HFBoards

Go Back   HFBoards > General Hockey Discussion > By The Numbers
Mobile Hockey's Future Become a Sponsor Site Rules Support Forum vBookie Page 2
By The Numbers Hockey Analytics... the Final Frontier. Explore strange new worlds, to seek out new algorithms, to boldly go where no one has gone before.

Advanced Stats: Corsi, QoC, etc

Reply
 
Thread Tools
Old
08-09-2011, 07:02 PM
  #76
HarryNealesGarden
Big Daddy Ted
 
HarryNealesGarden's Avatar
 
Join Date: Nov 2009
Location: BOS
Country: United States
Posts: 4,244
vCash: 500
Quote:
Originally Posted by Kaner Coffee View Post
wouldn't these essentially just be league averages? This isn't the NFL where not every team plays each other.
Not even close. It would (I'm assuming, anyway) be weighted to the actual 82-game schedule, so it would be weighted to account for a team's division and conference opponents.

e.g. Vancouver's AOPK would be depressed (thereby making their PP% look worse) when you consider they played the two worst PK-ing teams in the NHL a combined 12 times.

HarryNealesGarden is offline   Reply With Quote
Old
08-09-2011, 07:14 PM
  #77
Garfinkel1
Registered User
 
Join Date: Jun 2008
Location: CT
Posts: 3,446
vCash: 500
Quote:
Originally Posted by Kaner Coffee View Post
wouldn't these essentially just be league averages? This isn't the NFL where not every team plays each other.
Nope.

Because teams don't play the same opponents the same amount of times.

For example - these numbers aren't actual...

Rangers might play Islanders 6 times and play Vancouver 2 times
BUT a team like the Bruins play the islanders only 3 times and Vancouver 3 times.

The Rangers A.O.PP will be lower than the Bruins
It takes into account the # of times they play each team and what teams they play.
*assuming Vancouver has a better PP than the Islanders - Who am I kidding, no need to assume. It's pretty obvious it's going to be .

Garfinkel1 is offline   Reply With Quote
Old
08-09-2011, 07:53 PM
  #78
Malkin4Top6Wingerz
Can you like, shutup
 
Malkin4Top6Wingerz's Avatar
 
Join Date: Mar 2009
Country: United States
Posts: 4,950
vCash: 500
Quote:
Originally Posted by hockeyfreak7 View Post
However, where I completely disagree with MFT6W is that an indication is not the be all, end all of a player's performance. There are a very significant amount of variables that come into play that simply make observation the one and only way to assess a player's performance.
See, logic like this frustrates me. You say that advanced statistics are not the end all be all of a player's performance, where I agree, and then you say that as a result observation is the one and only way to assess player performance. Why do people think this way? Just because a metric is not perfect does not mean it isn't significantly valuable, and if something is significantly valuable there is no reason to dismiss it outright.

Quote:
So where advanced statistics indicates a certain notion about a player, observation must be what confirms that notion. If you simply go by your advanced statistics, those variables can easily throw you off. Maybe in the future, advanced statistics will give an exact measure of a player's value to his team, but for now, the strongest way to analyze a player's value is by observing him.
Observation can throw you off just as easily. Focusing on a player who happens have a bad stretch of games will color them in a negative light, and this presents a problem because first impressions are difficult to shake. Likewise, catch a player during the best stretch of hockey he plays that season and you'll more than likely overrate him in the future. Unless you watch a player play ever night, you'll always have an incomplete picture of their game. That's not even taking into account all of the inherent biases that come with observation, of which there are many.

The other problem with citing observation is that it's not all that useful in a debate even if those observations are sound. When somebody cites something they observed about a player it's like being told that the sky is a certain color. If it looks different to me, then what? You say it looks red and I say it looks blue. We're at an impasse, clearly.

Quote:
Also keep in mind, if that wasn't the best way to analyze a player's worth, then what would be the purpose of scouts?
Baseball still has scouts despite statistics being overwhelmingly more value than observation.

