumulatively Brodeur over the past 10-15 yrs has been at the top of some stats (Wins), near the top (GAA, shutouts), and above avg (sv%). If you think Wins are the most important stat, then he is "at the top". I suggest that those who look deeper into what determines a goalie's quality, would choose stats other than Wins, or even GAA or shutouts.

It's a difference of opinion. You can pick whatever criteria you feel determines who is the best. No one here is denying the basic "counting numbers" that Brodeur has accumulated.

Your accusations of hypocrisy, misleading etc. simply have to stop. They aren't accurate & just tends to make you look juvenile & petulant. I don't even know why you bother to comment. We all know what the basic stats (GAA, Wins, Shutouts) say.

Statman said...

"...with goalie A being great at preventing shots, and goalie B just stopping them. "

That wasn't made clear in your example.

If Goalie A was "great" at preventing shots, I (as GM) would want to have a good estimate of this prevention, apply it to how many shots my team currently allows, apply Goalie A's .90 sv% to this now-lower amt of shots, estimate the GAA, etc. And compare goalies that way.

Sure, if a goalie can prevent shots, & this can be reasonably documented, then that is something to consider.

But if you're just going to say ".90, 1.50 is better than .95 regardless of his gaa", then I'd strongly disagree.

Anonymous said...

Statman, when I address you as such, that is when I am talking to you. Almost every thing you just addressed, is directly towards the "Contrarian Fraud" as it was put. He is the one who made the misleading assertion that Brodeur's performance was "third" amongst goalie on the Devils, yet fails to do the same with Luongo, which is were the double standard, and thus bias and misleading information claims are stemmed from. Where you assume there is any "anger" or "rage" is beyond me. As said before, I expect the same measuring stick to be used for everyone, however the authors claims seem to be obsessively focused on Brodeur.

Statman said...

Ok, you're talking to CG not me.

How many other starters have played fewer games due to injury? There might be more than just Luongo & Brodeur. I don't know why you're jumping all over the author for not analyzing numerous players immediately. I assume he has a day job!

The "rage" comment is due to your various negative comments. There are basically 2 types of commenters here - the majority stop in now & then to discuss stats, assessing players etc., & a few who drop in now & then and scream "GARBAGE!!!" & but don't really provide proof other than refer to the NHL Record Book.

Anonymous said...

"But if you're just going to say ".90, 1.50 is better than .95 regardless of his gaa", then I'd strongly disagree"

No, but what I am saying is that it is certainly possible for a goalie with an average save percentage, let alone an above average one, to be a better goalie than the guy who led, or came close to leading the league in save percentage, which there disproves or invalidates the majority of the claims made on this blog.

Anonymous said...

"How many other starters have played fewer games due to injury? There might be more than just Luongo & Brodeur. I don't know why you're jumping all over the author for not analyzing numerous players immediately. I assume he has a day job!"

This is the point, these misleading claims are handpicked, and not assessed all the time, only when they support the point CG is trying to make. It has also been pointed out many times, by multiple people, how CG either ignores, or fails to address some valid criticisms or points that may disprove his theories, or call him out on his obvious bias.

Statman said...

Actually, I recall that there are some studies/discussion on here regarding shot prevention, probably dating back to last summer.

[I'm not going to do a search, but I recall something like a possible reduction of up to 2 shots per 60 mins. You'd have to apply it to shots/60, sv% etc. like I discussed. E.g. Who cares if a goalie prevents 10 shots per game, if his sv% is only 50%?]

So, the issue has been discussed, actually. Even if it hadn't, that wouldn't "disprove[s] or invalidate[s] the majority of the claims made on this blog". That's a huge overstatement & just not correct.

You need to dig around on here for awhile, read the stuff. That way, you won't be shocked & appalled if you just start reading recently. This is just a blog... none of us are getting paid to do this, we can't spend 50 hrs a week compiling research, publishing 100 page studies, etc. Those of us who have been on here for a couple of yrs tend to recall the various studies/discussion, & therefore don't have to question everything, make accusations etc. We remember that a lot of these facets (e.g. shot prevention) have been discussed in the past.

There are also a lot of other good blogs, start with hockeyanalytics.com, & check the Links. That's how I found this site.

Statman said...

