Your Personal Voting Record

[matnor]

Surprised to see so many people on HOH picking someone for the Hart who only played 57 of 82 games

But then, isn't it strange to give Kiprusoff the Vezina when he only played 38 games. I know the Vezina and Hart have different meanings but in my mind giving Kiprusoff the Vezina when Brodeur and Luongo played almost double the amount of games seems wrong.

 

[TheDevilMadeMe]

But then, isn't it strange to give Kiprusoff the Vezina when he only played 38 games. I know the Vezina and Hart have different meanings but in my mind giving Kiprusoff the Vezina when Brodeur and Luongo played almost double the amount of games seems wrong.

2003-04 was a strange/weak year for goalies because everyone who played had flaws. Ended up going for Kiprusoff because he singlehandedly saved Calgary's season. They weren't even close to the playoffs when he took over.

It wasn't the case of someone like Thomas, who was the team's starter all year but was rested regularly. Kiprusoff started almost every game for the team in the second half (I remember he had a really long streak of consective starts) as the team turned around completely.

 

[matnor]

2003-04 was a strange/weak year for goalies because everyone who played had flaws. Ended up going for Kiprusoff because he singlehandedly saved Calgary's season. They weren't even close to the playoffs when he took over.

It wasn't the case of someone like Thomas, who was the team's starter all year but was rested regularly. Kiprusoff started almost every game for the team in the second half (I remember he had a really long streak of consective starts) as the team turned around completely.

But then it's a little bit more like an MVP award rather than an award for best goaltender. Statswise, it looks as if Luongo had the best season but I know there are concerns about the shot counting in Florida. However, Luongo was 92.6% at home and 93.5% away so that does not seem to be a factor (but Brodeur was 90.4% at home and 92.7% away). He did face a lot more shots (32.0 per game) compared to Brodeur (24.6 per game) and I think it's reasonable to assume more shots against leads to a higher save percentage (the logic being that (i) teams that give up a lot of shots probably allow more perimeter shots and that (ii) facing fewer shots makes it difficult for a goalie to be alert when chances do appear).

 

[Master_Of_Districts]

But then it's a little bit more like an MVP award rather than an award for best goaltender. Statswise, it looks as if Luongo had the best season but I know there are concerns about the shot counting in Florida. However, Luongo was 92.6% at home and 93.5% away so that does not seem to be a factor (but Brodeur was 90.4% at home and 92.7% away). He did face a lot more shots (32.0 per game) compared to Brodeur (24.6 per game) and I think it's reasonable to assume more shots against leads to a higher save percentage (the logic being that (i) teams that give up a lot of shots probably allow more perimeter shots and that (ii) facing fewer shots makes it difficult for a goalie to be alert when chances do appear).

This could be determined by looking at the correlation between shots against and save percentage at the team level.

Have you looked at that?

 

[quoipourquoi]

I'd be surprised if Rask even got 1.75 goals per game while he was in net. It seemed like the Bruins had no interest in scoring for him. Thomas has no basis for the Hart at all that season.

71 Goals for Tuukka Rask; 175 for Tim Thomas

2.67 Goals/60 for Tuukka Rask; 3.12 Goals/60 for Tim Thomas

 

[Hardyvan123]

Surprised to see so many people on HOH picking someone for the Hart who only played 57 of 82 games

I agree here and just like baseball, goalies and pitchers already have their best player award so I'm hesitant to ever put one as my MVP.

57 games is also a huge problem , if he plays closer to 70 games it's one thing but the drop off to Rask that year wasn't huge either.

Okay so maybe stat wise it might look huge but he bookends 11 with much better stats in 10 and 12 and the small game sample might be part of simple variance IMO.

Thomas was very good in 11 but it's still only 57 games.

 

[Czech Your Math]

I'm tinkering with a couple of goalie metrics, similar to what others have done using SV% and GAA, and using replacement levels.