Further, none of us here are scouts. In fact, I think most hockey fans are pretty clueless. Why are their observations valuable, let alone to the degree that advanced metrics are?

Malkin4Top6Wingerz is offline   Reply With Quote
Old
08-09-2011, 07:59 PM
  #79
Malkin4Top6Wingerz
Can you like, shutup
 
Malkin4Top6Wingerz's Avatar
 
Join Date: Mar 2009
Country: United States
Posts: 4,950
vCash: 500
Here's a couple articles that do a good job of addressing the downfalls of relying on observation.

http://www.coppernblue.com/2010/3/29...firmation-bias

http://www.coppernblue.com/2010/12/2...ew-of-the-game

Malkin4Top6Wingerz is offline   Reply With Quote
Old
08-09-2011, 08:28 PM
  #80
Rob Scuderi
Registered User
 
Rob Scuderi's Avatar
 
Join Date: Sep 2009
Country: United States
Posts: 2,762
vCash: 500
Quote:
Originally Posted by Malkin4Top6Wingerz View Post
I'm not sure why you feel like Corsi supporters believe all shots to be equal. As another poster noted, it is simply a measure of puck possession.
First of all thanks for the feedback. Like I said I'm not super knowledgeable about the different metrics various groups maintain. And again I'm not suggesting Corsi is useless just that it may not hold up so well against defensive minded players like Malholtra as you said.

What I meant by it being based on flawed assumptions is that the I really do not agree that measuring puck possession, by shot counts not actual time of possession which doesn't exist in the NHL, is really measuring puck possession. It's measuring shot counts or even worse shot attempts, which is a much much less useful piece.

Quote:
Defensive minded players do get undervalued by Corsi, but for different reasons than you stated. Excellent defenders are almost always matched up against the opposing team's best forwards, and are often used to take defensive zone draws and the like. This leads to lower raw Corsi ratings. Manny Malhotra was a good example of this last year. He got many of Vancouver's most difficult assignments and Vigneault had him take an absurd amount of defensive zone faceoffs. Judging by raw Corsi, you'd think he was hurting his team with his -9.5 Corsi. It is not until you look at the extremely difficult role he played before you see his true value to the Canucks.

Your overall concern about shot quality is one that many Corsi skeptics share. And I'll agree to an extent, not all shots are equal. With that said, shooting percentages tend to regress in the long run. Very few players can consistently shoot at a higher rate than their opponents, let alone to a significant enough degree that Corsi loses all of it's value. If you take a look at the year to year data, you'll see that all of these great defensive defenseman do not have the innate ability to suppress save percentage. Over one year it might appear that way, but the next it might look like the opposite because of the fickle nature of shooting percentage. Even the Lidstroms and Webers of the NHL let in goals at about the same rate as anybody else when shots are equal, so why would Rob Scuderi and Tony Lydman possess this talent?
I'm not suggesting Rob Scuderi or Toni Lydman are so talented defensively that they do a better job of stopping the flow of goals over a year by year period when compared to the most elite defenseman in the game. Even if I was that'd be a pretty bad group to compare them to. All I'm saying is players with a particular playing style which focuses on blocking shots and limiting greater scoring areas are then unfairly punished by Corsi.

Look at Greg Zanon, his corsi was atrocious but he was still second in the league in blocked shot with 212 and had 20 minutes of TOI. I don't have his QualComp numbers but can dig them up if this conversation goes anywhere, assuming I can find it. Of his 21 minutes a night he spent about 3 on the PK. So he got about 18 minutes a night, ES only, and made out with a pretty solid -5. That's an impressive number, looking at his teammates +/-, especially when you consider he was relied on in a shutdown role seeing good ES TOI.

If we just look at his Corsi, which was a league worst -429 then you'd probably conclude he had a god awful year, which Minnesota fans may contend was not the case.

