THN's Top 100 Players Of All-Time By Position (2010)

[seventieslord]

using last year's goaltending numbers and the 41 goalies with the most minutes played, the correlation between sv% and sog/min is .27. that's really nothing.

 

[matnor]

using last year's goaltending numbers and the 41 goalies with the most minutes played, the correlation between sv% and sog/min is .27. that's really nothing.

How do you figure? To me that's a pretty large correlation. I estimate that facing 10 extra shots/game means on average almost 2 percentage points higher save percentage, that's a huge effect.

 

[ContrarianGoaltender]

How do you figure? To me that's a pretty large correlation. I estimate that facing 10 extra shots/game means on average almost 2 percentage points higher save percentage, that's a huge effect.

At the individual game level you might be right, but the effect mostly disappears at the seasonal level because goalies don't face the same number of shots against every game. All goalies will have some nights where they get lit up early and end up not facing many shots and some nights where they make a ton of saves and have a high save percentage, which ends up mostly balancing out.

The other factor is that those games are not typical starts for the majority of goalies. Most games will end up with the goalie facing 25-35 shots, where there is little effect on save percentage as your graph shows.

Secondly, why are you using modern observations to evaluate Johnny Bower? That was a very different era. For example, during Bower's tenure in Toronto, the correlation between seasonal shots against per game and save percentage for goalies with 25+ games played was -0.012. From a quick look at the game-by-game results, I'm not even sure there was a relationship at the individual game level either.

 

[matnor]

At the individual game level you might be right, but the effect mostly disappears at the seasonal level because goalies don't face the same number of shots against every game. All goalies will have some nights where they get lit up early and end up not facing many shots and some nights where they make a ton of saves and have a high save percentage, which ends up mostly balancing out.

Don't really know what you mean. As I said previously I only included goalies who played the entire game. The fact is that, on average, facing more shots tend to be associated with higher save percentage. Not sure why the seasonal level would be interesting.

The other factor is that those games are not typical starts for the majority of goalies. Most games will end up with the goalie facing 25-35 shots, where there is little effect on save percentage as your graph shows.

That's true. But the effect is still there. For instance, Miller faced almost 5 more shots per game than Brodeur last season.

Secondly, why are you using modern observations to evaluate Johnny Bower? That was a very different era. For example, during Bower's tenure in Toronto, the correlation between seasonal shots against per game and save percentage for goalies with 25+ games played was -0.012. From a quick look at the game-by-game results, I'm not even sure there was a relationship at the individual game level either.

Point taken. I just reacted to the double counting of sa and sv% and I didn't think I had the sv% numbers from Bower's era available. Turns out I do so I might take a look at it.

Edit: I just looked at the sv% and sa data from 52/53-66/67 and there is virtually no relationship. So, forget my comments about Bower. I do think it's very important to consider these things for modern goalies but apparently it plays a very little role for the O6 era.

Edit2: Actually, something seems to happen in the middle of the 70s. Until 75/76 there is almost no relationship but from 76/77 and onwards there is a very strong relationship. Did something happen with the data between these years or what else could be the reason?

 

[TheDevilMadeMe]

Skaters that can legitimately be placed ahead of Richard using the most common criteria:

Gretzky, Orr, Lemieux, Howe, Hull, Beliveau, Shore, Harvey, Bourque, Mikita,

Goaltenders who could legitimately be placed ahead of Richard:

Hasek, Plante, Roy

Players who could legitimately be placed ahead of Richard given that a person prefers prime to career:

Potvin, Jagr, Lafleur

I'm sorry. Placing Stan Mikita or Guy Lafleur over Maurice RIchard is revisionist history at its worst.

Most people who watched Mikita play also watched Richard (they only played 10 years apart), and do any of them actually think Mikita was better? Even Blackhawks fans?

And I'm sorry, tell any long time Canadiens fan that Lafleur has a case to be better than Richard, and you'll be laughed out of the conversation.

At some point contemporary opinion has to matter.

