Wednesday, October 25, 2006
Potpourri for .790
**Overdue props: Lowetide is a fantastic, fantastic blog. There are about twenty hockey blogs that I would describe as excellent, but no others come close to LT on this count: the possibility that the last thing he posted is fascinating. His personal photo and print archive is obviously considerable, as is his recall of hockey days gone by (Steve Tuttle? YGBSM). He's also pretty damn funny.
Which reminds me, he retold this oldie-but-goodie (and much varied) joke a couple of nights ago, which made me howl so I'm reprinting it here. The Bobby Hull Joke About Minsk:
**As noted, I didn't shell out for the PPV last night, so I missed seeing not only a righteous thrashing by the Flames, but what in all likelihood was Mayor Quimby's last appearance in an NHL uniform:
Quimby's SV% last season was .882; I just don't see that there's anything on his resume that would tempt another NHL team to take a flyer on him. The Grrrowl, on the other hand...
**Speaking of the Coyotes, credit is due to Tyler for his preseason prediction (15th in the W.C.). He used some cold, basic statistics (the whole series was riveting) to show that (A) they were average to poor in every element of the game last season, and (B) that the personnel changes they made -- or didn't make (CuJo) -- would probably make things worse.
A mini-rant on stats here: they are not the be all/end all. In terms of forecasting, they are helpful (an advantage) but they're far from perfect. I still like my own two eyes and gut feeling. (Example: Sens and Sabres. The underlying numbers say that Ottawa is and has been the superior team, but Buffalo had a better record for most of last season (the last 3/4) and beat Ottawa in the playoffs. Luck, maybe, but it looks to me like Buffalo is the better team, which is why I picked them to win the NE Division and the E.C.). But -- I don't think there's any doubt that statistical analysis is under-utilized by NHL teams and the hockey media.
Here's the thing: you can crunch and re-crunch numbers until your eyes go blurry and MSExcel crashes, and have reams of data on absolutely every player, team, and situation in the NHL. This does not prevent you whatsoever from considering less tangible things and giving them whatever weight you please.
I get the impression that some teams, GMs, writers, etc. think that too much stats might "infect" them. They are a freaking tool, and at times a darn useful one. Neglecting to make maximum use of them is a self-handicap, imposed on oneself or one's organization for no good reason.
**Guy Carbonneau, Fast Learner: "...Sergei Samsonov is on the fourth line with Steve Begin and Garth Murray." Next up: Jeff Friesen? (He, McCarty, & Richie were the only CGY fwds who didn't score a point last night).
**Go Flames.
Which reminds me, he retold this oldie-but-goodie (and much varied) joke a couple of nights ago, which made me howl so I'm reprinting it here. The Bobby Hull Joke About Minsk:
Man 1: Where you headed?
Man 2: Minsk.
Man 1: Minsk? Only things in Minsk are hookers and hockey players.
Man 2: My wife is from Minsk.
Man 1: What position does she play?
**As noted, I didn't shell out for the PPV last night, so I missed seeing not only a righteous thrashing by the Flames, but what in all likelihood was Mayor Quimby's last appearance in an NHL uniform:
The Phoenix Coyotes placed goalie Mike Morrison on waivers on Wednesday, less than 24 hours after a lopsided 6-1 loss to the Calgary Flames.
He is 0-3 in four starts with the Coyotes this season with a 6.13 GAA and .790 save percentage.
Quimby's SV% last season was .882; I just don't see that there's anything on his resume that would tempt another NHL team to take a flyer on him. The Grrrowl, on the other hand...
**Speaking of the Coyotes, credit is due to Tyler for his preseason prediction (15th in the W.C.). He used some cold, basic statistics (the whole series was riveting) to show that (A) they were average to poor in every element of the game last season, and (B) that the personnel changes they made -- or didn't make (CuJo) -- would probably make things worse.
A mini-rant on stats here: they are not the be all/end all. In terms of forecasting, they are helpful (an advantage) but they're far from perfect. I still like my own two eyes and gut feeling. (Example: Sens and Sabres. The underlying numbers say that Ottawa is and has been the superior team, but Buffalo had a better record for most of last season (the last 3/4) and beat Ottawa in the playoffs. Luck, maybe, but it looks to me like Buffalo is the better team, which is why I picked them to win the NE Division and the E.C.). But -- I don't think there's any doubt that statistical analysis is under-utilized by NHL teams and the hockey media.
Here's the thing: you can crunch and re-crunch numbers until your eyes go blurry and MSExcel crashes, and have reams of data on absolutely every player, team, and situation in the NHL. This does not prevent you whatsoever from considering less tangible things and giving them whatever weight you please.
I get the impression that some teams, GMs, writers, etc. think that too much stats might "infect" them. They are a freaking tool, and at times a darn useful one. Neglecting to make maximum use of them is a self-handicap, imposed on oneself or one's organization for no good reason.