I just think a guy like him who relies on shotblocking and playing strong in his own end ends up unfairly punished by Corsi. It doesn't matter at all how many times the other team attempts a shot while Zanon's out there. It matters how well he's doing in keeping the shots from a bad spot or if he's in a good position and can block it, which Corsi provides no credit for. What would be a much better indicator is one that accounts for those 200 blocks, like the Fenwick number you mentioned, but Corsi just does not do a solid job of measuring his ability in this isolated case.

Quote:
Again, Corsi is not based on that assumption.



Corsi only takes into account shot attempts, so players who block a lot of shots are undervalued to some degree. However another metric, Fenwick, remedies this concern by not counting blocked shots. Both have extremely similar predictive value, with Fenwick being a little more reliable over smaller samples and Corsi being a slightly better metric when the sample size is significant.
When I said Corsi treats all shots equal, I'm not sure how you disagree? What filters are there currently which account for weak shots versus higher percentage area shots? You even conceded it accounts for shots that don't even hit the net because they're blocked. How is that treating all shots equal? It's the exact opposite in my mind. It's making shots that clearly aren't equal, as they're in a bad area or got blocked, and treating them as if they're just as valuable as a shot coming off a breakaway. And if it counts all shots attempted like you said, then does that mean missed shots, not blocked shots, are included? If so it just gets less and less valid to me. I'm not saying it's useless but there's clear flaws with its application to certain players and I've read articles where these applications lend themselves to unfounded conclusions about players abilities.

Rob Scuderi is offline   Reply With Quote
Old
08-09-2011, 09:19 PM
  #81
The Fear Boners
Plz stop pucks
 
The Fear Boners's Avatar
 
Join Date: Jun 2007
Location: Scrip Club
Country: United States
Posts: 20,089
vCash: 500
They have them, but they're very much in infancy still.

The Fear Boners is offline   Reply With Quote
Old
08-09-2011, 09:40 PM
  #82
Malkin4Top6Wingerz
Can you like, shutup
 
Malkin4Top6Wingerz's Avatar
 
Join Date: Mar 2009
Country: United States
Posts: 4,950
vCash: 500
Quote:
Originally Posted by Bring Back Scuderi View Post
What I meant by it being based on flawed assumptions is that the I really do not agree that measuring puck possession, by shot counts not actual time of possession which doesn't exist in the NHL, is really measuring puck possession. It's measuring shot counts or even worse shot attempts, which is a much much less useful piece.
Corsi is highly correlated with zone time. More importantly, it is highly correlated with winning. Probably moreso than any kind of zone time data would be.

Quote:
I'm not suggesting Rob Scuderi or Toni Lydman are so talented defensively that they do a better job of stopping the flow of goals over a year by year period when compared to the most elite defenseman in the game. Even if I was that'd be a pretty bad group to compare them to. All I'm saying is players with a particular playing style which focuses on blocking shots and limiting greater scoring areas are then unfairly punished by Corsi.
So use Fenwick if it bothers you that much. Fenwick does not count blocked shots.

Quote:
Look at Greg Zanon, his corsi was atrocious but he was still second in the league in blocked shot with 212 and had 20 minutes of TOI. I don't have his QualComp numbers but can dig them up if this conversation goes anywhere, assuming I can find it. Of his 21 minutes a night he spent about 3 on the PK. So he got about 18 minutes a night, ES only, and made out with a pretty solid -5. That's an impressive number, looking at his teammates +/-, especially when you consider he was relied on in a shutdown role seeing good ES TOI.
Like Malhotra, Greg Zanon's horrid Corsi ratings are mostly a product of his role. He along with Schultz got the bulk of the heavy workload that is playing against the opposition's best players, and he did so while starting most of his shifts in the defensive zone. In addition to that, Minnesota was a poor possession team in general.

Quote:
If we just look at his Corsi, which was a league worst -429 then you'd probably conclude he had a god awful year, which Minnesota fans may contend was not the case.
This is why we apply proper context before analyzing possession metrics. Some players are put in situations that make outshooting the opposition very difficult. Zanon is one of the more extreme cases of this.