For 2004, using SV%, it seems Luongo is easily on top, Roloson is a distant second, and not far behind him are Aebischer, Raycroft and Kiprusoff. Using GAA, Turco eges out Brodeur, with Kiprusoff/Roloson battling for third.

I've looked at a few seasons, and it seems like some are cut and dried for best or best 3-4 goalies, and others are a free for all.

How do these replacement level numbers look for 2004?

SV% .910 or .905
GAA 2.43 or 2.76

It seems like Roloson really got the shaft that year, since his team finished last in a very tough division. However, he is at least top 4 in each metric (total, not per game).

 

[Czech Your Math]

I actually think this would be a fun summer project, even if it didn't have nearly the widespread interest of the "Top X" projects. I know I'd probably learn a lot in the process. Maybe start at expansion (or farther back) and go forward?

I'd like to see:

"Morenz" award for top forward
Norris
Vezina
Calder (not always sure who was eligible)
Hart
Pearson
Smythe
Playoff All-Star Team, but not as limited to Cup finalists... could have First Team for conference finalists (4 teams) and Second Team for conference semi-finalists (also 4 teams... or even include teams which lost first 7 game series)

I just don't think I can definitively say "player X should have won the Norris/Vezina" in many years, but will try to come up with some sort of list at some point (slowly working on it).

 

[Hardyvan123]

71 Goals for Tuukka Rask; 175 for Tim Thomas

2.67 Goals/60 for Tuukka Rask; 3.12 Goals/60 for Tim Thomas

Pro rated over a whole season Thomas (255GF) plays on the 5th highest scoring team in the NHL in 11 while Rask plays on the 22nd (218.94).

Boston was 8th overall.

 

[TheDevilMadeMe]

I'm tinkering with a couple of goalie metrics, similar to what others have done using SV% and GAA, and using replacement levels.

For 2004, using SV%, it seems Luongo is easily on top, Roloson is a distant second, and not far behind him are Aebischer, Raycroft and Kiprusoff. Using GAA, Turco eges out Brodeur, with Kiprusoff/Roloson battling for third.

I've looked at a few seasons, and it seems like some are cut and dried for best or best 3-4 goalies, and others are a free for all.

How do these replacement level numbers look for 2004?

SV% .910 or .905
GAA 2.43 or 2.76

It seems like Roloson really got the shaft that year, since his team finished last in a very tough division. However, he is at least top 4 in each metric (total, not per game).

Roloson only started 48 games and didn't distinguish himself from Manny Fernandez nearly as much as Kiprusoff distinguished himself from the other goalies in Calgary. There was a sense at the time that playing in Minnesota was very favorable to a goalie's numbers. The sense at the time was it was because of Jacque Lemaire, but does anyone have home/road splits for Minnesota?

Also keep in mind the massive difference between Brodeur's home and road save percentage (shown by Matnor above). Brodeur's save percentage was .023 higher on the road than at home, a massive difference and one that is highly unlikely to be explained by random variation. (Luongo was .009 higher on the road, which seems like it could be within random variation, or perhaps he was underrated himself but to a much lesser extent than Brodeur).

Brodeur's officially recorded save percentage in 2004 might be underrating him by 0.012 vs a goalie with an average shot recorder, which I'm sure you know is a massive difference.

 

[Czech Your Math]

Roloson only started 48 games and didn't distinguish himself from Manny Fernandez nearly as much as Kiprusoff distinguished himself from the other goalies in Calgary. There was a sense at the time that playing in Minnesota was very favorable to a goalie's numbers. The sense at the time was it was because of Jacque Lemaire, but does anyone have home/road splits for Minnesota?

Also keep in mind the massive difference between Brodeur's home and road save percentage (shown by Matnor above). Brodeur's save percentage was .023 higher on the road than at home, a massive difference and one that is highly unlikely to be explained by random variation. (Luongo was .009 higher on the road, which seems like it could be within random variation, or perhaps he was underrated himself but to a much lesser extent than Brodeur).

Brodeur's officially recorded save percentage in 2004 might be underrating him by 0.012 vs a goalie with an average shot recorder, which I'm sure you know is a massive difference.

