Martin Brodeur

[seventieslord]

That is what it shows. Where in my data does fewer shots cause a higher save %?

by the "other way around" I'm not saying fewer shots causes a higher sv%, I'm talking about which causes which.

Like HO said, this can easily be construed to mean one or or the other. Some say fewer shots causes lower sv%, some say lower sv% causes fewer shots. It's inconclusive at best.

 

[overpass]

Haakon84, my critique in post 103 still applies, but I'll be more blunt. Before you put any stock in these numbers, there are a couple of things that you need to do to fix them.

You need to fix the causation effect. Low save percentages can cause a playing to the score effect. See the data I posted on first period shots only. The range of save percentage is narrower than the range for the full game, because the playing to the score effect is not present to the same degree in the first period.

Why do you need to fix this? For goalies that play of bad teams, it's very likely that their low-shot games will come when they have allowed goals early, are behind, and the other team is playing defensively. For goalies that play on good teams, it's more likely that they are just facing few shots because of excellent defence. Goalies on strong teams are more likely to play with the lead, and it's easier to have a high save percentage when playing with the lead.

Second, you absolutely need to fix the selection bias effect - and I have no idea how you can do that. But it's a fact that when you select a group of games with low shots recorded, you are selecting 1) games in which few shots were actually taken, and 2) games in which too few shots were recorded. Underrecording of shots will cause save percentage to be lower. You must consider this effect before doing any adjusting!

Finally, why not look at seasonal numbers instead of game-by-game numbers? Seasonal numbers will eliminate the selection bias effect, as we would expect random variation in shot bias to basically cancel out after 50-60 games. We know the result is that we see no relationship between shots faced per game and save percentage. I think that is the best way of looking at this issue for now.

 

[haakon84]

Haakon84, my critique in post 103 still applies, but I'll be more blunt. Before you put any stock in these numbers, there are a couple of things that you need to do to fix them.

You need to fix the causation effect. Low save percentages can cause a playing to the score effect. See the data I posted on first period shots only. The range of save percentage is narrower than the range for the full game, because the playing to the score effect is not present to the same degree in the first period.

This is all inference and any assumption here is based on data less conclusive than mine. I don't see why it is my responsibility to adjust for a dubious claim. I'll be more willing when more evidence is piled up. Your numbers in the first period show a similar trend to my numbers. I expect that trend to continue in the 2nd period, and the 3rd period. Where is the data for this? Wouldn't that support your first period data?

Why do you need to fix this? For goalies that play of bad teams, it's very likely that their low-shot games will come when they have allowed goals early, are behind, and the other team is playing defensively. For goalies that play on good teams, it's more likely that they are just facing few shots because of excellent defence. Goalies on strong teams are more likely to play with the lead, and it's easier to have a high save percentage when playing with the lead.

Again all assumptions and I see no evidence besides mantra from TCG to back this up. There are multiple scenarios as to why a team faces fewer shots and it will be impossible to adjust to all of them. What is apparent that regardless of circumstance facing fewer shots results in a lower save % and a lower GAA. Of course you could play seventieslord advocate and just flip it around without providing any evidence and only quoting other people's points.

It's obvious as to why goalies have a much lower save % in both wins and losses and that is because a goal has a much larger impact on save % in relationship to shots. Id like to see the gap and save %'s in wins and losses above 35 shots. Don't you think that data would tell us something about the low shot totals? That is my whole point. Here is a chart showing th GAA a goalie needs to have a .920 save % when facing a certain number shots per game.

SPG | Save % | GAA
40 | .920 | 3.2 GAA
35 | .920 | 2.8 GAA
30 | .920 | 2.4 GAA
25 | .920 | 2.0 GAA
20 | .920 | 1.6 GAA
15 | .920 | 1.2 GAA

Second, you absolutely need to fix the selection bias effect - and I have no idea how you can do that. But it's a fact that when you select a group of games with low shots recorded, you are selecting 1) games in which few shots were actually taken, and 2) games in which too few shots were recorded. Underrecording of shots will cause save percentage to be lower. You must consider this effect before doing any adjusting!

I have seen very few cases where people adjust for this and again if there is one goaltender on that list who would benefit from this tremendously it would be Martin Brodeur and this may be why (along with his elite defensive teams) his SPG ratio is skewed towards the under 30 SPG category.

Finally, why not look at seasonal numbers instead of game-by-game numbers? Seasonal numbers will eliminate the selection bias effect, as we would expect random variation in shot bias to basically cancel out after 50-60 games. We know the result is that we see no relationship between shots faced per game and save percentage. I think that is the best way of looking at this issue for now.