Quote:
I just think a guy like him who relies on shotblocking and playing strong in his own end ends up unfairly punished by Corsi. It doesn't matter at all how many times the other team attempts a shot while Zanon's out there. It matters how well he's doing in keeping the shots from a bad spot or if he's in a good position and can block it, which Corsi provides no credit for. What would be a much better indicator is one that accounts for those 200 blocks, like the Fenwick number you mentioned, but Corsi just does not do a solid job of measuring his ability in this isolated case.
Again, use Fenwick if you really believe this to be an issue. Since Corsi and Fenwick have almost identical predictive value, I tend to believe that neither one is much better than the other. Corsi does punish shotblockers to a degree, but blocking a lot of shots also means that their opponents are spending a lot of time in their end with the puck, and blocking a shot is not like blocking a pass in football. Play does not end there. Often times after a successful blocked shot the puck deflects to another player which leads to a scoring chance. And of course when a player goes down to block a shot and the puck still gets through, the goalie will have a more difficult time picking up the puck. That means more screened goals and rebound opportunities. So while shot blocking has value, there are disadvantages to playing that style, and the best way to keep the other team off the board is still to prevent shots, as has been proven time and time again.

Quote:
When I said Corsi treats all shots equal, I'm not sure how you disagree? What filters are there currently which account for weak shots versus higher percentage area shots? You even conceded it accounts for shots that don't even hit the net because they're blocked. How is that treating all shots equal? It's the exact opposite in my mind. It's making shots that clearly aren't equal, as they're in a bad area or got blocked, and treating them as if they're just as valuable as a shot coming off a breakaway. And if it counts all shots attempted like you said, then does that mean missed shots, not blocked shots, are included? If so it just gets less and less valid to me. I'm not saying it's useless but there's clear flaws with its application to certain players and I've read articles where these applications lend themselves to unfounded conclusions about players abilities.
What I meant was that while Corsi treats all shots as equal, the application of that data does not lie under that assumption. How you want to interpret the numbers is completely up to the user.

You can go back and forth with me expressing your concerns about shot quality not being properly weighted within possession numbers, but I assure you that argument has been made by countless Corsi skeptics and is never substantiated. That is because, by and large, shot quality has minimal impact. Valuable defenders do not give up lower quality scoring chances, they just give up less of them. If you want to scour BTN for defenseman who consistently outperform their Corsi ratings to a significant degree, be my guest. I'm betting you won't find any.

Malkin4Top6Wingerz is offline   Reply With Quote
Old
08-09-2011, 11:08 PM
  #83
hockeyfreak7
Registered User
 
hockeyfreak7's Avatar
 
Join Date: May 2010
Location: Charlottesville
Posts: 8,873
vCash: 500
Quote:
Originally Posted by Malkin4Top6Wingerz View Post
See, logic like this frustrates me. You say that advanced statistics are not the end all be all of a player's performance, where I agree, and then you say that as a result observation is the one and only way to assess player performance. Why do people think this way? Just because a metric is not perfect does not mean it isn't significantly valuable, and if something is significantly valuable there is no reason to dismiss it outright.
Ah, here we go again.

I didnt dismiss them out right, I only said that they are an indication not a confirmation.

They can suggest that either a player is, say, bad defensively, or tremendously unlucky. There are too many variables involved for these advanced stats to be an conclusive measure of performance.

It's like a pucking crossing the goal line under a pile of players. Did the puck probably go in? Yes, but we can not conclusively say that the puck crossed the line unless we have an angle that out right proves it.

The same logic applies here. Are these stats useful? I suppose. It gives players a uniformed metric to judge them on. But do these stats actually prove anything? Not unless they are in conjunction with actual observation that proves out right what a player's performance is like.



