The Brodeur/Roy/Hasek debate

[Ol Dirty Bstrd]

As far as I understand, the consensus is Roy is the best of these three, then Hasek, the Brodeur. I never understood this thinking. I will try and break down my thinking.

Roy vs. Brodeur - the same old Brodeur excuses cant be used to downplay him because Roy played behind some stellar teams as well. Also, Brodeur has a slightly higher career sv%, so Brodeur haters can throw that out the window. So Roy is generally considered as better because of his playoff success. While I agree, I will just give some numbers to show brodeur isnt too far behind:

Win% SO SV% GAA Cups Conn Smythes
Roy .616 23 .918 2.30 4 3
Marty .582 21 .921 1.89 3 0

The only category Roy blows Marty away in is Conn Smythes and Marty EASILY could have won it any three times he won the cup especially when he was robbed of it by Giguere. Here are his stats in those playoffs:

Year SO GAA SV%
94-95 3 1.67 .927
99-00 2 1.61 .927
02-03 7* 1.65 .934
*record for shutouts, including 3 in finals

Those are phenomenal numbers and the "good team" excuse is only so valid because look at his save percentages. For example, he lost in the first round in 01-02 with a 1.42 GAA and a .938 sv%, so his team in front of him couldnt have been that good at all. I could see some people picking Roy for one game as more of a "money goalie," but Brodeurs postseason numbers are pretty close to Roy's and he was slightly better in the regular season. Keep in mind ROy never even started 70 games. Marty's durability and puckhandling should definitely be taken into account.

As the numbers show, its pretty close so I can see people taking ROy over Marty. But heres the real headscratcher. People then rank Hasek in between Roy and Marty. Hasek in his prime was the most dominant goalie ever, easily better than both of them. But his prime was very short and that causes you to choose longevity vs. prime when discussing these goalies. How can someone choose longevity (or clutchness) when comparing Roy and Hasek, but choose prime when comparing Hasek and Brodeur. Keep in mind Brodeur and Roy BOTH played behind great teams, while Hasek didn;t for the most part. Also, once again Hasek only started 70 games once while Marty does it ever year. Brodeur had Hasek like numbers through January last year, leading the league in every major goaltending category, but those numbers fell because he started many more games than the dominator typically would in his career.

Sorry for the rant, but do I make valid points? Does Brodeur get downplayed by some? Can someone without bias explain the Roy then Hasek then Broduer logic? THanks.

Click to read more...

 

[Augustus Pablo]

Brodeur is the best of the three.

 

[SabresFan26]

Hasek-Roy-Brodeur in that order.

Brodeur really benefited on his team with the defensive style of play. Hasek in his prime was the best of all three.

 

[Augustus Pablo]

Hasek-Roy-Brodeur in that order.

Brodeur really benefited on his team with the defensive style of play. Hasek in his prime was the best of all three.

Hasek benefitted from the Sabres defensive style of play. What makes Roy better then Brodeur?

 

[Bear of Bad News]

Looks like a good foundation for debate - I'll probably comment more in depth when I'm home in the evening.

One to start - you really need to normalize Roy's and Brodeur's save percentages and goals-against averages if you want to compare them directly - although a large portion of their careers overlapped, there's still a considerable portion that did not.

Based on the numbers I have, Roy over his career had a save percentage z-score of 8.317, meaning that he was about 8.3 standard deviations above average in that regard.

Brodeur is at 4.309 (through 2006-07), while Hasek is an eye-popping 9.749 (again, through 2006-07).

Edit - just realized that your numbers are all postseason. Roy's postseason z-score is 3.705, Brodeur's is 1.758 and Hasek's is 2.497.

 

[NOTENOUGHJTCGOALS]

As far as I understand, the consensus is Roy is the best of these three, then Hasek, the Brodeur. I never understood this thinking. I will try and break down my thinking.

