Sunday, August 16, 2009

Brodeur vs. Luongo, Part 2

Bruce over at Copper & Blue has laid out the case for Martin Brodeur as the best goalie of the 2000s. He picks Brodeur as the best combination of quantity and quality over the given time period, and describes him as a "win machine".

Brodeur's teams have won a lot of games, that is undeniable. However, the problem with using wins is that you are giving individuals credit or blame for team results. The best goalie is the one who did the most to help their team win, not the one who had their team win the most. I'm not picking Luongo as the best goalie of the 2000s because he ranks first in a few random statistics that I happen to like, I'm picking him because I think he did more than anyone else to help his teams win.

I have a method for quantifying a goalie's expected wins, based on their goal support, shots against, and team discipline. It compares the goalie's team to league average in each of those categories, assesses how many goals they scored or prevented compared to average, and then converts that goal differential into wins. Here are the averages of the key components during the period:

Brodeur's teams: 232 goals, 25.9 shots against/60, 298 power plays against
Luongo's teams: 207 goals, 31.4 shots against/60, 403 power plays against
League average: 229 goals, 28.3 shots against/60, 370 power plays against

For example, since Brodeur's teams averaged 25 more goals per season than Luongo's teams, we would expect that to translate into more wins. The standard rule of thumb is that a 6 goal differential is equivalent to one win, which means that Brodeur's teammates' offence is worth about 4 extra wins per season.

It's not too difficult to see from those numbers that it was easier to win games in Brodeur's situation than in Luongo's. Brodeur had a significant advantage in all three categories. I calculate Brodeur's teams as having 410.1 expected wins and 465.0 actual wins, counting each tie or shootout/OT loss as half of a win. Luongo's teams had 296.1 expected wins and 353.5 actual wins. If we subtract the two numbers, we can see how each team did relative to their expected numbers. Brodeur's teams won 54.9 more games than expected, while Luongo's won 57.4 more. It is even more impressive that Luongo's teams outpaced Brodeur's teams since Luongo played 80 fewer games. That's almost an entire season's worth of weaker backup goalies in net, which would likely substantially drag down the overall results.

If we convert the expected numbers into winning percentages, we can compare directly with each goalie's actual winning percentage to remove games played by backups from the equation. Brodeur's winning percentage was .630 while his teams had an expected .556, for a difference of .074. Luongo's winning percentage was .498 compared to .401 expected (+.097).

That's not the entire story, however, as there are two major variables we have to still account for. One is shootouts, and the other is Brodeur's shot prevention effect.

From 2006-2009, Brodeur's teams have gone 33-18 in shootouts, while Luongo's teams are just 18-25. If we dig deeper, we see that Luongo and his backup goalies actually had a better combined shootout save percentage (.696) than New Jersey did (.679). The difference in win/loss record is entirely because Brodeur's teammates scored on 40% of their shootouts while Luongo's teammates scored just 28% of theirs.

To factor this in, we need to adjust the expected win totals to reflect the rest of the team's shootout performance. Brodeur's teams won 7.5 more shootouts than average, while Luongo's won 3.5 fewer. Adding that to the win totals, we get Luongo's teams at 60.9 wins above expected and Brodeur's at 49.9.

Finally we come to the tricky issue of shot prevention effects. It's much tougher to assign a specific number here, so I'm going to rely on a sensitivity analysis. The goal is to identify the shot prevention effect Brodeur would need to surpass Luongo, and assess whether it is reasonable that the actual difference was that large.

First, let's look at how New Jersey's overall team wins above expected would vary based on different estimates of shots against prevented by their goalie:

1 shot prevented per game: +58.9 wins
2 shots prevented per game: +70.4 wins
3 shots prevented per game: +81.9 wins

Since Luongo's teams were at +60.9, the goaltender shot prevention difference between the two teams has to be 1.2 or more for Brodeur's teams to end up ahead. I'd say it is reasonable that Brodeur's effect relative to Luongo's is that high or higher. Keep in mind though that we are comparing Brodeur's teams to Luongo's teams, and Brodeur played in 80 more games. With these numbers it is certainly possible to make a case that Brodeur provided more total value than Luongo over the decade, but let's look at the individual numbers for each goalie:

Luongo: .396 expected, .498 actual, +.102

Brodeur:
1 shot prevented per game: .550 exp win %, .630 actual, +.080
2 shots prevented per game: .535 exp win %, .630 actual, +.095
3 shots prevented per game: .519 exp win %, .630 actual, +.111

Here the break-even point is 2.5 shots per game. I'm willing to concede a shot prevention difference between them of 1.0 - 1.5 shots per game, and maybe even as high as 2 shots per game, but 2.5 seems a bit high to me. As a result, my interpretation of these numbers is that on a per-game basis Luongo was slightly more likely to help his team win than Brodeur. To be sure, it is very close. When the numbers are close, we can also go to qualitative factors to break the tie. I generally put a high weighting on peak play in my rankings, and in my opinion in the 2000s Luongo's peak was higher than Brodeur's.

Another consideration is that each marginal goal becomes slightly less valuable the more a team's expected winning percentage deviates from .500, . A goalie on a very poor or very strong team would not see each additional goal saved have the same impact on winning, because it is more likely to come in a blowout game. On the other hand, a goalie on an average team will spend a lot of the time in tied or close games, which means that their effect is magnified. After factoring in shot prevention, Brodeur's expected winning percentage is close to .500, which means that his marginal goals saved likely had a strong impact in terms of winning games. On the other hand, Luongo's expected winning percentage was just .396. I don't think this has a large effect on the numbers, but it is something that would be a slight advantage to Brodeur.

Depending on your assessment of the method and goalie shot prevention effects, you may agree or disagree with the evidence that Roberto Luongo helped his team win more games than Martin Brodeur. I'm still working with the method so some of the assumptions may need to be revised slightly to get a better fit. Regardless, I think this exercise should be strong evidence against the view that Brodeur far surpassed all his goaltending peers between 1999-00 and 2008-09. If Brodeur can be described as a win machine, then the same thing should be said about Roberto Luongo.

Lawrence said...

This is an interesting breakdown, and you've obviously put a lot of good thought into this. Here's my simple beef, and why my own opinion aligns more with Bruce's. It's the team idea. Obviously good teams make it easier on a goalie than terrible teams, but I also think that a good goalie helps create this good team. Without getting bogged down by the numbers argument too much, I think it is unfortunate that Martin Brodeur will always have an asterix beside his name in the minds of hockey's greatest numbers fans. *Well, he did have one of the most dominant records and list of performances of all time, but really if we look at the numbers more closely, he really was only slightly better than average on a good team for a very long time.

The thing with Martin Brodeur that I find most impressive is that this is the only complaint you can find in him, numbers or not. He quite honestly is good or great in every single association with being a goalie. More so, in my mind than anyone else, and it can be argued against of course, but even face value from stopping pucks, to winning games, to statistical dominance, to championships, to international play, to championing and using reduced equipment sizes, to all the 'additionals' - puck handling, character, loyalty and unselfish behavior (not having an agent), he is just great...very very great. Of course for every Gretzky fan there will be a Lemeiux lover, who says "what if" but as with Gretzky...there is no if, he just did do it, as with Brodeur.

Obviously Luongo is amazing, and as I raved about Brodeur, it was then Luongo, and now it is Lehtonen added to the list, and soon ..., but the simple fact is Martin Brodeur will be the goalie every other goalie will be and is measured against for a very long time including Roberto Luongo, surpassing Patrick Roy and Hasek and that alone is enough to convince me.

The cycling equivalent would be Eddy Merckx, he won everything and dominates all disciplines. Eddy Merckx and Martin Brodeur are the bar.

Anonymous said...

I don't agree with either Bruce or you, but of the two I would have to give the nod to Luongo, albeit by only a slight margin.

With the exception of one playoffs (2001), Brodeur has been utterly rock-solid and often well above-average. His durability, consistency, and long peak must count for something. On the other hand, as you have already pointed out he has played on much better teams than Bobby Lou ever has.

I do give RL a lot of credit--he has seen stretches where he was literally the only good player on his team. I don't deny that he has more raw talent than Marty. Still, I'd like to see a little more outstanding clutch play out of Bobby Lou and fewer meltdowns when he is needed most.

(For me, J-S Giguere has to be goaltender of the decade.)

Anonymous said...

lol at giguere, that anonymous fanboy is still around?

The Contrarian Goaltender said...

"Obviously good teams make it easier on a goalie than terrible teams, but I also think that a good goalie helps create this good team. Without getting bogged down by the numbers argument too much, I think it is unfortunate that Martin Brodeur will always have an asterix beside his name in the minds of hockey's greatest numbers fans."

I agree that a good goalie helps his team. Brodeur certainly helped his teams, otherwise we'd have seen New Jersey win just about what they would have been expected to given their goals for, their shots against, and their team discipline. But I think you do have to "get bogged down by the numbers" because otherwise you really have no idea how much Brodeur actually contributed, and the argument becomes an unresolvable tradeoff of vague estimates and platitudes.

This analysis focuses on the main things that the rest of the team does, and then assumes that the rest of the effect must be the goalie. How did Brodeur help his team? Did he score goals? Did he cause his team to take fewer penalties? If you answered no to both of those questions, then the only team adjustment that Brodeur might have a hand in would be the shots against number, and I threw up a range of shot prevention numbers for you to make your interpretation on that one.

"He quite honestly is good or great in every single association with being a goalie."

Martin Brodeur does have a diverse skill-set. But it does not necessarily follow that an all-arounder is better than a very accomplished specialist. When you compare Hasek to Brodeur, for example, Brodeur might very well be better at most goaltending skills. However, Hasek's ability to stop the puck was on a completely different level. Since that is the singularly most important job of a goalie, that was also what made Hasek the superior goaltender.

Similarly, the evidence suggests that Luongo is also better at stopping the puck than Brodeur, not nearly to the same degree as Hasek of course, but enough that there is a legitimate debate between the two netminders from Montreal.

"Martin Brodeur will be the goalie every other goalie will be and is measured against for a very long time including Roberto Luongo, surpassing Patrick Roy and Hasek and that alone is enough to convince me."