Quote:
Observation can throw you off just as easily. Focusing on a player who happens have a bad stretch of games will color them in a negative light, and this presents a problem because first impressions are difficult to shake. Likewise, catch a player during the best stretch of hockey he plays that season and you'll more than likely overrate him in the future. Unless you watch a player play ever night, you'll always have an incomplete picture of their game. That's not even taking into account all of the inherent biases that come with observation, of which there are many.
Correct. So what's the solution? Extend your sample size. That's what scouts do, right? I'm not saying make a judgment based on one game, but certainly over the course of a season you can draw some conclusions. Conclusions that can not simply be made by looking at advanced statistics.

Quote:
The other problem with citing observation is that it's not all that useful in a debate even if those observations are sound. When somebody cites something they observed about a player it's like being told that the sky is a certain color. If it looks different to me, then what? You say it looks red and I say it looks blue. We're at an impasse, clearly.
You are 100% correct. Observation of performance is subjective, as you have suggested.

So why try to objectify it with advanced statistics? What I value in a player may not be what you value or even perceive in a player as you astutely observed with your "sky is not blue" analogy.


Quote:
Baseball still has scouts despite statistics being overwhelmingly more value than observation.

Further, none of us here are scouts. In fact, I think most hockey fans are pretty clueless. Why are their observations valuable, let alone to the degree that advanced metrics are?
Correct. That's why scouts get paid and we dont. They can judge talent better than you or me, and they do it by observing players over the course of the season, not by reading the advanced statistics at the very end.

hockeyfreak7 is online now   Reply With Quote
Old
08-10-2011, 12:59 AM
  #84
Malkin4Top6Wingerz
Can you like, shutup
 
Malkin4Top6Wingerz's Avatar
 
Join Date: Mar 2009
Country: United States
Posts: 4,950
vCash: 500
Quote:
Originally Posted by hockeyfreak7 View Post
Ah, here we go again.

I didnt dismiss them out right, I only said that they are an indication not a confirmation.
How is that different from observational analysis exactly?

Quote:
They can suggest that either a player is, say, bad defensively, or tremendously unlucky. There are too many variables involved for these advanced stats to be an conclusive measure of performance.
Corsi has very little variance because of the large amount of events that it tracks. Goals and assists are actually much more luck oriented than possession metrics and deal with just as many variables, yet you and all of the other critics of advanced statistics use them to draw significant conclusions. Why is that?

Quote:
It's like a pucking crossing the goal line under a pile of players. Did the puck probably go in? Yes, but we can not conclusively say that the puck crossed the line unless we have an angle that out right proves it.
No one metric 'proves' that a player is valuable. And even between two so called 'experts' who 'watch the games', they will often have very different views when asked to analyze the same player. Same thing applies to fans who watch every game and have drastically different opinions about players on their own team. Who is right and who is wrong?

Let me put it this way: How am I supposed to prove somebody else's observations to be false?

Quote:
The same logic applies here. Are these stats useful? I suppose. It gives players a uniformed metric to judge them on. But do these stats actually prove anything? Not unless they are in conjunction with actual observation that proves out right what a player's performance is like.
You keep saying that advanced metrics dont 'prove' anything when, actually, they have proven many things. Even if they hadn't, what has observational analysis ever proven?

Quote:
Correct. So what's the solution? Extend your sample size. That's what scouts do, right? I'm not saying make a judgment based on one game, but certainly over the course of a season you can draw some conclusions. Conclusions that can not simply be made by looking at advanced statistics.
The drawbacks to observation go well beyond a small sample size. Take a look at the articles I linked a few posts up. Derek Zona describes the fallacies of observation and group think better than I ever could.

Quote:
You are 100% correct. Observation of performance is subjective, as you have suggested.

So why try to objectify it with advanced statistics? What I value in a player may not be what you value or even perceive in a player as you astutely observed with your "sky is not blue" analogy.
Because we want to know more about the game and what drives success? Seriously, what the hell kind question of that?

What you or anybody else value in a player is irrelevant - what matters is their actual value on the ice. Advanced statistics help us get to the root of where that value lies with actual hard evidence to back it up.

