[News] USA Olympic Roster announced

[GirlySports]

if they're allowing reporters embedded into the process that will for sure lead to tough decisions then what is everyone apologizing for?

[BACKCHECK!!!]

Originally Posted by Textcritic

And that is that teams do not (and more appropriately should not) prioritise [advanced stats] as a tool for assembling their roster. In other words, we all recognise that the Toronto Maple Leafs presently enjoy an unsustainable level of puck possession time that shows they are a bubble playoff team. If they want to improve, it will not come by way of inserting players with great corsi or Fenwick ratings, but rather, by way of better, more skilled players.

You are now trying to make definitive claims of your own about causal relationships without providing any evidence. 19Yzerman19 is unable to produce convincing evidence that adding high-corsi players will result in a high-corsi team. But likewise you present no evidence to support your categorical claims to the contrary.


Adding a selection of high-corsi players will not improve a team's corsi?

Prove it.


My gut feeling is that the answer lies somewhere in between, but I'm not making statements like "will not", "don't", and "should not" because I don't have any convincing evidence to base it on.

[MissTeeks]

Paul Bissonnette‏@BizNasty2point02m

You can criticize me all you want @HockeyCanada. Just please don't ask me to spell anything.

[Textcritic]

Originally Posted by BACKCHECK!!!

...Adding a selection of high-corsi players will not improve a team's corsi?

Prove it...

I can't, since I don't have the means to conduct such an experiment. However, I would further argue that historically speaking, it does appear as though NHL teams are not constructed on the basis of individual players' corsi ratings. Until a NHL team clearly and unambiguously uses this criterion to select players as a means to improve puck possession, we can never know for sure, but so far, the fact that this does not seem to be a priority would minimally indicate something about the value of this method for improvement.

[Oling_Roachinen]

Originally Posted by 19Yzerman19

Well, there's the management team of the Toronto Maple Leafs.

I did say usually to leave room for arguing about quality shots, scoring chances and obviously the biggest difference being goalies.

I'm in the middle where I do think there's room for advanced stats, but I don't think you can discount watching players either. I realize there can be bias in watching players but you can still put heavy weight on it.

But when someone starts using Corsi as more than just support it's important to realize exactly what the data means and how to actually use it. Going by On-Ice Corsi this year Patrick Maroon is better than any forward on the Flames. Trying to fix that by going with Relative Corsi and you're left with Backlund being better than anyone on the Ducks. You can't easily compare Corsi across teams that's for sure. You may be able to use them to compare teammates, but then Hudler has a lower Corsi than Galiardi, and Monahan has one of the worst Corsi numbers.

Same with Fenwick % and the like, Team A has more shots than Team B with more offensive starts. They have more possession then and are therefore the better team. That's completely disregarding goalies though, if Team B has a 10M goalie stopping everything and Team A has a sieve it's not going to be an unexplained anomaly if Team B ends up above Team A in the standings.

[Tinordi]

Originally Posted by Enoch Root

I really don't want to get dragged into the stats debate, but I just couldn't let this quote go because it is so fundamental...

There is a difference between cause and correlation, which I am certain you understand, but you have erred here nonetheless.

There is absolutely no evidence whatsoever that building your team to have a higher corsi will lead to success (that would be a causal relationship).

What seems clear from the evidence is that if you build a better team, your corsi will go up (a correlative relationship).

GMs should focus on building better teams, not improving corsi.

I was waiting for you to chime in with exactly that point. Correlation is not causation.

Here's a timely article discussing just that. That yes, advanced stats are a very good predictor of future results and not simply describing what has recently happened.

http://www.sbnation.com/nhl/2013/10/...nalytics-teams

edit: wrong page

[Oling_Roachinen]

Originally Posted by Tinordi

I was waiting for you to chime in with exactly that point. Correlation is not causation.

Here's a timely article discussing just that. That yes, advanced stats are a very good predictor of future results and not simply describing what has recently happened.

http://www.sbnation.com/nhl/2014/1/2...el-assessments

Maybe all the folks who thanked your post will take a look as well.

What?

He's arguing that good teams may result in good advanced stats instead of good advanced stats resulting in good teams.

You're talking about past results helping to predict future results. It's not even remotely the same discussion.

[V]

I love it when people like strombad dispute stats by showing an observation that doesn't exactly fit the model. Like data is supposed to be homoscedastic, and if it isn't it's proven to be false. It just shows a very poor grasp of what statistics is, and is supposed to accomplish.

[strombad]

Originally Posted by Tinordi

I was waiting for you to chime in with exactly that point. Correlation is not causation.

Here's a timely article discussing just that. That yes, advanced stats are a very good predictor of future results and not simply describing what has recently happened.

http://www.sbnation.com/nhl/2013/10/...nalytics-teams

edit: wrong page

As I previously posted, it's been scientifically proven that statistics can not form a concrete predictor of human performance in sport. So, sorry, that article is garbage.

Personally, I'll always take information from actual scientists conducting actual experiments and writing actual scientific papers, over a blogger hyping corsi.

[strombad]

Originally Posted by V

I love it when people like strombad dispute stats by showing an observation that doesn't exactly fit the model. Like data is supposed to be homoscedastic, and if it isn't it's proven to be false. It just shows a very poor grasp of what statistics is, and is supposed to accomplish.