I have looked at seasonal numbers and they follow the same trend as career numbers! I will post more of that data when I get around to it but I don't think season to season we aren't going to see much a difference. When a goaltender faces a majority of games with above 30 shots his save % is higher than when he doesn't. For example Brodeur pre and post lockout, Hasek pre and post Buffalo. I have yet to see any solid evidence refuting that claim.

I appreciate the critique and will take much of the input into consideration when compiling more data. I do think it is harping on minutia based on inferences from incomplete data that TCG has posted. And you and seventieslord have clearly shown your allegiances to his body of work (nothing wrong with that) I think this may be fogging up your vision though. He used an 8% sample of ALL goaltenders on varying teams over 3 years. I challenge him to categorize it in a matter similar to mine and see what he infers from that.

 

[Master_Of_Districts]

Again all assumptions and I see no evidence besides mantra from TCG to back this up. There are multiple scenarios as to why a team faces fewer shots and it will be impossible to adjust to all of them. What is apparent that regardless of circumstance facing fewer shots results in a lower save % and a lower GAA. Of course you could play seventieslord advocate and just flip it around without providing any evidence and only quoting other people's points.

You completely neglected to address his point regarding playing to the score effects.

The effect of facing 35 shots in a game for a particular goaltender will vary according to how many shots that goalie's team surrenders on average. In other words, if a goalie playing for a team that gives up 23 shots/game on average faces 35 shots in a particular game, it's likely that playing to the score effects were in operation. Here, the effect on the goalie's save percentage will be positive. If, however, the goalie is playing on a team that surrenders 35 shots against on average, then it's unlikely that playing to the score effects were involved, and the effect upon save percentage is different or non-existent.

Quite simply, your analysis reflects a fundamental misunderstanding of the way that the game is played.

Moreover, if your position is correct, then one would expect to observe a positive correlation between seasonal save percentage and shots against/game at team level. As has been alluded to previously, no such relationship exists.

 

[haakon84]

You completely neglected to address his point regarding playing to the score effects.

The effect of facing 35 shots in a game for a particular goaltender will vary according to how many shots that goalie's team surrenders on average. In other words, if a goalie playing for a team that gives up 23 shots/game on average faces 35 shots in a particular game, it's likely that playing to the score effects were in operation. Here, the effect on the goalie's save percentage will be positive. If, however, the goalie is playing on a team that surrenders 35 shots against on average, then it's unlikely that playing to the score effects were involved, and the effect upon save percentage is different or non-existent.

Quite simply, your analysis reflects a fundamental misunderstanding of the way that the game is played.

No I understand the concept and I have yet to find any data that supports this. As seventieslord said it is inconclusive at best. Combine that with inconclusive data and there never will be a conclusion.

Moreover, if your position is correct, then one would expect to observe a positive correlation between seasonal save percentage and shots against/game at team level. As has been alluded to previously, no such relationship exists.

Yes when you take all teams and all goaltenders into account as I said earlier I do not expect to see a correlation. Again, I have looked at the seasonal data and posted snippets here. I expect that when shots per game is broken up into categories you will see a more glaring correlation.

 

[GNick42]

I always felt Brodeur was over rated where he played behind such a great defense in Scott Stevens and Nieyermayer. Besides the master of the trap as head coach.

Look at Scott Clemmensen and numbers he put up last year playing in that system? He never did nothing this year for Florida nor did he earlier in his career for Toronto before he played for Jersey.

While Roy singlehandely took a so-so team to 2 Stanley Cups.

 

[BM67]

I always felt Brodeur was over rated where he played behind such a great defense in Scott Stevens and Nieyermayer. Besides the master of the trap as head coach.

Look at Scott Clemmensen and numbers he put up last year playing in that system? He never did nothing this year for Florida nor did he earlier in his career for Toronto before he played for Jersey.

While Roy singlehandely took a so-so team to 2 Stanley Cups.

You mean the 86 Canadiens who tied for 6th in GF and had the 7th best record in the league? (I guess Robinson, Chelios, Gainey and Carbonneau doesn't compare to Stevens and Niedermayer either.) Or the 93 Canadiens who were 9th in GF and had the 6th best record in the league?

Yeah they are so much more so-so than the 95 Devils who were tied for 13th in GF and had the 9th best record in the league.

Montreal had home ice advantage in 3 of 4 series for both of their two Cups, while the Devils were on the road for all 4 series in 95.

 

[TheDevilMadeMe]

Second, you absolutely need to fix the selection bias effect - and I have no idea how you can do that. But it's a fact that when you select a group of games with low shots recorded, you are selecting 1) games in which few shots were actually taken, and 2) games in which too few shots were recorded. Underrecording of shots will cause save percentage to be lower. You must consider this effect before doing any adjusting!