"This is the point, these misleading claims are handpicked"

I'm not so sure about that... the blog has "Brodeur" in the title, so I assume he's going to bring him up from time to time!

There is nothing incorrect in saying that Brodeur has weaker stats this yr than his backups, albeit 10 games (which I think the author mentioned).

To go crazy on him for not picking a 10 game period of Luongo's 30-game season is a bit ridiculous. Why not pick 10 of the worst games this season of any goalie, then? Then most of them would look worse than Brodeur, or any goalie. The fact that Clemmenson's stats are quite good is of course going to be brought up. Luongo's complete 08-09 season is better than his backups, clearly.

"It has also been pointed out many times, by multiple people, how CG either ignores, or fails to address some valid criticisms or points that may disprove his theories, or call him out on his obvious bias."

Well, it appears there are only 1-2 people who are (unusually) virulent in their criticism, actually. Not "multiple" people. This "e" guy, you, & some other guy who merely types "GARBAGE!!" & other juvenile stuff.

"E" types on & on & on & on... but he doesn't really say anything.

However weak his pt is, or how small a dent he has made in the other's argument, he maintains the fact that he HAS a pt/made a very small dent negates the other person's entire argument.

Anonymous said...

"[I'm not going to do a search, but I recall something like a possible reduction of up to 2 shots per 60 mins. You'd have to apply it to shots/60, sv% etc. like I discussed. E.g. Who cares if a goalie prevents 10 shots per game, if his sv% is only 50%?]"


Obviously the 50% save percentage is not ever going to be compensated by any skill, however using extremes, if a goalie only stops 80% but has an unheard of talent in regards to preventing shots accounting for 10 fewer shots a game, it would still be possible for him to be a better goalie than a guy who stops 90%. If lets say he faces 15 shots a game,and thus 3 goals are scored against him, but his rebound control and specifically his puckhandling are so good that his is able to create offensive chances because of it, such as Brodeur or Turco have done, but to an even greater extent, then he is again changing the game and exponentially more valuable then a guy who just sits back and stops 90% of the shots. Going further with the obvious exaggeration, assume his skills directly or indirectly create 1 goal a game of extra offense for his team. Thus take this goalie and put him on a team that on average allows 25 shots a game, and scores 2.75 goals a game. Reducing the shots a game to 15, would result in a gaa of 3, however his adding contributions in puckhandling bump the goals for up to 3.75, resulting in more wins. Contrarily, a goalie with no shot prevention skills goes the the team and instead of facing 25 shots, he then faces 30, letting in 30. Without the added offense created, this team now gives up 3 per game and still average 2.75 for. This example as said is an obvious theoretical extreme, however thinking about it a similar situation has occurred in football many times, specifically with Michael Vick, in which a quarterback with an unorthodox but very effective skill set was able to be a very good and effective player without having the typical "good qb" stats.

Anonymous said...

To give a better example of the direct or indirect way in which this puckhandling extraordinaire could create offense, lets say his team knows his is superb at cutting off dump ins, and thus does not have to back check nearly as much, the goalie is effectively able to outlet the puck, 1) often creating odd man rushes, 2) resulting is forwards being less fatigued, and 3) not allowing the other team to change line as effectively resulting in them fatiguing.
All of these things are very possible effects of a goalie with good puckhandling or rebound control, and should be taken into effect, not just as "oh we need to quantify them or they do not count" but rather as talents the have a positive team effect.

Anonymous said...

@CG
"Many other people consider wins, GAA, Cups and international medals to be indicators of team success rather than goalie skill, and prefer to analyze goalie play within the context of the team rather than merely looking at unadjusted numbers. There is a lot of work done on this blog that addresses all of these points, look it up."

However, you seem to just attribute success to luck, which is rather silly, especially in terms of international play in which those that are members of the team are individually selected based on their value to the team.

Anonymous said...

WELL WELL WELL, MORE CONTRARIAN BULLSHIT, MORE UNANSWERED AND IGNORED PRO BRODEUR ARGUMENTS, LESS CREDIBILITY FOR THE CONTRARIAN FRAUD.

The Contrarian Goaltender said...

You refuse to acknowledge Luongo's poor performance in a sample of 10 games

Wait, what? Direct quote from a previous comment: Luongo has been poor since he came back (.884, 3.50), I definitely agree.