It may be true that the general hockey community sees Brodeur as the bar. It already happened once that popular consensus latched onto the goalie with the highest career numbers (Terry Sawchuk), despite evidence from the statistical record and contemporary accounts indicating that he was outplayed by other goalies during the same time period (Plante, Hall). It's probably going to happen again with Brodeur, and once again that doesn't mean they are right. I just don't understand how hockey fans who watched hockey during the 1990s could possible see Martin Brodeur as the guy setting the bar instead of Dominik Hasek.

Lawrence said...

"But I think you do have to "get bogged down by the numbers" because otherwise you really have no idea how much Brodeur actually contributed, and the argument becomes an unresolvable tradeoff of vague estimates and platitudes."

This though, is where 'numbers' get in the way of the big picture. The drill-hole analysis vs. the objectives and systems. No argument will be lucid enough to convince me the every season the objective is to win the Stanley Cup. It doesn't matter how, with what stats, with who, is winning the cup.

My goalie, hasn't won the cup, so he's not the best. Brodeur has. in the 2000's twice. You can't honestly have a conversation about the best goalie of the 00's without saying a requisite is the Championship. That leaves 7 goalies, of which Luongo is none of. From there, you can look into the numbers at more detail, but you can't confuse the how with the why. Otherwise you're arguing that Luongo's minuscule one tenth of a percent better stat is more important than a Stanley Cup...or two. I don't think so. I'll trade a .915% for a .905 and a cup any day. You have to beat the best to be the best, and that means winning. Winning everything. Then we'll talk drill-hole stats and contribution.

Lawrence said...

Whoops, that should say: No argument will convince me that the objective in every season is NOT to win the Stanley Cup.

The Contrarian Goaltender said...

Winning the Stanley Cup is the objective every season, of course. But there are two important things to keep in mind.

One, the Cup is a team objective that requires 20 guys. No player can win the Stanley Cup on his own. Judging any individual on any team result in any setting is always unfair.

Two, "winning the Cup" is not a process, it is an end result. Hockey players in a game aren't "winning the Cup" any more than a research chemist in a lab doing an experiment is "winning a Nobel prize". Teams play the games, and if they win all the series then they get awarded the Cup. Therefore the true specific objective for every team is to win enough games to qualify for the playoffs and then to be the last team standing. If you don't do that, then you don't win the Cup.

Hockey games are won by goals for and against, and therefore players help their teams win by scoring goals or preventing goals against. Teams that have players that are very good at either or both of those things are more likely to win games, playoff games, and championships.

So we have identified a result (the Cup) and the way to achieve that result (win hockey games by scoring/preventing more goals than the other team). Each of us has to then decide what makes more sense as a basis for evaluation, either to judge players based on whether they and 19 others collectively achieve a team result, or to evaluate players on whether they are individually successful at the processes that directly contribute to the overall team result.

"You can't honestly have a conversation about the best goalie of the 00's without saying a requisite is the Championship."

Of course I can. I don't care how good a goalie's teammates are, I care how good he himself is.

Who was the best goalie of the 1990s? If your answer is not Dominik Hasek, then your answer is wrong. How many Cups did the Dominator win in that decade? Zero.

"Otherwise you're arguing that Luongo's minuscule one tenth of a percent better stat is more important than a Stanley Cup...or two."

If Luongo save percentage shows that he played better than Brodeur, then that is more important than the fact that Brodeur's team won the Cup. Absolutely.

Again, what is important is how much the player individually helped his team win the Stanley Cup. Whether the goalie's teammates were good enough to close the deal is irrelevant.

If the goalie individually failed in an important situation then you can use that in an argument against them, if you interpret that result with a proper consideration of team context and sample size (something like "LUONGO is a CHOKER, he let in 7 goals in ONE GAME!!11!11!!!" doesn't quite fit the bill). Quite simply goalies don't win hockey games (and thereby Cups), hockey teams do.

"I don't think so. I'll trade a .915% for a .905 and a cup any day."

That trade isn't the straight-up trade you make it out to be. If you trade a .915 guy for a .905 guy, you're making your team more likely to get scored on, and therefore less likely to win hockey games, and therefore less likely to win the Cup.

What you're really saying is that I would trade the .915 goalie's team for the .905 goalie's team. Unfortunately, that doesn't tell us an awful lot about the two goaltenders involved.

Marc-Andre Fleury has a Cup while Roberto Luongo does not. If Vancouver traded Luongo for Fleury, they are a worse team next year. Fleury doesn't win the Cup in Vancouver unless the trade also includes Crosby and Malkin. Somehow I don't think that's too likely to happen.

Unknown said...

You know what Luongo hasn't done? Won anything of significance. You can throw all the stats out you want, but it won't change that fact. Sorry chief.

Anonymous said...

"One, the Cup is a team objective that requires 20 guys. No player can win the Stanley Cup on his own. Judging any individual on any team result in any setting is always unfair."

Two goalies--exactly two--have managed to very nearly win the Cup on their own: Hasek and Giguere. You can't say that about Luongo, Brodeur, Patrick Roy, etc. Both of those goalies won the Cup the very first year they got on truly good teams. Vanbiesbrouck, Kolzig, Irbe, etc. had great runs also, but they never brought their teams as close as the above two minders--who are I think, quite obviously, the #1 and 2 greatest goaltenders of all time.

"Who was the best goalie of the 1990s? If your answer is not Dominik Hasek, then your answer is wrong. How many Cups did the Dominator win in that decade? Zero."

Agreed completely. Likewise anyone who answers that somebody other than J-S Giguere is the best goalie of THIS decade is also wrong. Yes, the Dominator was a better goalie, but he had peaked already by the '00s.

"Quite simply goalies don't win hockey games (and thereby Cups), hockey teams do."

I will agree with you that Bobby Lou is better than Brodeur, but it is UNDENIABLE that he cost his team Game 6 vs. Chicago. Vancouver outplayed the Hawks that game, scored five goals, and took a lead into the third period.

Anonymous said...

Whats with this one random dude coming in and calling Giguere one of the all time greats like its some kind of given fact? Giguere is not even part of the discussion. get over it. you seem to be the only one who thinks he is.

Anonymous said...

Yes, yes, why not go offer a sacrifice to your god, Brodeur. If Giguere had played on the teams Brodeur did the first half of the '00s, he would have three or four Cups. I can promise you that CG would name Giggy goaltender of the decade long before your god.

Anonymous said...

So again, despite never being named the best goalie in the league by anyone EVEN for a single year, he still somehow comes out as the best of the decade? Its amazing what 3 or 4 good seasons do for a guy in the minds of some. You have a better case for calling Olaf Kolzig the best goalie of the 90's.

And funny you bring up Luongo's game 6 against Chicago as some big knock, but what do you call it when a guy was not even good enough to be on the ice during the playoffs... For the second time in his career?

The Contrarian Goaltender said...

"I can promise you that CG would name Giggy goaltender of the decade long before your god."

Hmm. That's a bit presumptuous. I don't think I would do that all, actually. Even if I somehow did find evidence in favour of Giguere I'm positive it wouldn't put him "long before" Brodeur.

Giguere is absolutely in the conversation when you are talking about the best goalie of the 2000s and his overall EV SV% record is very strong, but I think based on reasonable assumptions of the other things that Brodeur brings to the table that Marty would have the edge.

I don't put a huge weighting on playoff performance, so unlike you I don't think Giguere's 2003 is his passport to greatness. It was a terrific postseason, to be sure, but even if you do look at playoff play above all else I think you have to look at the whole picture, and Giguere unfortunately gave back a lot of his 2003 gains in 2006, 2008 and 2009.

Anonymous said...

Perhaps now that rat turd has been put to rest "with his god" gigeure

Anonymous said...

Contrarian, by your own usual measuring stats, Giguere has solidly outperformed Brodeur. He has a higher overall save percentage than Brodeur and with the exception of two years tops played on teams vastly worse than his. It's not much of a contest. Brodeur only is better than Giguere when it comes to counting and team-dependent stats. Yes, he is a better puckhandler and has better rebound control, but that doesn't make a whole lot of difference if the goalie who has inferior puckhandling and rebound control is more likely to make those saves.

If you think Brodeur is better than Giguere overall, logically speaking you must concede that he is also better than Hasek (which I do not agree with).

Anonymous said...

Wait... so since when does Gigueres 3 or 4 years make him equal to Hasek. Somebody is out of their mind. Giguere and Hasek have nothing in common outside of one playoff run. Not only that, but both seem to be overrated due to their "one miraculous playoff run". why is it that goalies on good teams deserve no credit for winning, but goalies on decent but not spectacular teams get all the credit for their teams runs? Roloson overall has a better case than Giguere, especially going by the current logic, that he played for a worse team than Giguere did in 2003. How about Olaf Kolzig as well. He Actually won a Vezina on top of his run to the finals.

Kent W. said...

I'll trade a .915% for a .905 and a cup any day. You have to beat the best to be the best, and that means winning. Winning everything. Then we'll talk drill-hole stats and contribution.

Would you employ this line of reasoning for skaters as well? Eric Godard won a cup last year. Jarome Iginla never has. I can cite you 50 "drill-hole" stats that prove Iginla is a superior hockey players,but would you throw them out because "Godard was won the cup"? I doubt it.

If you ignore the contributions made by players in winning games in favor of the ultimate outcome, you invert cause and effect.

The Contrarian Goaltender said...

"If you think Brodeur is better than Giguere overall, logically speaking you must concede that he is also better than Hasek (which I do not agree with)."

Are you talking about over their careers or in the last decade? There is little question that Brodeur's entire career beats Giguere's. Over the last decade is where there is an argument.

Since '99-00 Giguere has been .927 at EV and .872 on the PK, while Brodeur is .922 at EV and .876 on the PK. It does look like Giguere's the better puckstopper, but Brodeur likely closes most of that gap through other skills. Giguere might still edge him, but at that point it's tough not to pick the guy with 50% more minutes played.

The main point is it is close. If you want to take Giguere over Brodeur for the 2000s I'm not going to argue too hard with you. But I think you're overselling the difference between them.

I also wonder a bit about Giguere's team situation. Despite his strong results, he has actually been outplayed by his backups since 2002-03:

Giguere: 2.48, .915
Other ANA goalies: 2.32, .917

That at least raises the possibility of either some shot quality effect or shot bias or something going on there, although it could just be goalie talent since most of Giguere's teammates were pretty good. Still, you'd ideally like to see the starter outperform his teammates, and those numbers don't really suggest that Giguere played for especially weak teams.