Finally, why not look at seasonal numbers instead of game-by-game numbers? Seasonal numbers will eliminate the selection bias effect, as we would expect random variation in shot bias to basically cancel out after 50-60 games. We know the result is that we see no relationship between shots faced per game and save percentage. I think that is the best way of looking at this issue for now.

For this part, why is that any different for what he's doing than anyone else who tries to prove something using save %?

 

[overpass]

For this part, why is that any different for what he's doing than anyone else who tries to prove something using save %?

I'll start by stating my assumptions. Shots on goal are not recorded exactly as they are taken. The difference between actual shots and recorded shots may vary in the short run, but will zero out in the long run in most arenas, or at least even out to a consistent number.

If this is the case, we can assume that actual shots=recorded shots for a period of time or a season or longer*. Therefore this effect is not a concern.

But on a game-by-game basis, the actual shots will not always match the recorded shots. Since actual shots always match recorded goals, games in which shots are under-recorded will have recorded save percentages that are lower than the true save percentage. Games in which shots are over-recorded will have higher recorded saver percentages than true save percentages. We already know that single-game save percentages are far from perfect at evaluating single-game performance, as quality of scoring chances can vary from game to game, so in practice this isn't a big deal.

The real problem comes when dividing games into categories based on recorded shots. When we select for low-shot games, we are selecting both for "true" low-shot games, and for games in which shots were under-recorded. This means that low-shot games will tend to have lower "recorded save percentages" than "true save percentages", and vice versa.

This effect is only an issue when looking at save percentage for a selection of games, rather than at the save percentage for all games or for a random sample of games.

*In some cases there may be a bias, as has been suggested for New Jersey. If this is the case, we can add a constant to the shots recorded number to correct for it, based on the estimated size of the bias.

 

[GNick42]

Hasek = Orr
Roy = Gretzky
Brodeur = Howe

That's how I see it at least.

Brodeur is overrated. He wouldn't even be in top 10 goalies all time if played on a weak team his career. Brodeur played his whole career behind the trap and with 2 best defenseman in the league. Look at numbers Scott Clemmenson put up his one year in Jersey? He never did nothing this year in Florida nor before with the Leafs.

 

[barneyg]

First of all, this is great stuff, even if it's very far from the OP's heartfelt praise for Brodeur.

I have seen very few cases where people adjust for this and again if there is one goaltender on that list who would benefit from this tremendously it would be Martin Brodeur and this may be why (along with his elite defensive teams) his SPG ratio is skewed towards the under 30 SPG category.

Here's my $.02:

1) Even if some state effects (such as score effects) were somehow taken into account, Brodeur's SV% is likely understated because his home games were played in an arena that undercounts shots: http://objectivenhl.blogspot.com/2009/12/shot-recording-bias-florida-and-new.html. I don't know if there's any evidence on score effects and SV%. That sure would be interesting. I don't think we can assume that SV% is higher or lower when a team is leading or trailing.

2) BUT what's easily shown is that SV% is higher at even strength than on special teams, and NJ is notorious for having been among the least penalized teams throughout Brodeur's career (certainly since 97-98, I don't have the data before that). That biases Brodeur's SV% upwards.

 

[Ogopogo*]

I have Brodeur as the #4 goalie of all time right now. The way this year is going, he could end up #3 or even #2 by season's end.

 

[GNick42]

You mean the 86 Canadiens who tied for 6th in GF and had the 7th best record in the league? (I guess Robinson, Chelios, Gainey and Carbonneau doesn't compare to Stevens and Niedermayer either.) Or the 93 Canadiens who were 9th in GF and had the 6th best record in the league?

Yeah they are so much more so-so than the 95 Devils who were tied for 13th in GF and had the 9th best record in the league.

Montreal had home ice advantage in 3 of 4 series for both of their two Cups, while the Devils were on the road for all 4 series in 95.

See, what I mean by Brodeur being overrated?? Games like this expose him. He's not that good of a goalie.

People think he is a great goalie because he played in a defensive shell in Jersey and behind two great defensemen most his career.

 

[GNick42]

I have Brodeur as the #4 goalie of all time right now. The way this year is going, he could end up #3 or even #2 by season's end.

Brodeur I would have top 10...but not close to top 3-4.

He is not close to Plante, Hasek, Sawchuk, Hall, etc...

I would take Roy ahead of him, especially in the playoffs. I have to go with Parent ahead of him. Parent won two Cups by himself. But I would put Brodeur in the mix with Cheevers, Dryden, Moog, etc...

 

[Blades of Glory]

Brodeur I would have top 10...but not close to top 3-4.