Roy vs. Brodeur - the same old Brodeur excuses cant be used to downplay him because Roy played behind some stellar teams as well. Also, Brodeur has a slightly higher career sv%, so Brodeur haters can throw that out the window. So Roy is generally considered as better because of his playoff success. While I agree, I will just give some numbers to show brodeur isnt too far behind:

Win% SO SV% GAA Cups Conn Smythes
Roy .616 23 .918 2.30 4 3
Marty .582 21 .921 1.89 3 0

The only category Roy blows Marty away in is Conn Smythes and Marty EASILY could have won it any three times he won the cup especially when he was robbed of it by Giguere. Here are his stats in those playoffs:

Year SO GAA SV%
94-95 3 1.67 .927
99-00 2 1.61 .927
02-03 7* 1.65 .934
*record for shutouts, including 3 in finals

Those are phenomenal numbers and the "good team" excuse is only so valid because look at his save percentages. For example, he lost in the first round in 01-02 with a 1.42 GAA and a .938 sv%, so his team in front of him couldnt have been that good at all. I could see some people picking Roy for one game as more of a "money goalie," but Brodeurs postseason numbers are pretty close to Roy's and he was slightly better in the regular season. Keep in mind ROy never even started 70 games. Marty's durability and puckhandling should definitely be taken into account.

As the numbers show, its pretty close so I can see people taking ROy over Marty. But heres the real headscratcher. People then rank Hasek in between Roy and Marty. Hasek in his prime was the most dominant goalie ever, easily better than both of them. But his prime was very short and that causes you to choose longevity vs. prime when discussing these goalies. How can someone choose longevity (or clutchness) when comparing Roy and Hasek, but choose prime when comparing Hasek and Brodeur. Keep in mind Brodeur and Roy BOTH played behind great teams, while Hasek didn;t for the most part. Also, once again Hasek only started 70 games once while Marty does it ever year. Brodeur had Hasek like numbers through January last year, leading the league in every major goaltending category, but those numbers fell because he started many more games than the dominator typically would in his career.

Sorry for the rant, but do I make valid points? Does Brodeur get downplayed by some? Can someone without bias explain the Roy then Hasek then Broduer logic? THanks.

Roy also played more years in when scoring was higher and goalie stats in general were lower. Also, Roy and Brodeur both had stacked teams, the Avs were always primarily an offensive powerhouse while the Devils were a defensive one.

Watching all three goalies, Hasek stole so many games it wasnt funny, but he was a headcase and didnt always have consistent greatness, missing games due to mysterious injuries.

 

[Ol Dirty Bstrd]

Looks like a good foundation for debate - I'll probably comment more in depth when I'm home in the evening.

One to start - you really need to normalize Roy's and Brodeur's save percentages and goals-against averages if you want to compare them directly - although a large portion of their careers overlapped, there's still a considerable portion that did not.

Based on the numbers I have, Roy over his career had a save percentage z-score of 8.317, meaning that he was about 8.3 standard deviations above average in that regard.

Brodeur is at 4.309 (through 2006-07), while Hasek is an eye-popping 9.749 (again, through 2006-07).

Edit - just realized that your numbers are all postseason. Roy's postseason z-score is 3.705, Brodeur's is 1.758 and Hasek's is 2.497.

Thank you for pointing this out. I get the general idea of your calculations, but could you possibly explain what these specific numbers mean? Thanks.

 

[Datsyuk 13]

Roy played in a offensive system. Brodeur played in a defensive system.

 

[Bear of Bad News]

Thank you for pointing this out. I get the general idea of your calculations, but could you possibly explain what these specific numbers mean? Thanks.

No problem! It's a bit of an esoteric calculation, I must admit.

Basically, it's a method that you can use to estimate the probability that a goaltender is an average-ability netminder (obviously, none of the three goaltenders in this conversation is average by any stretch of the imagination).

I'll give a small example to explain how it works - suppose the league-average save percentage is 90%. Goaltender X makes three saves on three shots. To know if he's better than league average, we can calculate the probability that a league-average goaltender could make three saves on three shots simply by random chance.

This probability would be: (0.9)^3 = 72.9%. In other words, it's very likely that an average goaltender could do this. We would need more evidence to conclude that Goaltender X is better than average.

Now let's look at a real season - Martin Brodeur's 2006-07 campaign. Brodeur allowed 171 goals on 2182 shots for a 0.922 save percentage; the league-average save percentage was 0.905. What's the probability that a league-average goaltender could have (by random chance) put together a season as good as Brodeur's in 2006-07? Exactly Brodeur's season would be Comb(2182,171)*(0.905)^(2182-171)*(1-0.905)^(171), and then you'd have to add in every season better (170 GA, 169 GA, 168 GA, et cetera). It's a pain.