Those are 16 players I could potentially agree with being ranked ahead of Richard, which is not to say that I believe that they certainly are better. I should note that I am assuming that Richard was basically useless defensively and also that many of his NHL seasons featured a somewhat diluted talent pool due to WW2, both of which I have heard arguments for but cannot say with absolute certainty. I also don't value playoffs as much as most do, which is the main reason that some people rate him as high as 5

Yeah, i get it that you can't be valuing playoffs that much if you'd even consider Mikita over Richard.

 

[Hardyvan123]

It's not just "he won some Cups, therefore he's good." (Your words, not mine). He backstopped a dynasty. As I said before, there are probably eight goalies in the history of the game with that claim. Six of the others - Sawchuk, Plante, Broda, Fuhr, Smith and Dryden - are consensus top 20 goalies/top 100 players in NHL history. The other is Gump Worsley, and Worsley would be just outside of my top 20 NHL goalie list. (For the record, I have Bower at No. 10 on my NHL goalie list).

Save percentage helps, but goalie, more than any other position, is judged based on wins and championships.

I would hope that playing behind that Leafs team wouldn't hurt his cause. As I said before, to me, it augments his legacy. Obviously the Leafs team defence, their blue line, and their system, in the 1960s, was incredible. But it doesn't work without Bower. I would hope that the voters would be smart enough to realize the value of an all-time great goalie in piecing together the team that many consider to be the best defensively in NHL history.

I'm not in a position to speak for the panel who voted for the 1998 list, but I think longevity might have been a greater drawback for Bower's cause. He played at least 40 games seven times; at least 50 five times. Essentially, it's a 12-year NHL career that we're looking at. (Incidentally, am I the only one who finds it incredible that an over-40 goalie could have as many games played in the NHL as candles on his birthday cake?)

BM was also probably right when he noted that Bower played for so long in the AHL. While the AHL was a tremendous level of hockey for most of the Original 6 era, accomplishments from that league, in that time, probably didn't matter much to the voters for the 1998 list.

Couple of points here the 1st being that I wonder if we can really call the AHL a tremendous level of hockey during bower's playing days. compared to hockey in Russia sure but let's not overstate the level of play in the AHL here.

Also Bower was great late in his career but how can we honestly put him in the top 10 (top 20 is a stretch too IMO) when he plays only one and a half seasons in the top league in the world at that time before the age of 35?

Also with the point that goalies are judged based on wins and championships, I believe this to be true, that they are judged by many this way, but we really need to use some kind of filter to separate a team from the player and goalie is the hardest position to do this IMO.

The bottom line is that his NHL career started too late for him to be considered even in the top 20 IMO.

 

[Hardyvan123]

I'm sorry. Placing Stan Mikita or Guy Lafleur over Maurice RIchard is revisionist history at its worst.

Most people who watched Mikita play also watched Richard (they only played 10 years apart), and do any of them actually think Mikita was better? Even Blackhawks fans?

And I'm sorry, tell any long time Canadiens fan that Lafleur has a case to be better than Richard, and you'll be laughed out of the conversation.

At some point contemporary opinion has to matter.



Yeah, i get it that you can't be valuing playoffs that much if you'd even consider Mikita over Richard.

Mikita was hardly a horrible playoff performer( a line of 155-59-91-150 to Richard's 133-82-44-126) and once again a player is being elevated for his team's success in the playoffs. Richard was a great player but also pretty one dimensional, albeit goal scoring is probably the best dimension to have.

I for one believe that you can make pretty good arguments of the 15 players listed above and some others as well.

Personally I think Richard is somewhere in the top 15 but it is really subjective to how much one thinks all those cups were due to him or the team in general.

Of the last guys mentioned I have Jagr ahead of Richard.

 

[seventieslord]

Don't really know what you mean. As I said previously I only included goalies who played the entire game. The fact is that, on average, facing more shots tend to be associated with higher save percentage. Not sure why the seasonal level would be interesting.

because no one talks about how so-and-so had a .912 sv% last night, they talk about what their seasonal save percentage was. And larger sample sizes always mean more.

Edit2: Actually, something seems to happen in the middle of the 70s. Until 75/76 there is almost no relationship but from 76/77 and onwards there is a very strong relationship. Did something happen with the data between these years or what else could be the reason?

Wait, what do byou call a "very strong" relationship? 0.27? That is not the least bit strong.