Luongo has played poorly over the last 10 games. He also played poorly in the last 10 games of the 2007-08 season. There have been a number of other stretches during his career when he has played poorly over a 10 game span. Goaltending results are variable, goalies run hot and cold. That is why you need to look at larger sample sizes to evaluate them.

misleadingly state earlier "Martin Brodeur is 3rd on his team in save percentage" as if it is some sort of indication of performance.

See, I'm confused. Do you disagree that Martin Brodeur is 3rd on his team in save percentage? It seems that every time I post a simple fact, somebody is claiming that I am being misleading or biased. I guess it is impossible to expect that people aren't going to simply see what they want to see.

By the way, Luongo is still first on his team in save percentage. If he was third, then I would have pointed that out as well.

as a goalies job is not to stop shots, his job is to prevent goals.

Absolutely, a goalie's job to prevent goals against. The best way to measure his effectiveness at doing so is through save percentage, because the goal prevention method that far overwhelms all others in importance is stopping shots.

Yet when presented with the EXACT same claim for Martin Brodeur, you for some reason change your mind. Hypocrite? I think so.

No, I'm not a hypocrite, because my reasons for rejecting that claim are not because I don't think cumulative performance matters, but because I disagree that Brodeur's cumulative performance surpasses his peers. I look at more than wins and shutouts when I evaluate goalies - I take into account save percentage, team-adjusted and shot quality adjusted stats, and they do not tell me that Brodeur is more consistent or that his performance over his career has been better than everyone else. If Martin Brodeur was legitimately a top 5 goalie for the last 15 years in a row, then you could make a very good case for him as one of the best ever. I don't believe that to be the case, based on my observations and detailed study of the numerical record.

No, but what I am saying is that it is certainly possible for a goalie with an average save percentage, let alone an above average one, to be a better goalie than the guy who led, or came close to leading the league in save percentage, which there disproves or invalidates the majority of the claims made on this blog.

If you are arguing that is possible because of puckhandling skills or rebound control, then you are quite clearly wrong. I have done a lot of research in that area, and looking at pairs of goalies on the same team shows that the shots against gap is not nearly large enough to be able to bridge the gap between a league leader and an average goalie.

However, I do agree with your statement. It is possible that a goalie with an average save percentage is a better goalie than a guy who is close to the league lead in save percentage when there are extreme team factors involved. That in no way invalidates the claims of this blog, because the overall thesis of this blog is that goaltending is very dependent on team defence. I have argued all along that a strong or weak defence has a big impact on goalie stats.

All of these things are very possible effects of a goalie with good puckhandling or rebound control, and should be taken into effect, not just as "oh we need to quantify them or they do not count" but rather as talents the have a positive team effect.

I am working on researching these effects. I agree they are effects, but we can't take them into effect if we don't know how large they are. All the evidence I have seen suggests that they are much less significant than a goalie's ability to make the first stop, and that ranking goalies based on their ability to stop pucks is not exactly precise but leads to results that are mostly correct.

The Contrarian Goaltender said...

E:

because ignorance and those that prey upon it are my enemies.

Which one is it, do you think I am ignorant or I am preying on ignorance? I'm guessing it is the former, because ignorant people are not exactly my target audience here. Quite frankly I expect ignorant people to not understand the statistical arguments on this blog, to disagree with most of my points because they go against conventional wisdom and the glib explanations that get spouted in the mainstream media, and to leave a comment like "THIS SITE IS GARBAGE!"

The Contrarian Goaltender said...

Does that sound like I'm questioning the data itself?

Of course you aren't questioning my data on New Jersey shot prevention now, because I posted it all for you. Yet you were definitely questioning it before, and the record shows that your objections were unfounded.

Again, that you quoted it leads one to believe it exists. Faulting me for somehow being wrong in requesting your sources or for not doing my own legwork is where the disingenuity comes in.

You are getting confused between two different things, both of which I referred to when I talked about New Jersey's shot prevention record. One was the compiled shot prevention data for every team over the last 15 seasons, which I only had in an unpublished spreadsheet. I included New Jersey's numbers in a comment but didn't feel it was necessary to cite a source because the data was from publicly available numbers released by the NHL. The other source was my research on playoff save percentage. I made a tangential reference in that post to shots against numbers related to Brodeur. I cited that source when asked about it.