Quote:
Correct. That's why scouts get paid and we dont. They can judge talent better than you or me, and they do it by observing players over the course of the season, not by reading the advanced statistics at the very end.
You do know that Corsi wasn't invented by some stat nerd with too much time on his hands, right? Current goaltending coach for the Buffalo Sabres Jim Corsi came up with the stat, and he knows quite a bit about the game. Your assertion that scouts and talent evaluators rely purely on what they see and do not read advanced statistics is like most of what your argument is predicated on - hand waving. There is absolutely no evidence that teams ignore this kind of data, and in a game where exploiting every edge can make all the difference in a parity filled league, such a claim seems absolutely preposterous.

I can buy that clowns like Mike Milbury and Craig Button who recite hockey cliches in their sleep relied on this kind of analysis to make their decisions when they were GM. St. Louis and Savard were clearly too small to ever make an impact at the NHL level, Chara was just an ogre on skates who would never amount to anything, and DiPietro was going to be the next Patrick Roy. Gotta love the infallability of observation!

Malkin4Top6Wingerz is offline   Reply With Quote
Old
08-10-2011, 03:04 AM
  #85
Ogopogo*
 
Ogopogo*'s Avatar
 
Join Date: Apr 2005
Location: Edmonton
Country: Canada
Posts: 14,214
vCash: 500
Quote:
Originally Posted by Doctor No View Post
Do tell.
Corsi is simply +/- using shots on goal. Correct?

+/- is a team stat that tells you virtually nothing about an individual - taking that same flaw and applying it to shots on goal tags yet another team stat on an individual, giving us a number that is not representative of the individual at all.


Last edited by Ogopogo*: 08-10-2011 at 03:33 AM.
Ogopogo* is offline   Reply With Quote
Old
08-10-2011, 03:25 AM
  #86
MGIII
Legendary
 
Join Date: Jun 2010
Location: Buffalo, NY
Posts: 774
vCash: 500
The day we use a better system to predict future occurrences than raw statistics will be a glorious day indeed. An evolution of statistics.

MGIII is offline   Reply With Quote
Old
08-10-2011, 12:42 PM
  #87
Hipster Doofus
Registered User
 
Hipster Doofus's Avatar
 
Join Date: Sep 2006
Country: United States
Posts: 6,450
vCash: 500
Quote:
Originally Posted by hockeyfreak7 View Post
Also keep in mind, if that wasn't the best way to analyze a player's worth, then what would be the purpose of scouts?
In baseballs scouts map and chart pitches or confirm how plays got recorded (i.e. bad calls ect.). The advanced stats in baseball are sophisticated enough that anyone who attempts to use **** like wins and rbi's immediately comes off as outdated. Most teams use many different metrics all in an attempt to find the bets value.

For hockey, this would be like having scouts who map players movements. This is what Gretzky did as a kid, perhaps the most broken player in sports history. With hockey the need for scouts would be that much more important because you need someone to map and graph the game. Right now all our stats are from arena officials (hits, shots, ect.) or games crew (goals assists, PIM's) and such and most of these metrics generally involve using raw data recorded from these sources.

Hipster Doofus is offline   Reply With Quote
Old
08-10-2011, 03:03 PM
  #88
MathMan
Registered User
 
Join Date: Jan 2006
Posts: 17,037
vCash: 500
Quote:
Originally Posted by Hipster Doofus View Post
Right now all our stats are from arena officials (hits, shots, ect.) or games crew (goals assists, PIM's) and such and most of these metrics generally involve using raw data recorded from these sources.
There's an initiative being made to record scoring chances, defined as hard plays towards the net from an area extending from the posts to the faceoff dots, and from there to the top of the circles (missed shots count, blocked shots do not). But it's very labor-intensive so only a handful of teams are being so examined. It's generated a lot of valuable information, however, and it's also shown that these scoring chances are highly correlated to the Corsi/Fenwick metrics.

MathMan is offline   Reply With Quote
Old
02-03-2012, 05:57 PM
  #89
Erick
Registered User
 
Erick's Avatar
 
Join Date: Nov 2009
Location: Broward, FL
Country: United States
Posts: 12,251
vCash: 500
Sabremetric hockey stats?