Fortunately, we have the normal distribution approximation to the binomial distribution. For large number of shots faced, we can approximate the above calculation by using the normal distribution (bell curve) with mean equal to shots*save percentage and variance equal to shots*save percentage*(1-save percentage).

In the case of Brodeur's season above, the normal approximation gives us a z-score of z=2.633, meaning that Brodeur's campaign was 2.633 standard deviations better than average in 2006-07. The probability that an average goaltender could do what Brodeur did? About 0.42% (or one in 238).

Now we can look at Brodeur's career as a whole. Through 2006-07, Brodeur had allowed 1931 regular-season goals on 22167 shots faced. A weighted league-average save percentage over his career is 0.904. What's the probability that a league-average goaltender could have a career as good as Brodeur's (or better)? Doing the same calculations gives a z-score of 4.309, and the probability is about 0.0008% (or about one in 122,000).

Another way of looking at this is this - based on the odds calculated above, it's possible for an average NHL goaltender to come along every few years and have a single season as good as Brodeur's. But doing it year after year after year? That's exceedingly unlikely. Good times!

Note that this method has several flaws - the first one being that all shots faced are equal in terms of difficulty to one another (a requirement to use a binomial distribution). But it is what it is, as Todd Bertuzzi might say.

 

[Ol Dirty Bstrd]

Hasek-Roy-Brodeur in that order.

Brodeur really benefited on his team with the defensive style of play. Hasek in his prime was the best of all three.

Yes, I pointed that out, and it is also the point I am trying to refute. I will reiterate myself:

1. Roy also benefitted, although maybe to a lesser extent. Blake, Foote, etc. were no slouches either.

2. Brodeur had arguably his best season, in which he won a Vezina and had Hasek like numbers more than halfway through the year, behind the weakest defense of his career. It was also post lockout, which was aimed to increase scoring. THis shows that Marty is the best at adjusting his style to his team (he is not jsut successful with a low amount of poor quality shots)

3. Marty struggled early this year but has adjusted once again. This defense is even worse than last years, and he had to deal with a new system which would affect anyone. While the haters claim the new offensive system has slowed him down, it appears he will prove the doubters wrong once again as he only needed in adjustment period (we wont know for sure until the end of the year). Bear in mind, Marty played 2 games with this defense:

Greene-Rachunek
Vishnevski-Brookbank
Mottau-Oduya

That is the worst defensive corps Ive ever seen. Greene, an undrafted rookie with under 50 games of experience was the #1. Vishnevski and Rachunek are no more than third pairing defensemen. And the other three are, or at least should be, career AHLers. In those 2 games, Marty let up 3 goals, one of which was directly attributed to Mottau falling for no reason giving Vasicek a good 5 seconds to stand in frotn of Brodeur and deke back and forth before it finally trickled in (no joke it was that long). The following two games White who returned for the first time from a career threatenign eye injury replaced Rachunek and Brodeur only let up 1 goal in those two games, once again directly caused by an atrocious giveaway by a career AHLer (Oduya turnover ot Malkin). So he let up 4 goals in 4 games behind a defense that Roy couldn;t even dream up in his worst nightmare.

TO sum it up, it doesn't matter the team Marty has in front of him, HE WILL ADJUST. And thats what makes him the best or close to it. So everyone who complains about his defense, explain last years vezina trophy or this year strech of 4 GA in 4 games behind an AHL defense.

 

[pitseleh]

Hasek/Roy is one of those debates I flip flop on all the time. Roy was my favourite player when I was growing up and a tremendous post-season competitor but I don't think I ever feared a player as much as Hasek in his prime.

Personally, I think that Brodeur is the clear third.

 

[Bear of Bad News]

Here's another thread where I talk about some weird ways I have of doing this: http://hfboards.com/showthread.php?t=381137

It's less math heavy, and probably more accessible to a general audience.

 

[Ol Dirty Bstrd]

No problem! It's a bit of an esoteric calculation, I must admit.