He is not close to Plante, Hasek, Sawchuk, Hall, etc...

I would take Roy ahead of him, especially in the playoffs. I have to go with Parent ahead of him. Parent won two Cups by himself. But I would put Brodeur in the mix with Cheevers, Dryden, Moog, etc...

Brodeur is well behind the big three (Hasek, Plante, Roy), and is in the same group as a Sawchuk or Hall, albeit behind them (in my books at least). I think it's a little stretch to compare him to Gerry Cheevers and Andy ****ing Moog. Come on now.

 

[seventieslord]

Brodeur is well behind the big three (Hasek, Plante, Roy), and is in the same group as a Sawchuk or Hall, albeit behind them (in my books at least). I think it's a little stretch to compare him to Gerry Cheevers and Andy ****ing Moog. Come on now.

Agree, Moog or Cheevers, that's a major stretch.

 

[NOTENOUGHJTCGOALS]

Brodeur is well behind the big three (Hasek, Plante, Roy), and is in the same group as a Sawchuk or Hall, albeit behind them (in my books at least). I think it's a little stretch to compare him to Gerry Cheevers and Andy ****ing Moog. Come on now.

Why is Moog grouped in with Dryden?

 

[GNick42]

Are you kidding....Moog cleaned us in 4 times in the playoffs.

Dryden was like Brodeur overrated. He played on a great team with league best defensemen in front of him(the big three). See Dryden was relatively average in international hockey. Game 6 of the Summit Series was only game he played at a top level.

 

[seventieslord]

Are you kidding....Moog cleaned us in 4 times in the playoffs.

Dryden was like Brodeur overrated. He played on a great team with league best defensemen in front of him(the big three). See Dryden was relatively average in international hockey. Game 6 of the Summit Series was only game he played at a top level.

Yes, Dryden had a great team that helped him, no doubt, but despite that, he was able to shine individually like Brodeur rarely has. Dryden's sv% in both the regular season and playoffs was always at an elite/dominant level.

 

[TheDevilMadeMe]

See, what I mean by Brodeur being overrated?? Games like this expose him. He's not that good of a goalie.

People think he is a great goalie because he played in a defensive shell in Jersey and behind two great defensemen most his career.

Brodeur I would have top 10...but not close to top 3-4.

He is not close to Plante, Hasek, Sawchuk, Hall, etc...

I would take Roy ahead of him, especially in the playoffs. I have to go with Parent ahead of him. Parent won two Cups by himself. But I would put Brodeur in the mix with Cheevers, Dryden, Moog, etc...

I find it ironic that you take Brodeur's play in a game when he's almost 38 years old as proof that he's been "exposed," then immediately afterwards say that you'll take Roy over him.

 

[Dennis Bonvie]

I find it ironic that you take Brodeur's play in a game when he's almost 38 years old as proof that he's been "exposed," then immediately afterwards say that you'll take Roy over him.

Indeed, the one game analysis to reach a conclusion isn't well received on the History site.

I guess after the attempt to downgrade Bobby Orr based on a single playoff game everyone is vunerable to the bad game syndrome.

 

[GNick42]

I find it ironic that you take Brodeur's play in a game when he's almost 38 years old as proof that he's been "exposed," then immediately afterwards say that you'll take Roy over him.

Roy was a great money goalie...yes, I have Roy ahead of Brodeur. Brodeur is overrated. He is like Dryden, played behind such a great defense, they made him look 25% better than he was.

 

[TheDevilMadeMe]

Roy was a great money goalie...yes, I have Roy ahead of Brodeur. Brodeur is overrated. He is like Dryden, played behind such a great defense, they made him look 25% better than he was.

Whoosh.

 

[mrhockey193195]

Wow, I hate brodeur just about as much as anyone can, and I definitely do not think he's top 4 all time, but some of the stuff being thrown out here is ridiculous, even by my standards.

Comparing Brodeur to Moog and Cheevers is more than a stretch. And grouping Dryden in there as well is just not right.

I understand the arguments that Dryden and Brodeur are slightly overrated because of their "inflated" numbers: hell, I make that argument all the time. But come on now, at some point you need to just go by what you saw. And watching Brodeur play, there's little question he's a top 10 goalie of all time (a little hypocritical, I know, since I never saw Sawchuk, Plante, etc. play, but Brodeur is easily top 3 of the goalies I've seen play in the last 20 years).

No matter how good of a team you have, you can't put up the numbers Brodeur did for so long, so consistently without being damn good. Ditto for Dryden (though his early retirement takes away a big part of what is Brodeur's legacy: the fact that he did it every year for 15+ seasons).