Anyone interested in hockey analytics?

What's the best site?

I've seen hockeyprospectus but it's not updated, and you gotta pay $20 apparently for that.

Behind the Net seems pretty popular for the CORSI stat, but I was looking for other things (maybe I'm missing them).

Anyone recommend a better site if there is one? Thanks.

Erick is offline   Reply With Quote
Old
02-03-2012, 06:02 PM
  #90
Jason MacIsaac
MARS Officer
 
Jason MacIsaac's Avatar
 
Join Date: Jan 2004
Location: Victoria, BC
Country: Canada
Posts: 16,476
vCash: 500
Send a message via MSN to Jason MacIsaac Send a message via Yahoo to Jason MacIsaac
behindthenet.ca, timeonice.com, arcticicehockey and any sites affiliated with it.

Jason MacIsaac is offline   Reply With Quote
Old
02-03-2012, 06:02 PM
  #91
Loso
Registered User
 
Loso's Avatar
 
Join Date: Mar 2010
Country: Canada
Posts: 3,028
vCash: 500
Not a huge fan.

Hockey has many more immesurable aspects than baseball IMO.

Runs are scored at most 10 different ways, goals are scored zillions of different ways. Useful if used properly for hockey but there will never be a stat like WAR.

Loso is offline   Reply With Quote
Old
02-03-2012, 06:04 PM
  #92
Jason MacIsaac
MARS Officer
 
Jason MacIsaac's Avatar
 
Join Date: Jan 2004
Location: Victoria, BC
Country: Canada
Posts: 16,476
vCash: 500
Send a message via MSN to Jason MacIsaac Send a message via Yahoo to Jason MacIsaac
Quote:
Originally Posted by Loso View Post
Not a huge fan.

Hockey has many more immesurable aspects than baseball IMO.

Runs are scored at most 10 different ways, goals are scored zillions of different ways. Useful if used properly for hockey but there will never be a stat like WAR.
If you know how to use sabremetrics they are very telling and you can get a great grasp on a player without watching him play.

Jason MacIsaac is offline   Reply With Quote
Old
02-03-2012, 06:07 PM
  #93
MadLuke
Registered User
 
Join Date: Jan 2011
Posts: 2,299
vCash: 500
Quote:
Originally Posted by Jason MacIsaac View Post
If you know how to use sabremetrics they are very telling and you can get a great grasp on a player without watching him play.
But not as much of baseball, baseball must the sport with the stats describing the most accurately a player quality, soccer the less.

MadLuke is offline   Reply With Quote
Old
02-03-2012, 06:15 PM
  #94
Jason MacIsaac
MARS Officer
 
Jason MacIsaac's Avatar
 
Join Date: Jan 2004
Location: Victoria, BC
Country: Canada
Posts: 16,476
vCash: 500
Send a message via MSN to Jason MacIsaac Send a message via Yahoo to Jason MacIsaac
Quote:
Originally Posted by MadLuke View Post
But not as much of baseball, baseball must the sport with the stats describing the most accurately a player quality, soccer the less.
Oh I agree. Hockey stats are also 25 years behind baseball, we are only in the early stages.

Jason MacIsaac is offline   Reply With Quote
Old
02-03-2012, 06:17 PM
  #95
Roo Mad Bro
U havin a giggle m8?
 
Roo Mad Bro's Avatar
 
Join Date: Dec 2010
Location: PA
Country: United States
Posts: 6,015
vCash: 500
I'm a huge fan & proponent of baseball Sabremetrics, but I have never really got into hockey Sabremetrics.

Roo Mad Bro is offline   Reply With Quote
Old
02-03-2012, 06:30 PM
  #96
Unaffiliated
Registered User
 
Unaffiliated's Avatar
 
Join Date: Aug 2010
Location: Richmond, B.C.
Country: Canada
Posts: 10,235
vCash: 500
Quote:
Originally Posted by Jason MacIsaac View Post
behindthenet.ca, timeonice.com, arcticicehockey and any sites affiliated with it.
Good post.