Basically, it's a method that you can use to estimate the probability that a goaltender is an average-ability netminder (obviously, none of the three goaltenders in this conversation is average by any stretch of the imagination).

I'll give a small example to explain how it works - suppose the league-average save percentage is 90%. Goaltender X makes three saves on three shots. To know if he's better than league average, we can calculate the probability that a league-average goaltender could make three saves on three shots simply by random chance.

This probability would be: (0.9)^3 = 72.9%. In other words, it's very likely that an average goaltender could do this. We would need more evidence to conclude that Goaltender X is better than average.

Now let's look at a real season - Martin Brodeur's 2006-07 campaign. Brodeur allowed 171 goals on 2182 shots for a 0.922 save percentage; the league-average save percentage was 0.905. What's the probability that a league-average goaltender could have (by random chance) put together a season as good as Brodeur's in 2006-07? Exactly Brodeur's season would be Comb(2182,171)*(0.905)^(2182-171)*(1-0.905)^(171), and then you'd have to add in every season better (170 GA, 169 GA, 168 GA, et cetera). It's a pain.

Fortunately, we have the normal distribution approximation to the binomial distribution. For large number of shots faced, we can approximate the above calculation by using the normal distribution (bell curve) with mean equal to shots*save percentage and variance equal to shots*save percentage*(1-save percentage).

In the case of Brodeur's season above, the normal approximation gives us a z-score of z=2.633, meaning that Brodeur's campaign was 2.633 standard deviations better than average in 2006-07. The probability that an average goaltender could do what Brodeur did? About 0.42% (or one in 238).

Now we can look at Brodeur's career as a whole. Through 2006-07, Brodeur had allowed 1931 regular-season goals on 22167 shots faced. A weighted league-average save percentage over his career is 0.904. What's the probability that a league-average goaltender could have a career as good as Brodeur's (or better)? Doing the same calculations gives a z-score of 4.309, and the probability is about 0.0008% (or about one in 122,000).

Another way of looking at this is this - based on the odds calculated above, it's possible for an average NHL goaltender to come along every few years and have a single season as good as Brodeur's. But doing it year after year after year? That's exceedingly unlikely. Good times!

Note that this method has several flaws - the first one being that all shots faced are equal in terms of difficulty to one another (a requirement to use a binomial distribution). But it is what it is, as Todd Bertuzzi might say.

THanks for taking the time to explain that. Must have been a pain to type out, but it helped alot!

 

[Bear of Bad News]

THanks for taking the time to explain that. Must have been a pain to type out, but it helped alot!

Always my pleasure. I was a math professor for nine years, so I like digging into this stuff.

 

[SoundwaveIsCharisma]

As far as I understand, the consensus is Roy is the best of these three, then Hasek, the Brodeur. I never understood this thinking. I will try and break down my thinking.

Roy vs. Brodeur - the same old Brodeur excuses cant be used to downplay him because Roy played behind some stellar teams as well. Also, Brodeur has a slightly higher career sv%, so Brodeur haters can throw that out the window. So Roy is generally considered as better because of his playoff success. While I agree, I will just give some numbers to show brodeur isnt too far behind:

Win% SO SV% GAA Cups Conn Smythes
Roy .616 23 .918 2.30 4 3
Marty .582 21 .921 1.89 3 0

The only category Roy blows Marty away in is Conn Smythes and Marty EASILY could have won it any three times he won the cup especially when he was robbed of it by Giguere. Here are his stats in those playoffs:

Year SO GAA SV%
94-95 3 1.67 .927
99-00 2 1.61 .927
02-03 7* 1.65 .934
*record for shutouts, including 3 in finals

Those are phenomenal numbers and the "good team" excuse is only so valid because look at his save percentages. For example, he lost in the first round in 01-02 with a 1.42 GAA and a .938 sv%, so his team in front of him couldnt have been that good at all. I could see some people picking Roy for one game as more of a "money goalie," but Brodeurs postseason numbers are pretty close to Roy's and he was slightly better in the regular season. Keep in mind ROy never even started 70 games. Marty's durability and puckhandling should definitely be taken into account.