I have never criticized you for requesting sources, because you are entirely within your rights to do so and I will always provide them.

Slight difference: a journalist has objectivity and fact checkers.

Bloggers (who are not journalists by any stretch, but nice try) have neither.


Trust me, bloggers have fact checkers. Just because they aren't paid doesn't mean they aren't there. Post something incorrect on a blog with decent traffic and you will get called out on it right away.

On the other hand, sports journalists are not objective because they travel with the team and have to get their quotes from the players every day. There are certain players who are friends with reporters and others who don't play ball with the media, which impacts the way they get covered. Journalists also have to maintain a good relationship with the higher-ups on the team, try to develop inside sources on the team, etc.

By the way, I didn't mean to imply I see myself as a journalist. I see myself as a researcher, and nothing more.

Which I stand by when you cite that article as evidence of a further claim, knowing that the information is probably not entirely correct to begin with. That's the disingenuous part.

My claim and the reason for citing that post had to do with shots against. That was not the main point of the post, and I didn't publish most of the shots against numbers, but I pointed out that the top 2 goalies in era-adjusted shots against per playoff game were Martin Brodeur and Gerry Cheevers. My disclaimer, and your problem with the article, were both related to the impact of team defence on save percentage, something that is irrelevant to shots against numbers. The crucial fact that I referred to still stands, and supports the well-demonstrated position that New Jersey has an excellent shot prevention record.

I disagree with manipulation of figures to get desired results.

So do I. You must be remarkably prescient to know exactly what my desired results are. And I guess the possibility that my viewpoint is correct and supported by evidence is something that you aren't even willing to consider.

Of course, as per usual, you neglect to list option C, which would be, "Only someone who understood the difference between opinion and fact would consider the discussion of greatness pointless."

You're right, I did forget option C: "Only someone with a complete lack of intellectual curiosity and a desire to stifle debate and obstruct further study would consider the discussion of greatness pointless."

The fact that you, an admittedly biased person, decide on the worth of data and its being discarded or included nullifies any and all conclusions you reach based upon that information.

Ad hominem fallacy, anyone? You apparently have this idea that I have a treasure trove of data somewhere that proves Martin Brodeur is the greatest ever, yet I am hiding it away because it disagrees with my opinion. That is most certainly not the case. I have invited you many times to email me and ask for a full copy of my data and calculations on a particular topic. My email inbox remains empty, and that tells me all I need to know about whether you have a legitimate desire to scrutinize my work or if you are simply here to disagree.

Simple, the entire team plays differently when he's in there. I thought we covered this much earlier also...

Yes, the team probably plays differently. You seem to see that as the ending point to the debate, but I see it as just the starting point for analysis. Why would the Devils decide to allow more shots against when Clemmensen is in net? Is that because they are trying to create more offence, and therefore give up more chances against? Are they changing up their defensive zone coverage? What kind of shots are the extra shots against, or they perimeter shots or dangerous scoring chances? Even if you think these things aren't quantifiable, they are still observable by watching the team play. Changes to defensive play also affect offensive play, so we can look at shots for, goals for, and win/loss record to get a full picture of how the team play changes, as well as other defensive measures such rebound shots against, shots against broken down by game situation, and shot quality measures like average shot distance.

Those are interesting questions, far too interesting to dismiss with a simple "The entire team plays differently." If you are a GM wanting to know how your team will do with a particular goalie in net, you would like to have some idea of how the switch affects the play of the entire team.

Anonymous said...

@CG

This is really the best you can do?
"Yet when presented with the EXACT same claim for Martin Brodeur, you for some reason change your mind. Hypocrite? I think so.

No, I'm not a hypocrite, because my reasons for rejecting that claim are not because I don't think cumulative performance matters, but because I disagree that Brodeur's cumulative performance surpasses his peers. I look at more than wins and shutouts when I evaluate goalies - I take into account save percentage, team-adjusted and shot quality adjusted stats, and they do not tell me that Brodeur is more consistent or that his performance over his career has been better than everyone else. If Martin Brodeur was legitimately a top 5 goalie for the last 15 years in a row, then you could make a very good case for him as one of the best ever. I don't believe that to be the case, based on my observations and detailed study of the numerical record."