Unaffiliated is offline   Reply With Quote
Old
02-03-2012, 06:34 PM
  #97
serge2k
Registered User
 
serge2k's Avatar
 
Join Date: Sep 2006
Posts: 14,401
vCash: 500
Quote:
Originally Posted by Jason MacIsaac View Post
If you know how to use sabremetrics they are very telling and you can get a great grasp on a player without watching him play.
Or you can use them to get a viewpoint that makes absolutely no sense if you ever actually bother to watch the player.

serge2k is offline   Reply With Quote
Old
02-03-2012, 06:38 PM
  #98
Erick
Registered User
 
Erick's Avatar
 
Join Date: Nov 2009
Location: Broward, FL
Country: United States
Posts: 12,251
vCash: 500
Quote:
Originally Posted by Jason MacIsaac View Post
behindthenet.ca, timeonice.com, arcticicehockey and any sites affiliated with it.
Thanks, I'll look into them.

I agree that hockey is more of a team game than baseball and Sabremetrics may never reach the status they've reached in baseball.

However, I've been searching for the meaning of some advanced hockey stats and I personally thought it'd be interesting to see which players are overvalued/undervalued, etc.

Advanced stats in hockey may not be perfect, but there has to be a way to evaluate someone without personal bias or without using stats like +/- without context. Or even basic goaltending stats which are rather overvalued. GAA being too team-dependent at times and save % valuing every shot equally. They don't really tell me much in the grand scheme of things.

Erick is offline   Reply With Quote
Old
02-03-2012, 06:53 PM
  #99
Jason MacIsaac
MARS Officer
 
Jason MacIsaac's Avatar
 
Join Date: Jan 2004
Location: Victoria, BC
Country: Canada
Posts: 16,476
vCash: 500
Send a message via MSN to Jason MacIsaac Send a message via Yahoo to Jason MacIsaac
Quote:
Originally Posted by serge2k View Post
Or you can use them to get a viewpoint that makes absolutely no sense if you ever actually bother to watch the player.
That rarely happens if you use all stats to your disposal.

Jason MacIsaac is offline   Reply With Quote
Old
02-03-2012, 06:54 PM
  #100
Sivek
Registered User
 
Sivek's Avatar
 
Join Date: Apr 2011
Posts: 3,268
vCash: 500
Love how any posts concerning sabremetrics in hockey is quickly filled with posts decrying how useless they are. inb4 how using stats=living in mom's basement and never watched or played a game before and so on. People can enjoy stats and the actual game of hockey, it's true.

Anyways, the Hockey Prospectus annual is worth buying if you're interested in sabremetrics. Whether you agree with their opinions or not, having a wider breath of information can only help in evaluating players for your own amusement and gain a better understanding of what certain indicators in a players play may lead to. As noted already, hockey sabremetrics is really behind the ball when compared to other sports since their widespread discussion hasn't been really relevant until recent years, which in my opinion, makes it something interesting to follow since as a fan you can follow how sabremetrics are being crafted at the beginning rather than already having a nearly finished formula, as in baseball.

Also, some stats that really seem to have come on strong in the last couple of years are "Sheltered ice-time" for skaters and even strength save% for goalies if you're looking at some stats to start to with. Sheltered ice time would be extremely useful if there were only a quick & easy way to figure it out.


Last edited by Sivek: 02-03-2012 at 07:00 PM.
Sivek is offline   Reply With Quote
Reply

Forum Jump


Bookmarks

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off



All times are GMT -5. The time now is 12:15 AM.

monitoring_string = "e4251c93e2ba248d29da988d93bf5144"
Contact Us - HFBoards - Archive - Privacy Statement - Terms of Use - Advertise - Top - AdChoices

vBulletin Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
HFBoards.com is a property of CraveOnline Media, LLC, an Evolve Media, LLC company. 2014 All Rights Reserved.