 

[canucks4ever]

Mikita above Richard would make sense to me. Better two way player, more harts, more scoring titles and he had worse linemates.

 

[overpass]

Don't really know what you mean. As I said previously I only included goalies who played the entire game. The fact is that, on average, facing more shots tend to be associated with higher save percentage. Not sure why the seasonal level would be interesting.

The issue is that the relationship you are finding on an individual game level may not hold up at the season level.

It appears to be a result of teams playing to the score. So the arrow of causation is the opposite of what you might think - low save percentages cause low shots against.

Wait, what do you call a "very strong" relationship? 0.27? That is not the least bit strong.

That's an r-squared of about 0.07, suggesting that the relationship explains ~7% of the variation. I wouldn't say that's very strong.

 

[Derick*]

Don't really know what you mean. As I said previously I only included goalies who played the entire game. The fact is that, on average, facing more shots tend to be associated with higher save percentage. Not sure why the seasonal level would be interesting.



That's true. But the effect is still there. For instance, Miller faced almost 5 more shots per game than Brodeur last season.



Point taken. I just reacted to the double counting of sa and sv% and I didn't think I had the sv% numbers from Bower's era available. Turns out I do so I might take a look at it.

Edit: I just looked at the sv% and sa data from 52/53-66/67 and there is virtually no relationship. So, forget my comments about Bower. I do think it's very important to consider these things for modern goalies but apparently it plays a very little role for the O6 era.

Edit2: Actually, something seems to happen in the middle of the 70s. Until 75/76 there is almost no relationship but from 76/77 and onwards there is a very strong relationship. Did something happen with the data between these years or what else could be the reason?

If the effect exists in single games because of playing to the score and different defensive strategies, but not across whole seasons, that shows that over the course of a season those two factors vary enough game from game for each goalie that each goalie gets roughly the same amount of each and therefore it's irrelevant in judging a goalie's seasonal or career save percentage, so it doesn't need to be accounted for. If someone wants to say their goalie who stopped 9 of 10 last night did as well as your team's goalie who stopped 27 of 30, feel free to cite that.

 

[JackSlater]

I'm sorry. Placing Stan Mikita or Guy Lafleur over Maurice RIchard is revisionist history at its worst.

Most people who watched Mikita play also watched Richard (they only played 10 years apart), and do any of them actually think Mikita was better? Even Blackhawks fans?

And I'm sorry, tell any long time Canadiens fan that Lafleur has a case to be better than Richard, and you'll be laughed out of the conversation.

At some point contemporary opinion has to matter.

Yeah, i get it that you can't be valuing playoffs that much if you'd even consider Mikita over Richard.

When a player is as romanticized as Richard I find popular opinion can be misleading. That Mikita was a relatively uncharismatic player only exacerbates this problem. I won't pretend to have been watching either player playing during their respective careers (thanks for the tip that they played ten years apart by the way!!!), but all the evidence I have seen leads me to believe Mikita was better.

Defensively, as far as I can tell Richard was not a contributor. Mikita was quite competent defensively, and also one of the best faceoff men in the NHL. Offensively Richard never led the NHL in points, despite frequently having very strong linemates and playing against somewhat weak competition early in his career. Admittedly his competition later in his career was very strong. Mikita led the NHL in scoring four times, with relatively weak linemates, despite competition from Hull, a somewhat past his peak Howe, Beliveau, and later even Esposito and Orr. I can't see any evidence that Richard was better than Mikita either offensively or defensively, unless a person values goals tremendously over assists. I would say that Richard was a better playoff performer than Mikita, although this is somewhat tempered by the strength of Richard's teams compared to Mikita's.

As far as Lafleur and Richard, I still see no reason that a person couldn't consider Lafleur over Richard if they placed strong emphasis on prime... yes, even though it would upset Montreal fans.

 

[BM67]

Couple of points here the 1st being that I wonder if we can really call the AHL a tremendous level of hockey during bower's playing days. compared to hockey in Russia sure but let's not overstate the level of play in the AHL here.

The bottom line is that his NHL career started too late for him to be considered even in the top 20 IMO.