That is pathetic. Looking at New jersey's year by year rank at shots allowed there are year where they are not #!, or even outside the top #5. I am not arguing number for number but your logic used in one scenario vs the hypocritical change in logic in another. To say Brodeur has not been a top 5 goalie every year of his career is a pretty tough claim to make, and I have yet to see you prove that, however looking again at your assessment of New Jersey as the "best" at shots allowed, it would be fair to say that even if Brodeur was not top 5 every year that it would still be possible to have him as the cumulative best over the course of his career. You can't use certain lines of though when it helps make your point and then disregard it when it doesnt.

Another interesting point that went completely ignored was this
"Further discrediting your argument regarding save percentage only, IF there is ANY correlation between rebound control, and/or puckhandling, and the number of shots allowed per game, something almost anyone will acknowledge, then you must now take gaa into consideration as it is the only statistic which may be reflective of a goalie who excels at those things.

You seem to be blind to the effects of a goalie who can play the puck as well as the likes of Brodeur, Turco, or Dipietro, as I remember pretty often how people would jokingly refer to New Jersey's trap system as a 1-2-3. the 3 obviously accounting for the 2 defenseman, and then the 3rd, Brodeur."

Again, just because in your biased personal opinion, and maybe that of a few others, you seem to incorrectly think save percentage and other manipulated forms of it are the best methods for determining goalie performance, does not mean it is the only method. If there is an unquantifiable effect on preventing goals, then goals against average would naturally carry some weight as well.

Statman said...

Anonymous - Please list the top 10 (or at least top 5) goalies over the past 15 yrs, in regards to various statistical categories.

[After all, you're arguing that Brodeur is #1, cumulatively, in certain categories over the past 10-15 yrs, no?]

Anonymous said...

@CG
"If you are arguing that is possible because of puckhandling skills or rebound control, then you are quite clearly wrong. I have done a lot of research in that area, and looking at pairs of goalies on the same team shows that the shots against gap is not nearly large enough to be able to bridge the gap between a league leader and an average goalie."

Oh, well in that case then there is obviously no effect?!?! I guess because somebody with an admitted bias, and an admitted agenda, has decided to do some "studies" and as a result concludes there is not much of an effect I should then ignore the many studies done, some of which Bruce has pointed you in the direction of, most of which completely disagree with the "results" of the "studies" you have done. It is pretty self evident that rebound control and puckhandling can have a major effect on goal prevention, otherwise I doubt there would be emphasis on these skills in goalie development. You obviously have a tendency to work with certain numbers and scenarios in order to get the results to say what you want them to, while blatantly ignoring, discrediting and in some case purposely omitting information that would hinder the results you seem to desire.

Anonymous said...

@Statman
"Anonymous - Please list the top 10 (or at least top 5) goalies over the past 15 yrs, in regards to various statistical categories.

[After all, you're arguing that Brodeur is #1, cumulatively, in certain categories over the past 10-15 yrs, no?]"

I am not specifically saying Brodeur is a cumulative number 1 although it would be hard to say otherwise when using the "cumulative performance" line of thought, what I am addressing is the noted contradiction and inconsistencies in the general logic used depending upon what is trying to be proven by CG.

For the sake of following that line, and perhaps sparking further debate and analysis I would list my top 5 based on cumulative performance and productivity not over the past 15 years, but rather on the over course of their career as follows:
1. Brodeur
2. Plante
3. Roy
4. Hall
5. Hasek

In judging the PAST 15 years, there is an obvious overlap in Brodeur's case, which in an advantage for him, as well as retirement/skill drop off for other players that makes this list less relevant, but again, for the hell of it, here is my "top 5 over the past 15 years".

1. Brodeur
2. Hasek
3.Belfour
4. Kolzig
5. Luongo

Statman said...

Anon - Ok, thanks for the lists.

But how do you come up with them? What criteria? Is #1 a lot better than #2? What are the differences between #4 & #5, for example?

The purpose of this site & others like them is to, as best possible, quantify results. I'm not meaning to criticize you, but every hockey fan can come up with a Top 10/Top 5 list, based on what comes to their mind, their impressions from watching t.v. & reading mainstream media, etc.