Question for you: You've compared Giguere to Hasek a number of times, but did you ever actually watch late '90s Hasek? He was on a completely different level than J.S. Giguere. That's something that everyone who has watched both of them will tell you, and the numbers back that up.

Giguere played one memorable playoff at .945. Hasek played an entire season at .937 in a similar scoring environment. This with Hasek facing 23% of his shots on the penalty kill during the 1998-99 regular season, compared to Giguere facing just 11% of his shots in the 2003 postseason with his team down a man.

And that season was not even Hasek's best or second best. That's just how good the Dominator was.

Anonymous said...

"Are you talking about over their careers or in the last decade? There is little question that Brodeur's entire career beats Giguere's. Over the last decade is where there is an argument."

I meant the decade, though I do believe that overall Giggy is better too. I will grant Brodeur great longevity, durability, and a pretty long and late career peak. Giguere may well not be able to match that. That does not mean he is the less-skilled netminder.

"Since '99-00 Giguere has been .927 at EV and .872 on the PK, while Brodeur is .922 at EV and .876 on the PK."

Brodeur has had much better penaltykillers most of that time!

It does look like Giguere's the better puckstopper, but Brodeur likely closes most of that gap through other skills. Giguere might still edge him, but at that point it's tough not to pick the guy with 50% more minutes played."

You are favoring Brodeur purely on counting stats here. It's not fair to use counting stats in favor of Brodeur but dismiss them in the case of Roberto Luongo, who has a lot less to show for himself than Giguere. Use and stick to a consistent judging standard. You have said many times that additional games played after about 50/season do not matter much, so you shouldn't hold minutes against Giggy.

"I also wonder a bit about Giguere's team situation. Despite his strong results, he has actually been outplayed by his backups since 2002-03"

I will grant you that this is a fair and legitimate question. In the case of Martin Gerber, I think it's pretty safe to say that that was pure luck over a small sample size. Bryzgalov is trickier as the guy is outstanding at times, but doesn't seem to be very mentally tough. The only backup who at this point I would call pretty unequivocally better than Giggy is Jonas Hiller. Maybe Giggy has already peaked. It's hard to say.

"Still, you'd ideally like to see the starter outperform his teammates, and those numbers don't really suggest that Giguere played for especially weak teams."

Giguere has played on one elite team ('07 Ducks), one good but not elite team ('06 Ducks), and one defensively decent but offensively anemic and overly penalized team ('08 Ducks). All of his other teams have ranged from okay to pretty bad. Marty B has played for elite teams half the decade, and he has never been on a team worse than okay.

"Question for you: You've compared Giguere to Hasek a number of times, but did you ever actually watch late '90s Hasek? He was on a completely different level than J.S. Giguere. That's something that everyone who has watched both of them will tell you, and the numbers back that up."

In the 2003 run Giguere employed a number of Hasek-like flopping and rolling moves. Over that run he was not at all far off from the Dominator in his prime. I also do think that in their two very most similar series ('99 Senators and '03 Wings) that the Wings were a stronger foe.

"Giguere played one memorable playoff at .945. Hasek played an entire season at .937 in a similar scoring environment. This with Hasek facing 23% of his shots on the penalty kill during the 1998-99 regular season, compared to Giguere facing just 11% of his shots in the 2003 postseason with his team down a man."

I never even once suggested that Giggy is better than Hasek or near his level. I think that Giggy, at his peak, was the closest of any goalie to Hasek at his peak, but never implied that they were at the exact same skill level.

"And that season was not even Hasek's best or second best. That's just how good the Dominator was."

I have never even once disputed that the Dom was the greatest goalie ever.

Statman said...

I like Lawrence's & Chris' evaluation methods... Stanley Cups are the only thing that counts... so much more clear-cut than all this fancy math stuff.

For instance, now I know that 4-time Cup winner ('76-79) Rick Chartraw (420 career reg season games, 92 career reg season pts) is far better than Marcel Dionne, Gilbert Perreault, Brad Park, Borje Salming, Darryl Sittler, Peter Stastney, Dale Hawerchuk, & Pat Lafontaine combined (9,000++ career reg season points but 0 Cups total).

The Contrarian Goaltender said...

"Brodeur has had much better penaltykillers most of that time!"

That's certainly debatable.

1999-00 to 2008-09:
Anaheim: 83.4% on the PK
New Jersey: 84.2% on the PK

New Jersey is only ahead because the 2000 Ducks were the worst team in the league on the penalty kill while the 2000 Devils were 3rd. From 2000-01 to 2008-09 both teams were at 83.8% on the PK.

Let's look at how some of the other Anaheim goalies did on the PK compared to when they left town:

Bryzgalov in ANA: 327/364, .898
Bryzgalov in PHX: 541/644, .840

Gerber in ANA: 219/244, .898
Gerber in PHX: 801/939, .853

Hiller in ANA: 404/453, .892

Looks to me like the Ducks were pretty good on the penalty kill.

I see Giguere's PK play as a legitimate weakness. In a 2500+ shot sample his results are very average, even though his teammates' numbers are terrific. Giguere's a very good even strength goalie, but I don't think he's all that good on the penalty kill. That makes sense to me given his positional, blocking style. If Luongo had Giguere's PK save percentage he would have let in an extra 70 goals. That's a pretty significant gap, and just another way that Luongo is better than Giguere.

"I think that Giggy, at his peak, was the closest of any goalie to Hasek at his peak, but never implied that they were at the exact same skill level."

We've been through this before, but you apparently define Giguere's peak as "the 2003 playoffs". Unfortunately that is not a peak, that is a 21 game hot streak. In his next 21 NHL starts after losing game 7 of the '03 Cup Finals, Giguere was 5-14-2, 2.76, .910. Sometimes you get hot, and sometimes you get cold.

Pick a longer period to define his peak, like say a season or two, and there are certainly others who have peaked higher than Giguere in the 2000s. Kiprusoff, Luongo, Theodore, and Thomas would be fairly uncontroversial choices, and there are others who are in Giguere's range like Vokoun, Roloson, Cechmanek or even late-career Patrick Roy.

Anonymous said...

"New Jersey is only ahead because the 2000 Ducks were the worst team in the league on the penalty kill while the 2000 Devils were 3rd. From 2000-01 to 2008-09 both teams were at 83.8% on the PK."

RECENT Ducks teams have been pretty defensively sharp, but are you going to tell me that the juggernauts that were the '01, '02, and '03 Devils were no better than the so-so to awful Ducks teams in the same span on the PK?

"Let's look at how some of the other Anaheim goalies did on the PK compared to when they left town"

I already told you how I think Martin Gerber just got lucky one year. Bryzgalov is very good but just not tough (he did also post a .920 overall SP in Phoenix). Hiller probably is really better at this stage, but he has had only one full NHL season and playoffs to judge, too.

"I see Giguere's PK play as a legitimate weakness. In a 2500+ shot sample his results are very average, even though his teammates' numbers are terrific."

Fair enough and there is probably some truth to this, although at the same time, recent Ducks teams have taken a LOT of penalties, in contrast to the early '00s Ducks teams. Any team that takes that many penalties is going to get tired PKers and goalies.

"We've been through this before, but you apparently define Giguere's peak as "the 2003 playoffs". Unfortunately that is not a peak, that is a 21 game hot streak. In his next 21 NHL starts after losing game 7 of the '03 Cup Finals, Giguere was 5-14-2, 2.76, .910. Sometimes you get hot, and sometimes you get cold."

Giggy had posted two straight years of .920 overall save percentage on a mediocre-to-poor team, and let me remind you that Hasek's '99-00 season was also significantly worse than his '98-99 season and single-handed Cinderella run also. In fact, he never was really the same after that year.

"Kiprusoff, Luongo, Theodore, and Thomas would be fairly uncontroversial choices, and there are others who are in Giguere's range like Vokoun, Roloson, Cechmanek or even late-career Patrick Roy."

Kiprusoff had the benefit of a world-class defense for much of his peak, and I think it is pretty safe to say that Theodore just had a lucky year. Tim Thomas may be better, but he's just peaking now. He was not this good over the whole decade. Luongo is very good but has not been mentally tough at key times. Roloson is tough but in his best year was playing under the Jacques Lemaire trap of Minnesota. His Edmonton numbers have never been close to his Minn stats.

Lawrence said...

"Would you employ this line of reasoning for skaters as well? Eric Godard won a cup last year. Jarome Iginla never has. I can cite you 50 "drill-hole" stats that prove Iginla is a superior hockey players,but would you throw them out because "Godard was won the cup"? I doubt it.

Kent: no I wouldn't apply that same reasoning because that isn't even close to a fair comparison. That's comparing 400 odd goals to 2 or 3 (I haven't looked up the stats). What we are talking here is a difference between Brodeur and Luongo sv% of .005 (not even as large as the .010 I stated). That is 50 goals difference on 10,000 shots! You could just as easily discount this difference on shooters luck alone.

I would take Lecavalier over Thorton though, because Lecavalier can play in the playoffs and Thorton apparently cannot.

Regardless of this, I loathe the argument of team vs individual. There is no situation except Shootouts where a goalie plays as an individual. It's team from days one to the Cup.

As well, I think there is reasonable consideration that the strongest teams do not win the cup as often or less often than they do. There is always a Pittsburgh beatin Detroit, a San Jose upset, a miracle Calgary, Edmonton, Ottawa run (who all lost). It takes timing, luck, bounces, heart, goaltending. The years New Jersey won the cup they weren't even the best team. 2nd in the Atlantic (5th overall) and 1st (4th) and they beat Giguere who single handily carried them to the finals.

I'm sorry, but cups matter, whether it's end result or not. Besides, winning the cup is a goal, it's a vision of success and if you backcast from that vision your pick Lecavalier over Thorton (yes Kent, Iginla over Goddard) and you pick Brodeur over Luongo, like they did in the Olympics, World Cup, etc. etc. Luongo hasn't won anything important, not even the World Juniors, and that matters when you're talking, "Who is the Greatest?" vs "Who is really really Good?" .

Lawrence said...