Consider who was in the AHL when Bower won his 3 MVP awards: Camille Henry, Bronco Horvath, Ken Wharram, Ab McDonald, Harry Lumley, Don Simmons, Gump Worsley, Charlie Hodge, Bob Baun and a fair number of other decent future and past NHL players.

Bower's NHL career started late largely due to his playing for an independent team. The Rangers, or any other NHL club, couldn't just call him up, they had to trade for him.

 

[Hardyvan123]

Consider who was in the AHL when Bower won his 3 MVP awards: Camille Henry, Bronco Horvath, Ken Wharram, Ab McDonald, Harry Lumley, Don Simmons, Gump Worsley, Charlie Hodge, Bob Baun and a fair number of other decent future and past NHL players.

Bower's NHL career started late largely due to his playing for an independent team. The Rangers, or any other NHL club, couldn't just call him up, they had to trade for him.

Still to be one of the best of all time you should expect that player to find his way to the NHL earlier, it's not like it was in the 1920's or anything.

I also find the use of the word tremendous to describe the AHL during that time an over statement.

 

[matnor]

because no one talks about how so-and-so had a .912 sv% last night, they talk about what their seasonal save percentage was. And larger sample sizes always mean more.

Of course, but that's the level I have been looking at. A guy facing on average 5 less shots per game during the course of a season would have a two percentage points lower save percentage. That is important.

That's an r-squared of about 0.07, suggesting that the relationship explains ~7% of the variation. I wouldn't say that's very strong.

Wait, what do byou call a "very strong" relationship? 0.27? That is not the least bit strong.

R-squared or correlation are, quite frankly, next to useless measures. When I say the relationship is strong I mean that it is statistically significant on any conventional significance level (for you statistically inclined, it has a t-value of over 10 with a sample size of slightly more than 2000). R-squared depends entirely on aggregation level. For instance, the graph I showed before has an R-squared of .69 and a correlation of .83 but that is only due to aggregation.

The issue is that the relationship you are finding on an individual game level may not hold up at the season level.

It appears to be a result of teams playing to the score. So the arrow of causation is the opposite of what you might think - low save percentages cause low shots against.

You are absolutely right and that is the core question. I definitely don't believe this is a causal estimate and one should be careful before making any sort of adjustments based on the correlation. I only brought it up because it's common to double count sa and sv%. Nevertheless, controlling for opposition which would be a very crude way of testing for reverse causality, doesn't change the estimate in any way.

 

[Dennis Bonvie]

When a player is as romanticized as Richard I find popular opinion can be misleading. That Mikita was a relatively uncharismatic player only exacerbates this problem. I won't pretend to have been watching either player playing during their respective careers (thanks for the tip that they played ten years apart by the way!!!), but all the evidence I have seen leads me to believe Mikita was better.

Defensively, as far as I can tell Richard was not a contributor. Mikita was quite competent defensively, and also one of the best faceoff men in the NHL. Offensively Richard never led the NHL in points, despite frequently having very strong linemates and playing against somewhat weak competition early in his career. Admittedly his competition later in his career was very strong. Mikita led the NHL in scoring four times, with relatively weak linemates, despite competition from Hull, a somewhat past his peak Howe, Beliveau, and later even Esposito and Orr. I can't see any evidence that Richard was better than Mikita either offensively or defensively, unless a person values goals tremendously over assists. I would say that Richard was a better playoff performer than Mikita, although this is somewhat tempered by the strength of Richard's teams compared to Mikita's.

As far as Lafleur and Richard, I still see no reason that a person couldn't consider Lafleur over Richard if they placed strong emphasis on prime... yes, even though it would upset Montreal fans.

You are comparing a center to a winger. A center has much more defensive responsiblity than a wing. But I'm not sure what makes you think Richard played no denfese. Players in the original 6 all played with some defensive responsibilty.

Richard never led the league in scoring, but lost by a single point twice. He was second 4 times and third twice. 10 times in top 5 (to Mikita's 9).

Led the league in goals 5 times, second 3 times, 12 times in top 5.

In 1953-54, at the age of 32, Richard led the league in goals. He led again the next season and was second in the next 2 seasons. Not many players from that era scoring like that in their 30's.