One thing CG has written about (as have other site authors) is how sometimes when you look at the #'s & really dig into things, you can be surprised at how the results conflict with "conventional wisdom". And sometimes people get really upset (apparently) when the results conflict with long-held beliefs.

Anonymous said...

@ Statman

Obviously there are a number of things that I, or anyone else should consider, not just one statistic or area, but like I have said, many.

First for me, it came down to who gave his team the most consistent, and best performance for the longest period of time. This is something that is not appreciated at least by CG, however I find there to be far more value to a team if a goalie can give them consistent results for "the longest time possible", as opposed to the "Miikka Kiprusoff" or "Jim Carey" dominance for a few years. I would even go as far as saying 10 years of Hasek like dominance, is not equivalent to 15 years of "top 10" overall performance. To those inside the game, i.e. player, coaches, general managers, etc, ask them what is the one thing they want most from their players and it is consistency. Look at the problems Montreal is currently having with Kovalev. If a goalie gives his team consistently well above average performance for 20 years that is about all you can ask of him. Ideally, the team would like 15-20 years of Hasekesque dominance however that has never been done. In terms of consistencyit allows the team to focus on other areas, and thus improve, an additional benefit and "unquantifiable" effect on the team.

Secondarily, would be the players ability to fit in with the teams system. Using an obvious example in another sport, nobody will argue that Terrell Owens is not one of the most talented receivers in the league, however there are plenty of team who would not want him on their team. Roy's selfish attitude set Montreal back quite a bit, and perhaps cost team Canada a gold medal.

Also taken into effect would be the players peak in performance, however as stated earlier, consistency over time, far outweighs flash on the pan greatness. I personally weigh peak performance only as significant as the drop off in years in which that player is not at his peak yet seeks to be played as if he is. Kolzig is Washington comes to mind as a guy who has had his peaks, however has been almost forgotten during his years with the team, which is a sign that the team is content with his play.

In my opinion numbers are great, however no single number is the end all be all, and because of the fact that numbers change from era to era, I weigh the tenure of a player, specifically in regards to a goalie as something of significance because ultimately what does not change nearly as much, is the way in which management constructs a team. Sure a 4th liner who is with a team for 15 years, i.e. Draper, is not great in its most commonly used definition, but I would say he is one of the greatest 4th liners. With a goalie, especially one that starts for a prolonged period of time with one team, the results are a little bit more meaningful.

The Contrarian Goaltender said...

Anonymous:

To say Brodeur has not been a top 5 goalie every year of his career is a pretty tough claim to make, and I have yet to see you prove that, however looking again at your assessment of New Jersey as the "best" at shots allowed, it would be fair to say that even if Brodeur was not top 5 every year that it would still be possible to have him as the cumulative best over the course of his career.

Brodeur has two seasons where his save percentage was below league average on one of the best defensive teams in the league, so it is not particularly difficult to prove he has not been in the top 5 every season. However, my use of top 5 was just as an example, you are correct that the cumulative best is what is important and not necessarily whether he meets some arbitrary cutoff. However, refer to this post that compares save percentages to league average to see how Brodeur ranks against some of his peers, and that is not taking into account factors like shot quality and number of shots faced on the penalty kill. By my numbers, Hasek's cumulative performance compared to average is almost twice that of Brodeur's. The gap will be narrowed a bit once we are able to fully value things like rebound control and puckhandling, but that won't come even close to making up the 200+ goal gap that separates him from Hasek and Roy.

"Further discrediting your argument regarding save percentage only, IF there is ANY correlation between rebound control, and/or puckhandling, and the number of shots allowed per game, something almost anyone will acknowledge, then you must now take gaa into consideration as it is the only statistic which may be reflective of a goalie who excels at those things.

I don't think it makes any sense at all to take GAA into consideration, except when comparing goalies on the same team. I agree with you that it appears that goalies do have some small impact on shots allowed, but to me it makes more sense to try to isolate the shot prevention number and then use it to adjust save percentage by counting the shots he prevented compared to average as saves. Even based on very conservative estimates, over 90% of a goalie's shots against number is determined by the rest of the team, so it doesn't make sense to me to bring in the extra variable of team shot prevention through the use of GAA. I am currently working on ways to try to estimate shot prevention impact, through posts such as the one above.