Yes, I remember Minn. was considered a very defensive team, but so was NJ, and Roloson finished well in both metrics. I can see using the road/home splits somehow, but it seems like they could be influenced by difference in team play home/road (due to matchups), as much or more than shots being under/overcounted by home scorer.

 

[TheDevilMadeMe]

Just to show the massive difference that home/road splits could have.

This is what the actual top 20 in save percentage in 2004 looks like:

1. Miikka Kiprusoff-CGY .933
Dwayne Roloson-MIN .933
3. Roberto Luongo-FLA .930
4. Vesa Toskala-SJS .930
5. Andrew Raycroft-BOS .926
6. David Aebischer-COL .924
7. Evgeni Nabokov-SJS .921
8. Manny Legace-DET .919
9. Jose Theodore-MTL .919
10. Martin Gerber-MDA .918
11. Jussi Markkanen-TOT .918
12. Marc Denis-CBJ .918
13. Ed Belfour*-TOR .918
14. Martin Prusek-OTT .917
15. Martin Brodeur-NJD .916
16. Robert Esche-PHI .915
17. Manny Fernandez-MIN .915
18. Dan Cloutier-VAN .914
19. Jean-Sebastien Giguere-MDA .914
20. Martin Biron-BUF .913

This is what it would look like if you replaced the save percentages of Brodeur and Luongo with their road save percentages (assuming home recorders can't be trusted):

1. Roberto Luongo-FLA .935
2. Miikka Kiprusoff-CGY .933
Dwayne Roloson-MIN .933
4. Vesa Toskala-SJS .930
5. Martin Brodeur-NJD .927
6. Andrew Raycroft-BOS .926
7. David Aebischer-COL .924
8. Evgeni Nabokov-SJS .921
9. Manny Legace-DET .919
10. Jose Theodore-MTL .919
11. Martin Gerber-MDA .918
12. Jussi Markkanen-TOT .918
13. Marc Denis-CBJ .918
14. Ed Belfour*-TOR .918
15. Martin Prusek-OTT .917
16. Robert Esche-PHI .915
17. Manny Fernandez-MIN .915
18. Dan Cloutier-VAN .914
19. Jean-Sebastien Giguere-MDA .914
20. Martin Biron-BUF .913

Of course, for the study to be accurate, you'd need to use the road save percentage of all goalies, not just Luongo and Brodeur. Mathematically, there will be an equal overall decrease in the save percentages of other goaltenders as their is an increase in some.

Ignoring that, Brodeur's save percentage ranking goes from 15th to 5th if you just account for home arena effects, and thats' before even accounting for the value he adds to his team via puckhandling and playing almost every game. Maybe his 2004 Vezina (which he won in a landslide) isn't so undeserved as the official save % stat makes it look.

 

[TheDevilMadeMe]

Yes, I remember Minn. was considered a very defensive team, but so was NJ, and Roloson finished well in both metrics. I can see using the road/home splits somehow, but it seems like they could be influenced by difference in team play home/road (due to matchups), as much or more than shots being under/overcounted by home scorer.

It was shown that the officially recorded shooting percentages for both NJ and their opponents was much higher in NJ home games than road games, which is extremely strong evidence for undercounting.

 

[Hardyvan123]

I actually think this would be a fun summer project, even if it didn't have nearly the widespread interest of the "Top X" projects. I know I'd probably learn a lot in the process. Maybe start at expansion (or farther back) and go forward?

I'd like to see:

"Morenz" award for top forward
Norris
Vezina
Calder (not always sure who was eligible)
Hart
Pearson
Smythe
Playoff All-Star Team, but not as limited to Cup finalists... could have First Team for conference finalists (4 teams) and Second Team for conference semi-finalists (also 4 teams... or even include teams which lost first 7 game series)

I just don't think I can definitively say "player X should have won the Norris/Vezina" in many years, but will try to come up with some sort of list at some point (slowly working on it).