In the 1957-58 season, at the age of 36, Richard scored 11 goals in 10 playoff games, leading the playoffs in scoring also.

There was a reason that Richard was romanticized. He put the puck in the net like no one else had up to that time.

 

[JackSlater]

You are comparing a center to a winger. A center has much more defensive responsiblity than a wing. But I'm not sure what makes you think Richard played no denfese. Players in the original 6 all played with some defensive responsibilty.

I'm not going to give Richard credit for playing defence just because he played in the original six era. From what I've read about Richard from various sources including these forums the consensus was that he was disinterested defensively, although many times his contemporaries only hint at it. Everything I've read regarding Mikita (or seen) leads me to believe that he was quite responsible. Obviously it would be worse if Richard was a centre and didn't contribute defensively, but Mikita does get credit for being a better defensive player.

Richard never led the league in scoring, but lost by a single point twice. He was second 4 times and third twice. 10 times in top 5 (to Mikita's 9).

Led the league in goals 5 times, second 3 times, 12 times in top 5.

In 1953-54, at the age of 32, Richard led the league in goals. He led again the next season and was second in the next 2 seasons. Not many players from that era scoring like that in their 30's.

In the 1957-58 season, at the age of 36, Richard scored 11 goals in 10 playoff games, leading the playoffs in scoring also.

There was a reason that Richard was romanticized. He put the puck in the net like no one else had up to that time.

Mikita's four scoring titles to Richard's none obviously overstates Mikita's case offensively, but he did achieve better overall offensive results despite weaker linemates and tougher competition. Richard was not romanticized to the degree he was purely for his ability to put the puck in the net.

 

[Leaf Lander]

i was a lil shocked to see salming and kennedy fall of the top 100 list

 

[MXD]

Those are 16 players I could potentially agree with being ranked ahead of Richard, which is not to say that I believe that they certainly are better. I should note that I am assuming that Richard was basically useless defensively and also that many of his NHL seasons featured a somewhat diluted talent pool due to WW2, both of which I have heard arguments for but cannot say with absolute certainty. I also don't value playoffs as much as most do, which is the main reason that some people rate him as high as 5.

Many?!?!?!

Wow, that's a new way to play the semantics game.

A bit like in :

In the many seasons in which Richard and Mikita were both in the NHL, Richard badly outscored Mikita, even though Richard was BADLY past its prime.

 

[MXD]

i was a lil shocked to see salming and kennedy fall of the top 100 list

I think that's okay for Salming. Kennedy, on the other hand...

 

[JackSlater]

Many?!?!?!

Wow, that's a new way to play the semantics game.

A bit like in :

In the many seasons in which Richard and Mikita were both in the NHL, Richard badly outscored Mikita, even though Richard was BADLY past its prime.

Talent pool was weak from the start of Richard's career until the 1950s... a period slightly longer than Mikita and Richard's overlapping year. Good parallel though.

 

[seventieslord]

i was a lil shocked to see salming and kennedy fall of the top 100 list

It's not a top-100 list by any stretch. It's a top-20 by position. Kennedy is not a top-20 center and I guess Salming wouldn't be a top-20 defenseman either. But if it was an actual top-100 list, both would have to be there... at the expense of one of the 20 undeserving wingers.

 

[MXD]

Talent pool was weak from the start of Richard's career until the 1950s... a period slightly longer than Mikita and Richard's overlapping year. Good parallel though.

But the diluted pool didn't have much to do with WW2 for most of the 40ies...

 

[JackSlater]

But the diluted pool didn't have much to do with WW2 for most of the 40ies...

I'm guessing that the relative lack of talent into the 1950s had something to do with the young players who had to go to war/ take on greater responsibility at home instead of developing under normal circumstances. It could be that WW2 has little to do with it, but Richard still played many of his seasons against weak top end competition.

 

[Infinite Vision*]

I'm guessing that the relative lack of talent into the 1950s had something to do with the young players who had to go to war/ take on greater responsibility at home instead of developing under normal circumstances. It could be that WW2 has little to do with it, but Richard still played many of his seasons against weak top end competition.

Good point. I think that's honestly it.