I like Lawrence's & Chris' evaluation methods... Stanley Cups are the only thing that counts... so much more clear-cut than all this fancy math stuff.

Statman, you are intentionally mis-representing my argument so you can discount it. I didn't say that at all.

the question is "who is the greatest of the 2000's?"

So get a list of your top ten stats wise... sv% with a minimum games played so on an so forth. Boil it down to ten let's say. then ask "Who of these ten has won the cup?" Boil it down again. Then really get into the detailed drill-hole numbers. End result = Brodeur.

If the goal of a team each season is to win the cup, then winning the cup has to count for something, otherwise great regular season performers and poor to average playoff performers are better than great playoff performers and good regular season performers?

Sorry bud, Fail.

It's not Iginla vs Goddard or Chartraw vs Dionne, that's just being obtuse.

How about reframing it like this.

Luongo and Brodeur have been chosen to represent Canada on the same team. Brodeur was chosen over Luongo, as the best, by the best, why?

Second, Canada won, remember? Twice, 2002 and 2004. You going to tell me that they made the wrong decision choosing Brodeur over Luongo? Or that you would choose differently? That's a pretty big statement implying that all those minds in Canada hockey don't know what the hell they are doing.

Bruce said...

Interesting post, CG. Where to start?

For example, since Brodeur's teams averaged 25 more goals per season than Luongo's teams, we would expect that to translate into more wins. The standard rule of thumb is that a 6 goal differential is equivalent to one win, which means that Brodeur's teammates' offence is worth about 4 extra wins per season.

My take is that Brodeur's teams spent less time in their own zone in significant part due to Brodeur's contributions to flow of play. More time out of their zone = more offence? You won't find much in stats like goalie points, but it seems reasonable. Luongo's teams, meanwhile, seem to spend a lot of time mired in their own zone. The difficult question of course remains: how much of that is due to the team and how much to the goalie himself and how do you disentangle the two, since the goalie is undeniably part of his team.

Brodeur's winning percentage was .630 while his teams had an expected .556, for a difference of .074. Luongo's winning percentage was .498 compared to .401 expected (+.097).

Even accepting your numbers here (which I don't), it is disingenuous to represent the results as a straight percentage points over expected. Yes, .097 > .074, but Luongo had much more opportunity to win "unexpected" games. With an expected Pts% of .401, Luongo had (theoretically) the opportunity to improve his team's winning percentage by .599. He "achieved" .097; ergo .097/.599 = 16.2%. The equivalent calculation for Brodeur is .074/.444 = 16.7%.

Brodeur's teams won 54.9 more games than expected, while Luongo's won 57.4 more. It is even more impressive that Luongo's teams outpaced Brodeur's teams since Luongo played 80 fewer games.

Interesting you somehow find a way to give Luongo credit for playing fewer games. But let's again apply those results to opportunity. Both guys' teams played 738 games. Given his team had 296.1 expected wins, Luongo's teams had the opportunity to improve on that by a whopping 441.9 wins, of which BobbiLu got 57.4, or 13%. Brodeur's team had the opportunity to improve on his team's expected performance by just 327.9 wins, of which MB30 got 54.9, or 16.7%.

Luongo: .396 expected, .498 actual, +.102

Brodeur:
1 shot prevented per game: .550 exp win %, .630 actual, +.080
2 shots prevented per game: .535 exp win %, .630 actual, +.095
3 shots prevented per game: .519 exp win %, .630 actual, +.111

I don't understand why you apply all the shot prevention effect to one side of the ledger (Brodeur's). At least an equal factor when comparing the two goalies is Luongo's very poor record of shot prevention; of the 21 goalies I studied who played at least 250 regular season and 20 playoff games 1999-2009, Luongo ranked 21st and last, facing fully 2.2 shots more per 60 than any other goalie on the list (31.7; 20th was Ryan Miller at 29.5) . That's a huge differential. I'd be interested to see calculations like the above calculating Luongo's teams expected Pts% assuming he faced 1, 2, and 3 fewer shots.

If Brodeur can be described as a win machine, then the same thing should be said about Roberto Luongo.

The original comparison was Brodeur as "win machine" compared to Wayne Gretzky as "point machine" cuz his margin over his peers in the raw counting stat was similarly wide. (e.g. Brodeur 1999-2009 = 420 wins; Luongo 1999-2009 = 241 wins)

The equivalent to your above statement might be "If Gretzky could be described as a point machine, the same thing should be said about Bernie Federko".

Bruce said...

If Luongo save percentage shows that he played better than Brodeur, then that is more important than the fact that Brodeur's team won the Cup. Absolutely.

Well doesn't that nicely sum it up. It all boils down to Save Percentage.

At best his Sv% might indicate Luongo stopped pucks better than Brodeur, but certainly not that he played better. The goalie is not just a two-dimensional cardboard cutout, he's an actual hockey player.

Lawrence said...

"Since '99-00 Giguere has been .927 at EV and .872 on the PK, while Brodeur is .922 at EV and .876 on the PK. It does look like Giguere's the better puckstopper, but Brodeur likely closes most of that gap through other skills. Giguere might still edge him, but at that point it's tough not to pick the guy with 50% more minutes played."

How is it fair to say this about Giguere and not about Luongo? The difference between the three is: L-.920, G-.916, B-.915 or 0,40,50 goals on 10,000shots. That's about 8,10 goals per season. And of course, Brodeur is the most complete 'other skills player' than both of them.

I don't put a huge weighting on playoff performance, so unlike you I don't think Giguere's 2003 is his passport to greatness.

This is a major point of contention. You're reducing the value of a large body of work to prove your point. Half of Luongo's worst performances come when the games mean the most...and he ended this year's playoff with a worse sv% than in the reg. season (.920R - .914PO). Most goalies step up, like Brodeur (.916R - .929PO). It favours your arguement to discount playoffs if Luongo has lesser stas and only 22 career PO GP.

The last item is value, which should be considered when assessing a goalie's Greatness from a wholistic perspective.

Brodeur - 5.2 million cap hit
Luongo - 6.75 million cap hit

By extension, NJ can sign more players or better players to ice a better team to ensure further success. Stopping pucks isn't the only way your help your team win as a goalie.

Anonymous said...

"2nd in the Atlantic (5th overall) and 1st (4th) and they beat Giguere who single handily carried them to the finals."

They were the best defensive team in the league, no doubt about it, and had very decent offense as well. I am assuming you will say Ottawa was the best team in the league that year, but not by much, and Lalime was the weak link. Nobody here is disputing that Brodeur isn't a lot better than Lalime.

Statman said...

"At best his Sv% might indicate Luongo stopped pucks better than Brodeur, but certainly not that he played better. The goalie is not just a two-dimensional cardboard cutout, he's an actual hockey player."

Hmmm, sounds like someone is/was a goalie who felt underappreciated. There's no shame in a goalie simply stopping the puck! haha

Anonymous said...

Statman said...

"At best his Sv% might indicate Luongo stopped pucks better than Brodeur, but certainly not that he played better. The goalie is not just a two-dimensional cardboard cutout, he's an actual hockey player."

Hmmm, sounds like someone is/was a goalie who felt underappreciated. There's no shame in a goalie simply stopping the puck! haha

...............

While using gigantic equipment right? haha good to see you back here continuing on with your jockhold on CG's work. Why dont you continue on telling us how Brodeur wears the largest equipment in the league instead on bringing up obscure and completely irrelevant analogies about how Stanley Cups do not matter.

The Contrarian Goaltender said...

Lawrence:

Who is your pick as the best goalie of the 1990s? Patrick Roy? Martin Brodeur?

It obviously can't be the guy who won half of the Vezinas and 20% of the Hart Trophies and put up a save percentage that was .013 ahead of everyone else, because he didn't win the prerequisite Stanley Cup championship.

Anonymous said...

for the record, perhaps some should go check on the discussion on HFBoards, specifically about this post. The overwhelming majority found it laughable. summed it up as more or less a "ridiculous augmentation" of skewed stats intentionally used to fulfill an agenda.

Anonymous said...

The Contrarian Goaltender said...

Lawrence:

Who is your pick as the best goalie of the 1990s? Patrick Roy? Martin Brodeur?

It obviously can't be the guy who won half of the Vezinas and 20% of the Hart Trophies and put up a save percentage that was .013 ahead of everyone else, because he didn't win the prerequisite Stanley Cup championship.

...........

See this is where the bogus assertion proposed by statman comes from. The idea by narrow minded people that its either one or the other. The idea that a 4th line grinder was better than Marcel Dionne because he has a cup or 2. Its not that hard to see were Lawrence, or Bruce or anyone else who values a player (specifically a goalie) winning a cup, is coming from. When comparing similar players, with similar attributes, obviously the guy with the cup has the edge over the guy without one. Dominik Hasek had unrivaled accolades during the 90's, enough to were it was not even an argument. Thus a cup for Roy or Brodeur would not be enough to swing the scales. However had Roy won 5 Vezinas, 2 Harts, and a Cup + Smythe, then perhaps he would have the a case for the 90's.

The Contrarian Goaltender said...

Bruce:

Do you consider the post-lockout Devils to be similar to the post-lockout Canucks? Which team do you think has been stronger, and in particular which team do you think was better offensively and which one was better at shot prevention? (Trying to exclude any contribution from the goalies, of course).

The Contrarian Goaltender said...

Anonymous:

"Skewed stats"? Is it disingenuous to suggest that scoring goals helps a team win hockey games? Am I coming from way out of left field with the assertion that allowing more shots against leads to more goals against, and therefore more losses? Am I mistaken in claiming that power play shots are more dangerous than even strength shots? If none of the above are wrong, then what exactly is skewed?

If you don't like the results, that's fine. But accusing someone of skewing stats without pointing out exactly what they find wrong with the method is not particularly constructive criticism.

And of course people who rate goalies based on their Stanley Cup rings and career win totals are always going to find these kinds of posts laughable. That doesn't bother me in the slightest.

Statman said...

I never once said "Brodeur wears the largest equipment in the league", jackass. I said that -- compared to goalies from years ago -- he wears huge equipment (as do all of today's goalies, to a greater or lesser degree). It was a comment about how much easier it is to have a low GAA & accumulate lots of shutouts.

Odd that you keep bringing that up.

Statman said...

HFBoards... yeah, cutting edge analysis over there. Say hello to 2002 for me.