Very interesting idea and some of the wacky Hart voting Clarke/Esposito and some guys in the late 40's and 50's would wash out quite a bit differently I would think. Or maybe not.

 

[Czech Your Math]

Roloson only started 48 games and didn't distinguish himself from Manny Fernandez nearly as much as Kiprusoff distinguished himself from the other goalies in Calgary.

Brodeur's officially recorded save percentage in 2004 might be underrating him by 0.012 vs a goalie with an average shot recorder, which I'm sure you know is a massive difference.

Roloson did outplay Fernandez by a good amount though, and it was pretty close to what Kiprusoff did:

Roloson 2847, .933, 1.88
Fernandez 2166, .915, 2.49

Kiprusoff 2301 min, .933, 1.69
Turek 1031 min, .914, 2.33
McLennan 1446 min, .910, 2.20

Roloson played 24% more minutes than Kipper, and at a similar level. I wasn't implying that Roloson should have won, only that he finished a lot lower than he should have IMO.

Brodeur does seem to be underrated, perhaps by inequities in measuring SV% and also by the defensive nature of his team. I'm not sure why Lidstrom is clearly rated higher than him, when Brodeur had similar longevity and durability, while having a more direct and measurable impact on the defensive side of the game. I'm not a fan of his, either, I'd rather have a beer with Lidstrom than Hasek, Roy or Brodeur any day.

 

[Michael Farkas]

71 Goals for Tuukka Rask; 175 for Tim Thomas

2.67 Goals/60 for Tuukka Rask; 3.12 Goals/60 for Tim Thomas

That works just the same. They didn't score for Rask, they did score for Thomas, Thomas won and Rask didn't. This will look a lot better (not this little tete-a-tete, but the board's perception of Thomas in general) when Rask effectively sandwiches him with two Vezina-caliber seasons (2010 and 2013).

 

[Czech Your Math]

Very interesting idea and some of the wacky Hart voting Clarke/Esposito and some guys in the late 40's and 50's would wash out quite a bit differently I would think. Or maybe not.

When is the "top 60 centers" project starting? Hockey Outsider will be on vacation soon, and I'm sure many others will be as well. Could submit lists by labor day or something and then have threads for 2,3 or even 5 seasons simultaneously, otherwise it would take the better part of a year. Or might be better to wait and do it right, rather than rush it.

For the playoff all-star teams, we could split between Cup finalists and non-finalists in the O6. Might be best to do that for the 70's too, and then use the other format for 80s to present.

 

[TheDevilMadeMe]

When is the "top 60 centers" project starting? Hockey Outsider will be on vacation soon, and I'm sure many others will be as well. Could submit lists by labor day or something and then have threads for 2,3 or even 5 seasons simultaneously, otherwise it would take the better part of a year. Or might be better to wait and do it right, rather than rush it.

For the playoff all-star teams, we could split between Cup finalists and non-finalists in the O6. Might be best to do that for the 70's too, and then use the other format for 80s to present.

This board voted to do the goalies list before any of the forwards position. I started a preliminary research thread in May and it fizzled out in June. Hoping to ressurrect the project in the second half of the summer when people are jonsing for something hockey-related to do.

 

[Hardyvan123]

When is the "top 60 centers" project starting? Hockey Outsider will be on vacation soon, and I'm sure many others will be as well. Could submit lists by labor day or something and then have threads for 2,3 or even 5 seasons simultaneously, otherwise it would take the better part of a year. Or might be better to wait and do it right, rather than rush it.

For the playoff all-star teams, we could split between Cup finalists and non-finalists in the O6. Might be best to do that for the 70's too, and then use the other format for 80s to present.

It might be fun to work backwards chronologically (and use the same template and thinking) and do 1 year a week type of thing perhaps in regards to playoff all star teams?

I actually like the idea of making a playoff all star team top 4 at each forward position and 6 Dmen and 2 goalies per year but maybe that's too involved or it would be better to do 6 wingers and 6 centers?

Or where you thinking more traditional with just a 1st and 2nd all star team for the playoffs?