**Guy Carbonneau, Fast Learner: "...Sergei Samsonov is on the fourth line with Steve Begin and Garth Murray." Next up: Jeff Friesen? (He, McCarty, & Richie were the only CGY fwds who didn't score a point last night).
**Go Flames.
# posted by Matt @ 1:13 pm 
Comments:
Three Reasons why there will never be a "Moneypuck":
1) Inability to isolate individual performance.
2) Inability to categorize and grade "hockey actions" without reverting to a great deal of subjectivity.
3) Too many players, too many variables, too many things going on that are impossible to substantiate.
That doesn’t mean statistical analysis is meaningless in hockey. It simply means that there are substantial limitations to what you can accomplish with it relative to sports that lend themselves better to number crunching, like baseball.
There is some excellent work being done in basketball right now by the guys at the Sports Economist, and it has a similar fluidity and series of actions to hockey. I hear what you are saying, Mclea, but I think the possibility is there. Maybe I'm just hoping.
I think that statistical analysis in hockey has come a long way and probably still has a way to go but I also agree with McLea
As an example Mudcrutch had a post the other day about the value of a player to his team and one question was "how many games does a guy like Steve Staios actually win for the Oilers?" That's how I understood it anyhow.
Statistically not many probably.
But what value does a guy who plays with that much grit and determination have to his team? I noted that Lupul's play certainly reached a new level of effort the game after Staios took slapshot in the head and didn't miss a shift.
Maybe coincidence. I don't think so.
Of course, I guess that's why they call them intangibles.
Or to rehash an old argument that popped up constantly during the summer - the value of Dvorak - who I like btw - and his rates and stats.
A valuable player, I agree, but stats could show him to be comparable to a whole whack of guys who I would much much rather have then him on my team. Just common sense dictates it.
I think.
bdhs: Take any mc79 post about Staios or Moreau with a grain of salt. He can't stand them whether he'll admit it here or not.
Same goes for any post I make that comments on Stoll, Torres, or Andy Grabia's depth charts. Oh, throw Bill Guerin in there too. I always hated that guy. I hate him more than Brad Gilbert as a left side dman that's for sure.
In his defense, he never bothered to actually figure out Staios' contribution. If he had, he may have told a different story.
On the Dvorak thing - bdhs, how many of those other Dvorak comparables were getting paid considerably more than Dvorak? My bet is just about all of them, which was probably the crux of Tyler's argument and the ones I have made. Why pay twice the money for the names, when Dvorak will get you virtually equivalent results at half the price?
As for the stats thing, I agree with Matt that it's just a very useful tool. However, you have to be looking at the right stuff. If you look at the things that matter, they're better than 95% of the subjective crap that anyone spews on TV or on the internet.
RE: Moneypuck
1) Inability to isolate individual performance.
There may at this point be an inability to do this, but certainly I (like many others) are trying to find ways to isolate the player from the team. The icetime data should help.
2) Inability to categorize and grade "hockey actions" without reverting to a great deal of subjectivity.
Correlate to wins! If you can explain every win 100% of the time (impossible) then there is no subjectivity...
3) Too many players, too many variables, too many things going on that are impossible to substantiate.
Too many, ... never! Just needs more processing time.
RE: Phoenix:
Their problems are everywhere, and the net is not the worst of these problems (Cujo's been average and Morrison has been a lot better than his ~0.8 save percentage might indicate).
Biggest problem: they've allowed 3.9 worth of statistical goals based on shots against (independant of goaltender) per game (over 1 goal above average). They're scoring @ 2.15. So even with average goaltending they'd play at 23% with those expected goals. With bad goaltending a little lower.
It appears every player is bad (no line or group is doing well).
They need some defensive structure (and you don't always need great players for a great defensive system) to their game.
riversq - I am with you, don't get me wrong - and talking with a friend the other day about hockey I found myself talking a lot about what I have learned in the past year
and one of the guys who I did not get it about was Dvo, this is true; I just remember one post in particular where Tyler was throwing out some names as comps - now the money thing is a good point but I would still say I'd rather have 90% of the names he was tossing out there then Dvorak - some of that is old habits dying hard - I will admit that -
maybe if he had scored a goal last spring ...
I heard some guy talking about Phoenix last night and how Nolan was a plus player as if this made him a good signing; before I would have been impressed
instead I was "well, wait just a second here"
what about Jovo, MC? He's the other guy they were going on about, once again because he is a plus.
I presume the same - he is likely playing the Winchesters of the world and coming out slightly on top.
Am I right?
cynical joe - I was being facetious, a little
what I was saying is that my guess is that Jovo is likely a plus on a horrible team is that he is likely getting some easy minutes, which he should in some cases. At home Gretzky should be running him out against other teams' weak links to try and create some offence.