I guess because somebody with an admitted bias, and an admitted agenda,

Oh please, you're starting to sound like E. Yes, I have biases. So do you, and so does everyone else. If I was making completely subjective arguments based on my own views, then that could be a legitimate complaint. If I use an argument based on official NHL numbers, then it doesn't matter what I think or what anyone else thinks - the argument is what it is, and you need to address the premises and conclusions to defeat it.

And yes I have an agenda - to correctly evaluate NHL goalies. Shocking, I know.

It is pretty self evident that rebound control and puckhandling can have a major effect on goal prevention, otherwise I doubt there would be emphasis on these skills in goalie development.

No, it is not self-evident. Look, I'm a goalie, I love handling the puck and I am always working on improving my rebound control. It makes intuitive sense that those things are important, but the numbers just aren't there to support them having a "major effect on goal prevention".

You obviously have a tendency to work with certain numbers and scenarios in order to get the results to say what you want them to, while blatantly ignoring, discrediting and in some case purposely omitting information that would hinder the results you seem to desire.

Not true. I blatantly ignore certain statistics, that is clear. The reasons for that have nothing to do with the results I desire, but are based on my philosophy of goaltending evaluation. When I do research, I define the evaluative criteria, run the numbers, and post whatever comes out. If you want to argue that wins or GAA or something like that is meaningful, then argue that point. Don't accuse me of omitting data and making up numbers to support my views, because that is just not true.

By the way, we agree on 4 of the top 5 goalies all-time. Kolzig over Luongo is a curious choice, but otherwise I don't think your choices are too bad, other than that I obviously think you are substantially overrating Brodeur.

Statman said...

Anon - I don't have much time to reply. I will note that your lists are fine... for subjective lists.

You have jumped all over CG for supposedly distorting & leaving out stats... but I will say that your lists have zero backing behind them - at least, you haven't listed the various ways you've determined your list. Yes, you did mention some broad things like "consistency" & "fitting in with the team". But what is witin this things, what are they made up of? Is Consistency twice as important as Fitting In With The Team? Etc.

There's nothing wrong with your lists, if you want to say that you came up with them by sort of using the various impressions you've stored over the years, recalling some broad stats like a lot of wins & all-star selections, etc. I think every sport fan starts out that way. But I really don't get the attacks on this site, when at least one side is presenting numbers & the other presents nothing other than a subjective list.

G'night all.

Anonymous said...

@Statman
"But how do you come up with them? What criteria? Is #1 a lot better than #2? What are the differences between #4 & #5, for example?"

In addition to the things I mentioned, I further would like to add emphasis to a players tenure with a team, and ability to stay at the position, by using an over rated goalie like Evgeni Nabakov as an example of somebody who is supposedly good, yet has had trouble proving even to his own organization that he is the best they can do, as he has lost his starting job numerous times. Also as in the case of other goalie, bouncing from team to team is an obvious negative; not definitively negative, but in correspondence with circumstance. For instance, Wayne Gretzky bounced around quite a bit for the "Greatest player ever", however was he even replaced on the teams that got rid of him? The same thing is important to "look" at because it shows cases, as in the example of Hasek with Ottawa, and with Detroit, that the team felt they did not need him, and did fine without him. It is also a detriment to the team and another "unquantifiable" negative effect if every x amount of years the team has to worry about their goalie running off for a payday. How reliable really is a guy who is only looking out for himself? Again, I look at this as supplementary value analysis, and admit there are times when there is no value at all to using this approach, however it does give insight into a players value to a team.

Anonymous said...

@ Statman

The problem with the way in which results are gathered on this blog is that they are completely dependent on very few, handpicked statistics. I do not have the time to write a blog such as this one, therefor I will not have time to break down values and weighting of every single method by which I evaluate a goalie, however my approach differs from the one used here because it accepts the reality that not everything can be quantified, and further goes on to give credit for unquantifiable things, where as here they are ignored until they can be assigned a statistical value.
I am already expecting to hear the question "How can you give weight to something that can not be quantified?" its simple, it is partially quantified in team success, but more importantly, it shows itself over time in relation to individual success.