Anonymous said...

Statman do you really want me to have to go back and copy and paste your exact quote. You said that brodeur wore gigantic equipment compared to most goalies IN THE LEAGUE. not in the league decades ago. I'm pretty sure your boy CG was the one who directly called you out on being wrong as well.

Statman said...

That should be 1982, not 2002, since I don't think HFBoards has begun to use save pct. yet. They'll get there someday.

Sure, scour the CG archives & take part of one of my old sentences to make it appear I was saying that Brodeur wore the biggest equipment in the league. I did say that considering he's 6'1 & 215 lbs, that it is very doubtful he wears some of the smallest equipment in the league (smaller than a 5'8" 170 lb guy? doubt it).

Was your mistaken interpretation of what I said one of the highlights of your life? Seems like it.

Anonymous said...

Statman said...

That should be 1982, not 2002, since I don't think HFBoards has begun to use save pct. yet. They'll get there someday.

Sure, scour the CG archives & take part of one of my old sentences to make it appear I was saying that Brodeur wore the biggest equipment in the league. I did say that considering he's 6'1 & 215 lbs, that it is very doubtful he wears some of the smallest equipment in the league (smaller than a 5'8" 170 lb guy? doubt it).
.............

Really? As already known to pretty much everybody but you, Brodeur wears some of the smallest equipment in the league. And yes, there are 5'8 goalies who wear significantly larger equipment. Check Manny Legace

"Was your mistaken interpretation of what I said one of the highlights of your life? Seems like it."
..............

No just further glaring proof that you have no idea what you are talking about. Especially when you get bold and try to speak on your own, which is rare considering how much time you spend regurgitating CG viewpoints.

Statman said...

Really, there are 5'8" goalies who wear bigger hockey pants than 6'1 215 lb Brodeur?

Anonymous said...

wow. I am starting to question whether or not you've even been watching hockey. Perhaps if you get off the computer and put the calculator away you'll notice these things.

Statman said...

... so ALL 5'8" NHL goalies wear bigger hockey pants than Brodeur?

Anonymous said...

Haha how predictable. Proven wrong again so you start splitting more hairs. I was actually waiting for you to actually go off on a tangent about how Legace does not count because he is 5'9 162 and not 5'8 170. Nevertheless. off course I am sure there is somebody you could find who has played a minute or two in the league who may have worn similar sized equipment. However, as stated by even your buddy CG, and common knowledge to even the casual hockey fan, Brodeur wears some of the smallest equipment in the game, and always has.

So specifically to answer your question, there are NO goalies in the NHL currently who are 5'8. The shortest current NHL goalie is Manny Legace at 5'9 who wears full round, league max equipment. Anything else?

Anonymous said...

"While using gigantic equipment right? haha good to see you back here continuing on with your jockhold on CG's work. Why dont you continue on telling us how Brodeur wears the largest equipment in the league instead on bringing up obscure and completely irrelevant analogies about how Stanley Cups do not matter."

Tell me how many Cups your god has won after the Devils went from elite to just good. EVERY cup Marty has was won behind a phenomenal team. The '00 Devils were within the top two or three best all-around teams of the decade and the '03 Devils were probably the best defensive team of the decade. The '01 Devils were the best all-around team of the decade and yet your boy Marty STILL couldn't hold off an Avs team that was missing its best player.

I want you to tell me the last time Marty won a cup since '03. That's right, he hasn't. Tell me why he didn't win a Cup this year (last season's Devils team was the worst of any this decade, and yet it was still better than the '02-03 Ducks), but rather was booted in the first round of the playoffs by an upstart underdog team that wasn't really that good, giving up two fairly soft goals in the last minute and a half of the game.

Anonymous said...

haha yes too bad he does not have 4 cups to his name, only 3. I mean the whole "how many cups since ..." has no merit. how many does gigeure have since 07? How many did he have at 23 years old? This is why players have careers, not just a season or two... no one wins every year.

and again with the whole "he blew it this year in the playoffs" thing. He was on the ice. when you actually play, things like that have a chance of happening. too bad he wasnt on the bench with J.S this year or the year before with the Dominator.

Corey Pronman said...

Hello CG, very good post, my comments:

1. For expeceted wins, are you taking the teams goals for + (shots per game * league average shooting percentage) / (PP against & leagues shooting % on the PP)? If so do you differentiate ES shots from PP shots?

2. Shot Prevention, can you detail that a bit, its a new stat to me. Is it some sort of Shot Quality metric?

3, Speaking of Shot Quality, that is the obvious weakness in the analysis. Obviously its hard to obtain a concrete form of it, but I wouldn't draw a conclusion without that information, just my opinion.

4. Good call on the marginal effect weakening from .500, but wouldn't most of that be random luck and not a indicator of goalie performance (even though I understand this is a wins alaysis)?

5. Great post!

Corey Pronman said...

To the people saying how many cups Marty has:

1. Marty didn't win a cup, the New Jersey Devils as a team won those cups.

2. The playoffs are a crapshoot.

Lawrence said...

CG said "Who is your pick as the best goalie of the 1990s? Patrick Roy? Martin Brodeur?

It obviously can't be the guy who won half of the Vezinas and 20% of the Hart Trophies and put up a save percentage that was .013 ahead of everyone else, because he didn't win the prerequisite Stanley Cup championship."

Can we have this discussion without everyone taking the statements to the point of absurdity? I have never claimed that Martin Brodeur is only better than other goalies in the past decade because he won a cup. I am saying that should be a consideration.

Anonymous said it best "When comparing similar players, with similar attributes, obviously the guy with the cup has the edge over the guy without one. Dominik Hasek had unrivaled accolades during the 90's, enough to were it was not even an argument. Thus a cup for Roy or Brodeur would not be enough to swing the scales. However had Roy won 5 Vezinas, 2 Harts, and a Cup + Smythe, then perhaps he would have the a case for the 90's."

CG, your argument hinges on the fact that Luongo has a sv% that is .005 better than Brodeur's over this period, or 50 goals on 10000 shots, which I am saying can be discounted by luck alone at such a small difference. This is where Luongo has been superior. Brodeur on the other hand has superior counting stats and individual accolades in basically every category, from wins, to minutes played, to cups, to vezinas, to goals scored, to stickhandling, to wearing smaller equipment, to having a better value contract, to basically every arguable stat except the sv% difference. You are trying to make this difference enormous and it's not. It's not even to the tune of more than 10 goals per year. It's also not like Brodeur's sv% isn't excellent as well. You may have something if it was a difference between .920 and .880, but .005 is not large enough to be of greater importance than every other statistical/counting stat advantage Brodeur has.

The cup point only furthers the distance between two very similarly regarded keeper's. It doesn't mean that if you won the cup you're great. I'm not about to say that Osgood is better than Luongo, because it's not even close. But with Brodeur and Luongo it's very close....005% close.

Corey Pronman said...

Lawrence,

While you formulate some good points, using acolades to strengthen your argument is futile. Awards in all sports are stupid, because its just peoples opinions who vote on them.

Vezinas, Harts, Smythes are all a terrible way to judge a player.

Cheers
MP

Anonymous said...

I have to agree with Moneypuck. The writers often don't know their head from their tuchis. Look at this year's Calder winner--probably the most undeserving Calder winner in 15 years.

aislephive said...

Very interesting post CG and I agree with your theory and conclusion that Luongo is the better goalie of the two.

With that said, you have a ton of useful stats to use for goalies. One that I thought would be useful would be save percentage during a tie or one goal game. Pro-Brodeur people love to downplay stats and act as if Brodeur has some clutch qualities that other goalies including Luongo don't. Of course, Brodeur's innate clutch abilities are ignored when you bring up the last 90 seconds of Game 7 this year or Luongo's superior playoff career save percentage, but I think a stat as I mentioned would do wonders to showing how effective both goalies are when it comes to "clutch" saves.

overpass said...

CG, your argument hinges on the fact that Luongo has a sv% that is .005 better than Brodeur's over this period, or 50 goals on 10000 shots, which I am saying can be discounted by luck alone at such a small difference.

Lawrence - one standard deviation for 10 000 shots against is 28 goals. A difference of 50 goals is almost 2 standard deviations. So yes, it's possible that it's luck alone, but it's highly unlikely and not anything I'd want to bet on.

The Contrarian Goaltender said...

MoneyPuck:

1. I'm making a separate adjustment for each category, figuring out the added goal differential and then converting that to wins. The power play shots against adjustment is based on 1.5 shots x (league average power plays against - power plays against) x .050.

(This is because .050 is the typical save percentage gap between EV and PK)

2. This is still a fairly new area of research, but here are a few posts that deal with the issue for Brodeur, Belfour, and just in general.

3. The evidence lately seems to suggest that shot quality measures aren't as important as we once thought. Most of the shot quality difference between Brodeur and Luongo is because Brodeur faced way fewer opposing power plays. Even if there is some additional EV shot quality difference, I doubt Luongo faced easier shots than Brodeur, so it's not likely to change my conclusion.

4. It's not equally easy to add wins to your team in all scenarios, that was my point with the marginal goals effect, and whenever you are comparing guys on the same metric it's best to level the playing field.

You are correct that there is probably some luck involved, because we are dealing with wins. Teams that tend to do well in close games will win more than their share, for example. But I think that most of the differential between actual wins and expected wins is likely to be from the effects of goaltending.

5. Thanks!

The Contrarian Goaltender said...

Lawrence:

I'm not trying to take it to the point of absurdity. I'm following the algorithm you outlined below:

"So get a list of your top ten stats wise... sv% with a minimum games played so on an so forth. Boil it down to ten let's say. then ask "Who of these ten has won the cup?" Boil it down again. Then really get into the detailed drill-hole numbers. End result = Brodeur."

Do you agree that Hasek is the best goalie of the 1990s? In that case, how come he survived step 2? It could be I'm just misunderstanding you. If a Cup only matters when guys are close that makes more sense and I don't have as much of a problem with it.

I still don't like that method of rating goalies though, not just because Cups are a team result but because you aren't defining the relative weighting of your criteria. How much does the Cup matter? If Patrick Roy had 4 Cups in the 1990s, does that make him better than Hasek? Are Giguere or Khabibulin better than Luongo in the 2000s because they have Cups? What about someone like Kiprusoff who has a Vezina and a deep playoff run, does that make up for Luongo's overall higher level of play?