As the numbers show, its pretty close so I can see people taking ROy over Marty. But heres the real headscratcher. People then rank Hasek in between Roy and Marty. Hasek in his prime was the most dominant goalie ever, easily better than both of them. But his prime was very short and that causes you to choose longevity vs. prime when discussing these goalies. How can someone choose longevity (or clutchness) when comparing Roy and Hasek, but choose prime when comparing Hasek and Brodeur. Keep in mind Brodeur and Roy BOTH played behind great teams, while Hasek didn;t for the most part. Also, once again Hasek only started 70 games once while Marty does it ever year. Brodeur had Hasek like numbers through January last year, leading the league in every major goaltending category, but those numbers fell because he started many more games than the dominator typically would in his career.

Sorry for the rant, but do I make valid points? Does Brodeur get downplayed by some? Can someone without bias explain the Roy then Hasek then Broduer logic? THanks.

I think the biggest thing for me is that in two cases Roy just had magical post-seasons. That aura of invincibility he had during the 86 and the 93 Stanley Cup runs was just insane, I don't think either Hasek or Brodeur have had such flat out dominating post-seasons from a goaltending point of view.

 

[The Thomas J.*]

We can go on & on forever who was better when & for how long, I say throw the math equations wins & loss totals out the window, Game 7 of any playoff series game on the line, going into OT , who would you rather have? I'll take Roy.

 

[ALine9900]

We can go on & on forever who was better when & for how long, I say throw the math equations wins & loss totals out the window, Game 7 of any playoff series game on the line, going into OT , who would you rather have? I'll take Roy.

Well, if you were starting a franchise which would you take?

 

[NOTENOUGHJTCGOALS]

Well, if you were starting a franchise which would you take?

Roy. Egotistical maniac, but he was a huge driving force in what made his teams so competitive for years.

 

[The_Hockey_Guy18]

Regular Season

1. Hasek
2a. Brodeur
2b. Roy

Playoffs

1. Roy
2. Hasek
3. Brodeur

 

[Augustus Pablo]

Playoffs

1. Roy
2. Hasek
3. Brodeur

Hasek above Brodeur for playoffs

 

[ALine9900]

Roy. Egotistical maniac, but he was a huge driving force in what made his teams so competitive for years.

 

[Bear of Bad News]

Hasek above Brodeur for playoffs

It's certainly defensible. What makes it so hilarious to you?

 

[Blades of Glory]

The thing that makes me favor Roy is his passion. He may be full of himself, cocky, and egotistical, but his teammates ALWAYS could rely on him. I remember Kirk Muller or Guy Carbonneau after the 1993 Cup Finals saying that "Patrick took over the dressing room during the Finals."

It paid to have a goalie like Roy behind you when it mattered most. Brodeur is no slouch, with an amazing playoff resume. Hasek is Dominik ****ing Hasek, easily the most dominant of the three, in the regular season.

If I had one game, with everything at stake, there's not even a doubt which goalie I would want. Remember Game 6 of the 2001 Cup Finals, when the Avalanche were down 3-2 and could have been eliminated at New Jersey? Roy played one of the best games of his career, shutting out the powerful Devils, and followed that up with an oustanding Game 7.

Non of the goalies had a playoff run like Roy in 1993. Along with Giguere's 2003 and McLean's 1994, I consider it among the greatest performances of the past 20 years.

Roy was money. He was big, cocky, and IMO, instilled a sense of fear in opponents, especially in the playoffs.

Roy never had any Hall of Fame defensemen in front of him. He won 3 Conn Smythe trophies, and that's the final straw for me.

Hasek won the Stanley Cup on maybe the most stacked hockey team in NHL history, the 2001-02 Detroit Red Wings. They had 8 Hall of Famers. Nothing special about that in my mind. He was well past his dominant days of the late 1990's.

 

[reckoning]

Hasek benefitted from the Sabres defensive style of play.

Huh??? The Sabres had one of the worst defences in the league during Hasek's time there in the 90s. In `96 and `98 he faced more shots per game than any other #1 goalie in the NHL.

 

[Big Phil]

It's certainly defensible. What makes it so hilarious to you?

Hey Hasek was no choker in the playoffs but Brodeur has won three Cups to Hasek's one. In a Game #7 I'll still take Brodeur over Hasek.