 

[Czech Your Math]

It might be fun to work backwards chronologically (and use the same template and thinking) and do 1 year a week type of thing perhaps in regards to playoff all star teams?

I actually like the idea of making a playoff all star team top 4 at each forward position and 6 Dmen and 2 goalies per year but maybe that's too involved or it would be better to do 6 wingers and 6 centers?

Or where you thinking more traditional with just a 1st and 2nd all star team for the playoffs?

I'm open to anything, just throwing some ideas out there. I like the full team idea, but have it more spread out over the rounds. 13 forwards, 6 D and 3 goalies?

The only concern I have with working backwards, is that there would likely be many participants at the start... and they would dwindle as it went back in time. If it starts way back when, it will probably be more limited to those that may tend to take it more seriously and so have more consistent participation.

 

[Master_Of_Districts]

Roloson only started 48 games and didn't distinguish himself from Manny Fernandez nearly as much as Kiprusoff distinguished himself from the other goalies in Calgary. There was a sense at the time that playing in Minnesota was very favorable to a goalie's numbers. The sense at the time was it was because of Jacque Lemaire, but does anyone have home/road splits for Minnesota?

Also keep in mind the massive difference between Brodeur's home and road save percentage (shown by Matnor above). Brodeur's save percentage was .023 higher on the road than at home, a massive difference and one that is highly unlikely to be explained by random variation. (Luongo was .009 higher on the road, which seems like it could be within random variation, or perhaps he was underrated himself but to a much lesser extent than Brodeur).

Brodeur's officially recorded save percentage in 2004 might be underrating him by 0.012 vs a goalie with an average shot recorder, which I'm sure you know is a massive difference.

You're underestimating the effect of random variation on save percentage over the course of a single season.

Only about 0.01 of Brodeur's road-home difference in 2003-04 is due to undercounting. The remaining 0.013 is due to random variation. This we know because we can estimate the magnitude of the recording in N.J over the course of Brodeur's career through a variety of methods. And those methods suggest that Brodeur was screwed to the tune of 0.01 at home.

Brodeur faced 1845 shots in 2003-04. A goalie facing 1845 shots will, on average, show a home-road SV% split of 0.01 (assuming half of all shots are faced at home; the other half on the road).

And that's assuming no recording bias.

 

[TheDevilMadeMe]

You're underestimating the effect of random variation on save percentage over the course of a single season.

Only about 0.01 of Brodeur's road-home difference in 2003-04 is due to undercounting. The remaining 0.013 is due to random variation. This we know because we can estimate the magnitude of the recording in N.J over the course of Brodeur's career through a variety of methods. And those methods suggest that Brodeur was screwed to the tune of 0.01 at home.

Brodeur faced 1845 shots in 2003-04. A goalie facing 1845 shots will, on average, show a home-road SV% split of 0.01 (assuming half of all shots are faced at home; the other half on the road).

And that's assuming no recording bias.

In English, please.

Where does the 0.013 "due to random variation" come from?

What does the "average goalie will show a home-road SV% split of 0.01" mean exactly?

 

[Big Phil]

Surprised to see so many people on HOH picking someone for the Hart who only played 57 of 82 games

Simple. 2011 was one of the weakest years for forwards in years. Maybe since 2004 or 2002. No one stood out. You could have picked St. Louis to win it and it wouldn't have been any worse than Perry. I would have picked Thomas either way but I think what is impressive about Perry was 19 goals in the final 20 games which helped Anaheim get into the postseason.

 

[Hardyvan123]

I'm open to anything, just throwing some ideas out there. I like the full team idea, but have it more spread out over the rounds. 13 forwards, 6 D and 3 goalies?

The only concern I have with working backwards, is that there would likely be many participants at the start... and they would dwindle as it went back in time. If it starts way back when, it will probably be more limited to those that may tend to take it more seriously and so have more consistent participation.

Starting way back has it's merits but there is already a built in bias or point of view in the history section when looking at older players but then again on a strict yearly look that would not matter either.