When I rate goalies, I don't just try to look at everything and synthesize somehow in my head according to some unconscious heuristic which guy is more impressive. I go straight to the "drill-hole numbers", define some criteria that I think are important, run the numbers based on that, and see who comes out on top. That's what I did in this post, I reasoned that goalies who win more than expected must be good, set the criteria, calculated how much they would be expected to win and then compared against actual.

"CG, your argument hinges on the fact that Luongo has a sv% that is .005 better than Brodeur's over this period, or 50 goals on 10000 shots, which I am saying can be discounted by luck alone at such a small difference."

Re-read the post. I don't think I even mentioned save percentage in it. My argument does not hinge on a save percentage differential, it hinges on the fact that the difference between actual wins and expected wins is greater with Luongo in net than with Brodeur in net.

Secondly, I have to correct your estimate of the save percentage gap between the two goalies. You aren't taking special teams into account. Luongo's teams averaged over 100 extra power plays against per season, and as a result Luongo faced 23% of his shots against on the penalty kill compared to just 17% for Brodeur. If we break it down by game situation, we get the following:

Luongo: .929 EV, .889 PK
Brodeur: .921 EV, .875 PK

That is not a slight difference. That shows that Luongo is significantly better at stopping the puck, and that difference is over 12,083 even strength shots and 3,721 penalty kill shots for Luongo.

I am also intrigued that you will readily discount a persistent save percentage gap between Luongo and Brodeur over a 500+ game sample for both of them, yet conclude that Vinny Lecavalier is a better playoff performer than Joe Thornton from a 45 and 76 playoff games respectively and a scoring difference between them of just 0.03 points per game.

The Contrarian Goaltender said...

Just a clarification from something brought up in one of Bruce's comments, the shot prevention effect shown was intended to reflect the effect of both Brodeur and Luongo. I just put them both on one side to make the math easier. My esimate of 1.5-2 shots difference was for that reason, i.e. minus one for Brodeur and plus one for Luongo.

Corey Pronman said...

The evidence lately seems to suggest that shot quality measures aren't as important as we once thought. Most of the shot quality difference between Brodeur and Luongo is because Brodeur faced way fewer opposing power plays. Even if there is some additional EV shot quality difference, I doubt Luongo faced easier shots than Brodeur, so it's not likely to change my conclusion.

If you're referrign to Ferrari's study thats nowhere near conclusive. He didn't measure the extremes he measured the league as a whole so therefore the "results" obviously would look small or insignificant because you're looking at technically the league mean with uncertainty and sample size issues.

Bruce said...

Pro-Brodeur people love to downplay stats

Hahahahahahahahahahahahahahahahaha!!!!!!

Thanks for the laugh, Aislephive. That's the funniest comment I've read in weeks.

[word verification: junkies]

Bruce said...

using acolades to strengthen your argument is futile. Awards in all sports are stupid, because its just peoples opinions who vote on them.

Yeah, Moneypuck, those 30 GM's who vote on the Vezina, what the hell do they know about hockey? Meanwhile, I'm supposed to put a lot of stock in your Opinion?

A man whose opinion I greatly respect, Bill James, places a lot of stock in award voting, not just as a Hall of Fame criterion (awards = reputation) but as a legitimate analysis tool (awards = performance).

To me when one guy has received 44% of the Vezina Trophies in a decade and everybody else has received 0-11%, that speaks to both reputation and performance.

Lawrence said...

MoneyPuck: "Awards in all sports are stupid, because its just peoples opinions who vote on them.

Vezinas, Harts, Smythes are all a terrible way to judge a player."

Moneypuck, while I agree with you this is exactly what we are doing, just to likely more detail. We are formulating our opinion, analysing it, and 'awarding' the title of top goalie of the 00's. Which I agree with you is approaching the point of stupidity, because we are anuable to agree on the 'weighting' of criteria.

I'm not arguing that Roberto Luongo's sv% isn't better then Martin Brodeur's, I know that it is. I am arguing that the difference is not large enough to be weighed as more valuable than Marty's list of achievements, as well as his statisical dominance in nearly every category over Luongo.

Now, of course, it is possible to discount all of these items if you wish, or pick on one or the other, but the sv% stat has it's flaws as well.

CG said: "a few bad games can have a disproportionate impact on a goalie's [sv%] stats."

Overpass: "one standard deviation for 10 000 shots against is 28 goals."

Then there is the whole shot quality concept, which has been discounted, but at least, you have to consider a Standard Deviation on that calculation as well.

When it comes down to it, the difference between sv% for Luongo and Brodeur is not large enough to make a solid argument to put Luongo over Brodeur. Sure it puts him up there with him and Giguere as well.

This is the 'algorithm' that I have decribed. Then from here, it can be ranked a very tight 1-2-3 Luongo, Brodeur, Giguere, for example. Then you have to consider all the other counting stats in Brodeurs favour - Games, Wins, Win %, Playoff Games Played, Increase in Playoff Sv%, Goals Against Avg., Goals Scored, Stickhandling Superiority, Value Contract (which can increase likelihood of better team and more success), Vezina's, etc., etc. any number of things. When looking at all of those items, it favors Brodeur by a landslide.

There are two camps, and I need to just stop having this debate. Those who try to discount or reduce the importance of everything but sv%, and those who consider sv% as well as all other stats. If you base your argument solely on sv%, you need to acknowledge the weaknesses in that arguement.

Lastly, the only way we can adequately measure for team effects fairly is when the played on the same team. Which they have done twice. Once for the World Cup and once for the Olympics. Brodeur was also chosen when Luongo wasn't even considered to be invited, which tells you something about measuring peak-play, and this arbitrary 00's cutoff which may not even been considering Brodeur's best years.

Anyway, Brodeur was chosen over Luongo on three ocassions, twice when Luongo was on the team, and he was also favoured to play, while injured (his wrist) in the finals vs Finland (and Kipper) over Luongo and his same team international statistics are vastly superior, including sv%.

Brodeur 8GP 11GA GAA 1.38 SA 179 sv% .953
Luongo 3GP 6GA GAA 1.96 SA 82 sv% .927

Small sample size yes, but that's not even close, also considering Luongo played weaker QualComp. Game against Germany w/ 12 SA.

Even if you use a .008 difference in sv% thats 80 goals on 10000 on the 2000's stats, where I Bruce and I agree on .005. The difference when they were playing on the same teams are .026 or 260 goals on 10000. This is the best indication of the two in the same 'controlled' condition and again, its not even close. Add to that, the Brodeur was favoured by his peers over Luongo on three occassions in the 2000's and the answer is pretty clear.

overpass said...

When it comes down to it, the difference between sv% for Luongo and Brodeur is not large enough to make a solid argument to put Luongo over Brodeur. Sure it puts him up there with him and Giguere as well.

Lawrence, it's a huge argument to put Luongo over Brodeur. CG posted the SV% stats by game state, which is the only fair way to evaluate them! Luongo's edge here is very significant by any measure. That's just basic statistics.

Then you have to consider all the other counting stats in Brodeurs favour - Games, Wins, Win %, Playoff Games Played, Increase in Playoff Sv%, Goals Against Avg., Goals Scored, Stickhandling Superiority, Value Contract (which can increase likelihood of better team and more success), Vezina's, etc., etc. any number of things. When looking at all of those items, it favors Brodeur by a landslide.

Some of these are reasonable to include, including puckhandling/shot prevention and durability. They give a fuller picture of the value provided by each goaltender. But when you start throwing GAA, wins, win%, Cups won, playoff games played, etc. into the mix, can't you see that each of those contains a massive bias in the same direction, and strongly favors goalies on stronger teams? You aren't getting a fuller picture of value at all by including these, you are selecting goalies on good teams. Including everything in the comparison is not a reasonable thing to do in this case.

As a matter of fact, CG's post above was based on wins, but he properly adjusted for team strength. I guess if you wanted to go through and adjust every stat you suggest for team strength, you could, but it would be a very long and largely redundant exercise.

Your suggestion of contract size as a measuring stick seems to me to miss the point. We're interested in hockey here, not contract negotiations. Also, goals scored? Are you serious?

Brodeur 8GP 11GA GAA 1.38 SA 179 sv% .953
Luongo 3GP 6GA GAA 1.96 SA 82 sv% .927

Small sample size yes, but that's not even close, also considering Luongo played weaker QualComp. Game against Germany w/ 12 SA.

Even if you use a .008 difference in sv% thats 80 goals on 10000 on the 2000's stats, where I Bruce and I agree on .005. The difference when they were playing on the same teams are .026 or 260 goals on 10000. This is the best indication of the two in the same 'controlled' condition and again, its not even close.

http://en.wikipedia.org/wiki/Sample_size

Bruce said...

You know what Luongo hasn't done? Won anything of significance.

I was going to respond to this comment by Chris earlier, and now Lawrence's comment immediately above has reminded me. Luongo has indeed won three championships of signficicance, namely the Worlds in both 2003 and 2004 as well as the World Cup, also of 2004. In two instances he was the back-up but was called into action in the medal round and delivered some very good games.

Going from memory here, at the 2003 WC BobbiLu came on for an injured Sean Burke at some point during the Semi-Final, finished the job, then started and won the Final. In 2004 he was the main guy as Canada repeated.

At the 2004 World Cup Brodeur was the starter, but got hurt in the quarter-finals. Luongo played the Semi-Final against Czech Republic, facing something like 40 shots and winning 4-3 in OT (on Vinny Lecavcalier's winner). Come the Final and Brodeur was ready to go, Quinn went right back to him, and Canada went on to beat Finland 3-2. Brodeur faced considerably fewer than 40 shots, but he too got the job done, in regulation, with the same goal support.

I would dearly love to have access to game tape of those two games, featuring Luongo and Brodeur behind the exact same team, playing against a real tough foe in a loser-goes-home scenario. (Anybody know if it's avaiable somewhere?) My strong recollection of the games is that Canada had a real poor game defensively against the Czechs, the defence looked disorganized and a step slow, that they hung Luongo out to dry a few times and he came through with some big stops. Whereas in Brodeur's game the defensive crew spent most of the night avoiding serious trouble; while tense, the game was much more under control.