Plus my understanding is Michalek plays a lot of minutes against the other team's best players.
And Jovo is taking advantage of that, which he should.
I'd be interested to know if my theory is right - anyone? Tyler?
Post a Comment
<< Home
Three Reasons why there will never be a "Moneypuck":
1) Inability to isolate individual performance.
2) Inability to categorize and grade "hockey actions" without reverting to a great deal of subjectivity.
3) Too many players, too many variables, too many things going on that are impossible to substantiate.
That doesn’t mean statistical analysis is meaningless in hockey. It simply means that there are substantial limitations to what you can accomplish with it relative to sports that lend themselves better to number crunching, like baseball.
There is some excellent work being done in basketball right now by the guys at the Sports Economist, and it has a similar fluidity and series of actions to hockey. I hear what you are saying, Mclea, but I think the possibility is there. Maybe I'm just hoping.
I think that statistical analysis in hockey has come a long way and probably still has a way to go but I also agree with McLea
As an example Mudcrutch had a post the other day about the value of a player to his team and one question was "how many games does a guy like Steve Staios actually win for the Oilers?" That's how I understood it anyhow.
Statistically not many probably.
But what value does a guy who plays with that much grit and determination have to his team? I noted that Lupul's play certainly reached a new level of effort the game after Staios took slapshot in the head and didn't miss a shift.
Maybe coincidence. I don't think so.
Of course, I guess that's why they call them intangibles.
Or to rehash an old argument that popped up constantly during the summer - the value of Dvorak - who I like btw - and his rates and stats.
A valuable player, I agree, but stats could show him to be comparable to a whole whack of guys who I would much much rather have then him on my team. Just common sense dictates it.
I think.
bdhs: Take any mc79 post about Staios or Moreau with a grain of salt. He can't stand them whether he'll admit it here or not.
Same goes for any post I make that comments on Stoll, Torres, or Andy Grabia's depth charts. Oh, throw Bill Guerin in there too. I always hated that guy. I hate him more than Brad Gilbert as a left side dman that's for sure.
In his defense, he never bothered to actually figure out Staios' contribution. If he had, he may have told a different story.
On the Dvorak thing - bdhs, how many of those other Dvorak comparables were getting paid considerably more than Dvorak? My bet is just about all of them, which was probably the crux of Tyler's argument and the ones I have made. Why pay twice the money for the names, when Dvorak will get you virtually equivalent results at half the price?
As for the stats thing, I agree with Matt that it's just a very useful tool. However, you have to be looking at the right stuff. If you look at the things that matter, they're better than 95% of the subjective crap that anyone spews on TV or on the internet.
RE: Moneypuck
1) Inability to isolate individual performance.
There may at this point be an inability to do this, but certainly I (like many others) are trying to find ways to isolate the player from the team. The icetime data should help.
2) Inability to categorize and grade "hockey actions" without reverting to a great deal of subjectivity.
Correlate to wins! If you can explain every win 100% of the time (impossible) then there is no subjectivity...
3) Too many players, too many variables, too many things going on that are impossible to substantiate.
Too many, ... never! Just needs more processing time.
RE: Phoenix:
Their problems are everywhere, and the net is not the worst of these problems (Cujo's been average and Morrison has been a lot better than his ~0.8 save percentage might indicate).
Biggest problem: they've allowed 3.9 worth of statistical goals based on shots against (independant of goaltender) per game (over 1 goal above average). They're scoring @ 2.15. So even with average goaltending they'd play at 23% with those expected goals. With bad goaltending a little lower.
It appears every player is bad (no line or group is doing well).
They need some defensive structure (and you don't always need great players for a great defensive system) to their game.
riversq - I am with you, don't get me wrong - and talking with a friend the other day about hockey I found myself talking a lot about what I have learned in the past year
and one of the guys who I did not get it about was Dvo, this is true; I just remember one post in particular where Tyler was throwing out some names as comps - now the money thing is a good point but I would still say I'd rather have 90% of the names he was tossing out there then Dvorak - some of that is old habits dying hard - I will admit that -
maybe if he had scored a goal last spring ...
I heard some guy talking about Phoenix last night and how Nolan was a plus player as if this made him a good signing; before I would have been impressed
instead I was "well, wait just a second here"
what about Jovo, MC? He's the other guy they were going on about, once again because he is a plus.
I presume the same - he is likely playing the Winchesters of the world and coming out slightly on top.
Am I right?
cynical joe - I was being facetious, a little
what I was saying is that my guess is that Jovo is likely a plus on a horrible team is that he is likely getting some easy minutes, which he should in some cases. At home Gretzky should be running him out against other teams' weak links to try and create some offence.
Plus my understanding is Michalek plays a lot of minutes against the other team's best players.
And Jovo is taking advantage of that, which he should.
I'd be interested to know if my theory is right - anyone? Tyler?
Post a Comment
<< Home