Now as an ex-goalie (yes, Statman) I like to think the goalie himself has an effect on the play around him, not just in stopping or even handling the puck but in communicating with and organizing the defence. Now that is something that is very difficult to see let alone judge, but all the arrows of his career-long efficiency curve suggest that Marty Brodeur is likely exceptional at those things too. But good luck finding it in the stats, other than the discredited (by some) Goals Against Average, and by extension the even-more-discredited Wins. Some highly-organized defensive clubs even win Stanley Cups.

Corey Pronman said...

"Yeah, Moneypuck, those 30 GM's who vote on the Vezina, what the hell do they know about hockey? Meanwhile, I'm supposed to put a lot of stock in your Opinion?"

Because they get to watch all the games of all the top goalies right?

The Contrarian Goaltender said...

Lawrence, I don't base my argument entirely based on save percentage. Can you please stop arguing against the strawman?

The argument is that Luongo is a better goalie because he prevents more goals and thereby helps his team win games. Save percentage is the best proxy for that contribution, but it is not perfect and nobody is claiming it is.

The above post has nothing to do with save percentage. It was in fact a direct attempt to debate the issue of Brodeur vs. Luongo with people who do not agree with me on the importance of save percentage, hence the use of wins as the comparative metric. The only reason we're talking about save % now is that the guys who supposedly dislike it so much keep bringing it up in the comments.

I have no problem debating Brodeur vs. Luongo in terms of wins, shutouts, GAA, or whatever else. I showed in this post that even with optimistic assumptions about Brodeur (e.g. no shot quality effects, shot prevention of 3 per game) Luongo is still very close in terms of his contributing to helping his team win hockey games. This post shows that Luongo's shutout record is better than Brodeur's after adjusting for his level of shots against. I've also done comparisons to backups that show Luongo's GAA edge is higher than Brodeur's.

In short, there is lots of evidence that suggests Luongo is better. My interest here is not to twist save percentage numbers around to promote my favourites, but to find better ways to evaluate all aspects of goalie play, and I'll use all available tools and data to do so.

Statman said...

"I like to think the goalie himself has an effect on the play around him, not just in stopping or even handling the puck but in communicating with and organizing the defence. ... Marty Brodeur is likely exceptional at those things too."

Oh man, give me a break... do you seriously think that at the NHL level a goalie is "organizing the defence", barking out plays, telling the d'men where to go, etc.? [Of course, all goalies will be communicating with the d'men, "heads up" etc.]

Has a NJ player ever said to the press, "Marty's defensive game plan worked perfectly tonight, he called all the right plays, allowing us to win" ? haha

That might be somewhat conceivable if goalie was relatively much much MUCH better than the skaters on his team (e.g. an ex-junior/CHL goalie playing with ankle-skaters in a no-contact beer league), but in any sort of competitive league? NO WAY. You're giving way way WAYYY too much credit to Brodeur.

Lawrence said...

Overpass: I suggest you read this:
http://en.wikipedia.org/wiki/Scientific_control

CG: "Can you please stop arguing against the strawman?

I am including previous arguments and Bruce's work, however, I don't think it's a stretch to state this basis. You are trying to present your argument as deductive logic, and it's totally unsound, but you keep going back to it as sound.

What I understand you say:

The goal of a hockey team is to win games.
Goalies role in winning games is solely by stopping pucks
Roberto Luongo has a greater percentage of pucks stopped than Martin Brodeur in the 00's.
Therefore, he is a greater goalie than Martin Brodeur.

Now, in this post you deduce this:

The goal of a hockey team is to win games.
Goalies role in winning games is solely by stopping pucks.
Roberto Luongo had worse teams than Martin Brodeur in the 00's.
Therefore, he is a greater goalie than Martin Brodeur.

Both of these arguments are based on very dubious data where the small factors of error, bias, deviations cloud the argument.

The more sound argument is this:
We are comparing two goaltenders to conclude who is the greatest of all goalies in the 2000's.

Find a control group:

Have they both have played for the same team at the same time? (thus ideally eliminating the team argument, a main point of contention)

Yes, twice 2004 and 2006. Also worthy to note, they could have played for the same team in 2002 but Luongo was not even selected.

Who played better under this same team environment?

Brodeur. Significantly. He was also chosen over Luongo as the #1, thus, better, goalkeeper.

Brodeur 8GP 11GA GAA 1.38 SA 179 sv% .953
Luongo 3GP 6GA GAA 1.96 SA 82 sv% .927

So then Brodeur is better.

Are there any contentions to this logic?

Yes, sample size is too small. Ok, this is true, but it cannot be discounted because it is our most pure comparison.

Further investigation.
What is the objective or vision of a hockey team?

To win, namely, the championship or Stanley Cup.

Have either of these goalies been one of the six players on a team that has won the cup?

Yes, Brodeur... twice in the 2000's and Luongo:0. In fact, Luongo has not advanced past the round of 8 and has twice been part of teams that failed to qualify for the post-season, although being regarded as being playoff-bound teams.

Therefore, Brodeur is better.

Contentions?

Goalies don't single-handedly win the Stanley cup, there is a huge team bias now introduced.

Ok, noted. However, this cannot be disregarded as Brodeur has twice been on teams that have completed their season objective in the 2000's. Luongo - never.

Then you go through wins. Where Brodeur has more wins, and the contention is Goalies only win by making saves. This is also contended by Goalies control flow, yes can score (it's another + not matter how small), stickhandle and pass, have a lower cost...allowing for the creation of a better team etc.etc. Again, Brodeur has more wins.

Then you get to save%. Where Luongo finally is ahead, and the Contention is: Team bias, Shot Quality Bias, Scorekeeper Bias, Statisical Average being misleading, etc.etc.

Then it's individual accolades, anecdotal data etc etc.

However, the further down the rabbit hole you go, the more dubious the data because it is less sound.

So the result is: It appears that Martin Brodeur is the Greatest Goaltender in the 2000's. However, it should be noted that Roberto Luongo is a very good puck stopping goalie, even superior to Martin Brodeur and perhaps had he been on the teams Martin Brodeur was on he would have performed even better. There is compelling data which shows that Luongo has performed at a higher level with worse teams than Brodeur, but it is an unsound assumption that he would do better if the situation were reversed. There is not enough sound evidence to prove that Luongo is actually better and due to the overwhelming evidence in favor of Martin Brodeur, one must conclude that Martin Brodeur is the greater player of the two.

overpass said...

Yes, sample size is too small. Ok, this is true, but it cannot be discounted because it is our most pure comparison

Lawrence, you are wrong about the sample size issue. The sample size you are using is far too small to distinguish between two NHL goalies.

Your sample includes 82 shots against for Luongo and 179 shots against for Brodeur. Over 82 shots, Luongo's SV% was 0.927. However, a confidence interval of 2 standard deviations says that our best estimate of his true save% from this data is 0.927 +/- 0.056, or somewhere between 0.870 and 0.983.

Your mistake is thinking that because the small sample of "scientifically controlled" data is all you have, it must be used. No, if all you have is a very small sample, you have nothing.

The Contrarian Goaltender said...

Lawrence:

I believe our fundamental disagreement has little to do with save percentage. It is that you don't think we can adjust for team effects on a goalie, while I think that we can. If you simply throw out every single attempt I make to assess a goalie's team context as "unsound" then I understand why you aren't being convinced by the evidence presented, because you're still giving Brodeur credit for being on a better team.

If you think the post above is arguing that a goalie's role in winning games is solely by stopping pucks, then I must not have described my method clearly enough. I think it is fair to claim that my method was in fact the exact opposite of what you suggest (picking one stat and claiming that is the only significant one).

My argument was actually an argument by subtraction, i.e. estimate what the rest of the team contributed, subtract that from the team's total result, and what you have left is the goaltender's contribution. That might be from stopping the puck or handling the puck or controlling rebounds or directing traffic or making HUGE saves at KEY times or inspiring confidence in his teammates or however he managed to do it, it doesn't matter as long as it resulted in his team winning hockey games.

You can argue all you want about the relative merits of save percentage, but that has nothing to do with what the above post says because it is not based on save percentage at all. If you want to prove my numbers wrong, you either have to point out some unaccounted for variable (others have discussed shot quality, for example) or you have to say why you disagree with my expected wins estimates based on the performance of the rest of the team. Given your obvious hang-ups with comparing players on different teams, you probably do have issues with the latter one. But if we're talking about that, we're at least debating the issue directly.

The suggestion that we can't quantify goaltending in hockey because we don't have a control group would probably cause an economist or a social scientist to burst out laughing. Either that, or they'd start crying because they would have realized that everything they had ever done was completely useless by the same logic.

Compared to the kind of data collection problems and sample size issues and lurking variables and other statistical issues researchers in those fields always run into, it is comparative child's play to estimate the contribution of a hockey goalie. He's the one guy dressed differently than everyone else on a hockey team, he has a specific job to do, he spends nearly the whole game standing in the same blue semicircle, and every time he goes to work there are people who are employed to watch him and keep a numerical record of his activities. There really is not that high a degree of complexity here.

I'm not saying we can know everything with 100% certainty. Just that people who say that everything is shrouded in impenetrable layers of fog and mystery do not understand statistics.

Bruce said...

you don't think we can adjust for team effects on a goalie, while I think that we can.

Can you also adjst for goalie effects on the team? It's a two-way street. Great players make the players around them better, and I see no reason why that doesn't apply equally to goalies as "out" players. As a result, great goalies, like great skaters, have a much greater than random chance of playing on excellent teams.

Lawrence said...

Overpass: "Your mistake is thinking that because the small sample of "scientifically controlled" data is all you have, it must be used. No, if all you have is a very small sample, you have nothing."

What is the confidence interval on 'selection as the starter'? I'm agreeing that the sv% is not the strongest data point, but it also comes with the fact that Brodeur was chosen over Luongo on all three occassions, and twice Luongo watched from the bench more than he played. Same team, same control, result? Brodeur.

The superior play, however small, only reinforces that result, but disregard it if you will and address the selection of.

overpass said...

What is the confidence interval on 'selection as the starter'? I'm agreeing that the sv% is not the strongest data point, but it also comes with the fact that Brodeur was chosen over Luongo on all three occassions, and twice Luongo watched from the bench more than he played. Same team, same control, result? Brodeur.

Yes, Brodeur was the starter. Obviously Hockey Canada and the coaches thought he was their best bet. I wouldn't necessarily say they were wrong to pick Brodeur - he was the reigning Vezina winner and was the obvious and conservative choice. Hockey Canada also loves continuity, and Brodeur was in goal when Canada won in 2002.

At the same time, I thought Luongo was the better NHL goalie in 2004 and 2006. I thought he would have been a better pick for Canada, but there's nothing wrong with Brodeur either.

The Contrarian Goaltender said...

Bruce, I assume you mean "equally" as in the logic is the same. I certainly hope you don't mean "equally" in that the goalie's effect on the team is equal to the team's effect on the goalie, i.e. the goalie is responsible for 50% of the total team result.

I agree goalies have an effect on their teams. I just don't think that effect is as big as you think it is. Post lockout Vancouver and post-lockout New Jersey are fairly similar teams, are they not? They score about the same, and the Canucks face about 1 extra shot against per game. The main differences seem to be that the Canucks take more penalties and play in a tougher conference. Let's say Brodeur prevented 2 shots per game compared to average and Luongo created 2 shots per game compared to average. Do you seriously think that the Canucks and the Devils would be within 1 shot of each other in shots against?

I think our problem with goalie effects on the team is mainly a lack of information. If the NHL still tracked zone times, for example, then I would be able to tell you whether the Devils spent more time in their own zone with Clemmensen in net than with Brodeur in 2008-09. Unfortunately we don't have that kind of data so we have to try to tease it out in other ways. It's a work in progress, but until then I'm going to go with what I know. It's just as easy to overestimate as to underestimate (just ask Martin "I prevent 8-10 shots per game" Brodeur")

Lawrence said...

"Yes, Brodeur was the starter. Obviously Hockey Canada and the coaches thought he was their best bet."

So this was my point for the first line of inductive reasoning outlined two of my comments ago. When looking for a sample which negates team effects (by them being on/or available for the same team) Brodeur is the answer to the question.

This is strong induction:
In three possible occasions in the past ten years Brodeur has been selected over Luongo as the better (starting) goalie by Hockey Canada.

Therefore, Hockey Canada believes Brodeur is a better (starting) goalie than Luongo.

If we are examining all the 'evidence' if you will, especially interested in negating goalie-team effects because "our problem with goalie effects on the team is mainly a lack of information."

Then this is a first and very important piece of the puzzle. It also furthers to solidify that argument by their performances in these same-team games, where sample size can be called into debate but contradicting stats cannot. Regardless, this doesn't change the result. Now, we can move onto number 2 that I suggested as an alternate way to argue Luongo vs Brodeur instead of attempting to 'take away' team effects which I do not think is "shrouded in impenetrable layers of fog and mystery" but is suffering from "lack of information", and significant bias.

The only other way I can think of to soundly compare without team effect is the shootout, where the goalie truly plays as an individual. One would expect, if Luongo was truly a vastly superior puck-stopper (sv%) that he would be far ahead. Nope. The are basically equal .716 (37 SO's) vs .715 (42 SO's)

Therefore we can say that both goalies are equally good at stopping pucks when team effects are removed, or the slight edge leans to Brodeur.

Trying to argue goalie-team effects, subtract goalie-team effects, or average out goalie-team effects discounts systems dynamics, and this becomes an unsound argument.

Statman said...

George W Bush was twice elected as President of the USA.

Therefore, GWB was the best presidential choice.

Bruce said...

//Great players make the players around them better, and I see no reason why that doesn't apply equally to goalies as "out" players.//

Bruce, I assume you mean "equally" as in the logic is the same. I certainly hope you don't mean "equally" in that the goalie's effect on the team is equal to the team's effect on the goalie, i.e. the goalie is responsible for 50% of the total team result.

Nope, I mean the idea a great player makes the players around him better applies equally to goalies as skaters.

Lawrence said...

George W Bush was twice elected as President of the USA.

Therefore, GWB was the best presidential choice.

Again...going to the ultimate point of absurdity. That would be weak induction because there is not matter of "best" involved. there were only two candidates. So you could say that: Therefore electors preferred Bush over the other candidate.

Besides, all you've proven is Canadians know how to select goalies and win and Americans no nothing of their own political system/international politics.

overpass said...

Lawrence, it seems to me that you are hand-waving away a ton of information in your quest to find "a sample that negates team effects."

Brodeur and Luongo have each faced over 17 000 shots in the past decade in NHL games. That's a very good-sized sample. It's well worth taking the effort to adjust for team effects.

It seems bizarre to throw that out in favour of a decision made by Hockey Canada twice, the reasons for which we know nothing. In my mind, the simplest explanation for their decision to pick Brodeur is continuity and going with the guy who won it in 2002, not a considered evaluation of who was the best goalie of the decade.

Anonymous said...

Blogger Statman said...

George W Bush was twice elected as President of the USA.

Therefore, GWB was the best presidential choice.

..............

priceless. maybe you should stick to trying to convince us brodeurs goalie equipment is the largest in the league. or maybe just continue on twisting things to make it seem like everybody just "misinterprets" what you say.

Statman said...

I said that Brodeur has the largest equipment in the league?

Once again, you spelling-deficient ass-face: "no".

Anonymous said...

Wow, name calling!!! Great. Go find all those 5'8 NHL goalies so you can finally make your point about how there might be a goalie who wears smaller equipment.

Anonymous said...

CG: I would also like to hear how Luongo playing in what you consider to be "the tougher conference" has any relevance in comparing him to Brodeur. Even if we acknowledge the West is stronger (which is not necessarily the case), you still can not argue that Luongo has it as difficult as Brodeur does. The Atlantic is widely regarded as the strongest division in the NHL, regularly sending 4 teams to the playoffs. Seeing as divisional opponents are the ones a goalie sees most frequently, conference is kind of irrelevant.

The Contrarian Goaltender said...

"I would also like to hear how Luongo playing in what you consider to be "the tougher conference" has any relevance in comparing him to Brodeur."

It doesn't. In that quote I was comparing the relative strengths of Vancouver and New Jersey, not Luongo and Brodeur. The Atlantic is not the strongest division in the NHL, but that's a separate argument. Basically my contention is that Vancouver and New Jersey are similar teams, and anyone who claims Brodeur is far better than Luongo would have to substantially disagree with that assessment.

Anonymous said...

Similar yes, but defensively Vancouver is better. They play a defense first system. New Jersey has a blue line made up of minor leaguers and undrafted players like Mottau and Oduya. Outside Paul Martin their next best dman is a guy whos half blind. Watching the playoffs, outside of the final game were Luongo was awful, it was clear how good Vanbouver could be defensively. Especially that game in Chicago (game 4) where they trapped the entire game, and Luongo blew an easy shutout by giving up a late softy to Havlat, only to lose in OT.

Also just for shits, the Central is imo currently the strongest division, however it is hard to argue that there has been a division as strong as the Atlantic year in and out since the lockout.

The Contrarian Goaltender said...

Vancouver has the better defence corps, sure, but New Jersey has better forwards, who also have an effect on shots against. Vancouver also takes a ton more penalties, which results in both more shots and more dangerous shots against.

Whatever our subjective opinions are, what really matters is performance. Here are the average shots against per 60 minutes, in games they started only, for the backup goalies on both teams over the last 3 seasons:

Brodeur's backups: 3612 minutes, 1756 shots, 29.2 SA/60
Luongo's backups: 2395 minutes, 1158 shots, 29.0 SA/60

That's pretty much identical, and because of the injuries we're getting into some unusually high sample sizes here for both.

Compare that to Luongo at 28.9 and Brodeur at 27.7. That puts Luongo about dead even with his backups and Brodeur 1.5 shots ahead of his.

Most of Brodeur's backup sample is made up by Scott Clemmensen, who I highly doubt is an above average "shot preventor". But, as I said in the post, I'll go along with a 1.5 shot per game difference between Brodeur and Luongo, although there is little evidence to suggest anything beyond that.

Re: Atlantic Division, Jeff Sagarin has the Atlantic Division ranked 5th in 2005-06, 5th in 2006-07, 4th in 2007-08, and 2nd in 2008-09 (source). So I'd say it's not particularly difficult to argue that the Atlantic has not been the league's best division since the lockout.

Lawrence said...

"Lawrence, it seems to me that you are hand-waving away a ton of information in your quest to find "a sample that negates team effects."

Overpass, I'm not hand-waving away a ton of information and more than you are reading my response. My 5th previous response including this one, I see no hand waving at all. I have included it in the second, third and forth points. I have never attempted to say Martin Brodeur is far better than Roberto Luongo. I have also never attempted to hand-wave away Brodeur's or Luongos performances. On the contrary, I have been saying look at all the evidence, from large to small, and include that, right down to Marty's goals (because it is an indication of his superior puck handling). As I have said numerous times:

If "it's well worth taking the effort to adjust for team effects." which I agree it is if one acknowledges that the assumptions make for an unsound argument because of bias, confidence intervals, standard deviation, styles of play, goalie-team effects, etc.etc. then it is also well worth taking the effort to consider situations without team effects, which I have named two: playing for the same team, shootouts.

Both of these situations which eliminate the team effect better than any other suggested to this point indicate Brodeur is as good, if not ever-so-slightly better puck stopper than Luongo. Then you need consider that when you start analyzing the 17,000 collective shots and start saying Brodeur is not as good a puck-stopper as Luongo and actually has helped his team win less.

Jonathan said...

"Brodeur 8GP 11GA GAA 1.38 SA 179 sv% .953
Luongo 3GP 6GA GAA 1.96 SA 82 sv% .927

So then Brodeur is better."

You are forgetting the 2005 world championships, which took place during the lockout so all of the best players were there--including Bobby Lou and Brodeur.

Brodeur posted a 2.86/.908 in 7 games and 218 shots
Luongo was 1.50/.950 in 2 games and 43 shots

So overall, Brodeur is 2.07/.922, while Luongo is 1.79/.928.

Luongo's averages would probably be closer to Brodeur's if he played more games on Team Canada. Bottom line is that there is no reason to believe that Brodeur has outperformed Luongo on the international stage--or vice versa. It's a wash.

yanmaneee said...
Anonymous said...