I love it when people like V think that statistics predict the future, rather than represent historical data, which is all that they do. It shows a very poor grasp of what statistics is, and is supposed to accomplish.

[V]

Originally Posted by strombad

Personally, I'll always take information from actual scientists conducting actual experiments and writing actual scientific papers, over a blogger hyping corsi.

Do you think that's any different? Practically every scientific experiment I've ever seen is using current data to predict future results.

Originally Posted by strombad

I love it when people like V think that statistics predict the future, rather than represent historical data, which is all that they do. It shows a very poor grasp of what statistics is, and is supposed to accomplish.

Are you 8?

[MissTeeks]

So is this thread even about the US Olympic team anymore? Poile on Hockey Central at Noon today:

Hope_Smoke‏@Hope_Smoke12m
Kypreos "the onus is on Scott Burnside to not publish things that will damage players"

Hope_Smoke‏@Hope_Smoke11m
Poile "I was assured that things like that wouldn't be published. You should have never heard those Burke comments about Ryan"

Hope_Smoke‏@Hope_Smoke10m
Poile "I talked to Bryan Murray and Bobby Ryan agents. Burke reached out to Ryan himself"

Hope_Smoke‏@Hope_Smoke9m
Kypreos "how do we avoid this from happening again? You thought you had an understanding and you got screwed"

Hope_Smoke‏@Hope_Smoke8m
Poile "I could have been knocked over by a feather when these article came out. I couldn't believe what was published"

Hope_Smoke‏@Hope_Smoke3m
Kypreos "Burke is devastated by this. They were invited in. And they as journalists know they are going to have to work in the future"

Hope_Smoke‏@Hope_Smoke2m
Kypreos "was it necessary to the story to print Ryan can't spell"

[sureLoss]

Originally Posted by GirlySports

if they're allowing reporters embedded into the process that will for sure lead to tough decisions then what is everyone apologizing for?

Apparently they were under the impression that harsh comments like that wouldn't make it into the article for public consumption.

[Tinordi]

Originally Posted by strombad

I love it when people like V think that statistics predict the future, rather than represent historical data, which is all that they do. It shows a very poor grasp of what statistics is, and is supposed to accomplish.

Wow, completely ignorant.

As I previously posted, it's been scientifically proven that statistics can not form a concrete predictor of human performance in sport. So, sorry, that article is garbage.

No true scotsman:

NOTHING is a concrete predictor. So therefore we shouldn't try to predict things? Ridiculous.

Personally, I'll always take information from actual scientists conducting actual experiments and writing actual scientific papers, over a blogger hyping corsi.

You know that "actual science" uses statistics as the basis for "actual experiments" very frequently?

This series of posts could not make you look worse on your understanding of the issue we're talking about.

[browna]

I thought it odd on how frank that article was at the time I was reading it, and would've assumed/thought something that behind the scenes would've passed through someone on Team USA's filter, first.

I am sure (given the Thornton thing earlier this year) beat reporters (which Burnside is) hear a lot of stuff that isn't for public consumption...and for their own self preserverence, don't dare publish that stuff.

But, who cares. Burke said what he was going to say to who he, and others, thought was behind a closed door audience, and he's probably said worse things face to face to players (albiet not as publicly). He doesn't have to apologize for anything...I'd be surprised if Burnside gets any more scoops in the league now though.

[kyuss275]

I feel sorry for Burke. If anything it sounds like he wanted Ryan on the team as a top 6 forward, but not as a bottom 6 forward. If anybody has watched the "Bobby Ryan story" at the end Burke is pretty proud of Ryan and even say's something along the effect of "having bragging rights" going forward. Sounds like a beat writer trying to make a name for himself publishing things he said he would not.

[sureLoss]

Originally Posted by browna

I thought it odd on how frank that article was at the time I was reading it, and would've assumed/thought something that behind the scenes would've passed through someone on Team USA's filter, first.

I am sure (given the Thornton thing earlier this year) beat reporters (which Burnside is) hear a lot of stuff that isn't for public consumption...and for their own self preserverence, don't dare publish that stuff.

But, who cares. Burke said what he was going to say to who he, and others, thought was behind a closed door audience, and he's probably said worse things face to face to players (albiet not as publicly). He doesn't have to apologize for anything...I'd be surprised if Burnside gets any more scoops in the league now though.

It is interesting that Kevin Allen was given the same access Burnside got, and he did not put the Burke quote into the article he published for USA Today.

[Resolute 14]

Some thoughts.

On Bobby Ryan: I agree that Burnside needs to be held accountable. Not necessarily for reporting on what was said - only an idiot would invite reporters in and not expect a story to be written - but for evidently reporting it out of context. That's dirty pool.


On Advanced Stats: A lot of the arguments on this thread are meaningless. The truth is, hockey's "advanced" stats are primitive and the people arguing that they aren't a replacement for direct observation are pretty much right. Corsi and Fenwick, for example, are variations on measuring net shot attempts. In short, they pretty much say "the more you have the puck, the better your chances of scoring, and the lesser their chance of scoring." Well no sh**. And we call this "advanced"? These stats are nice in that they offer a way of quantifying an individual's contribution to this on some level, but they don't actually tell us anything new in terms of how to build a team. For that reason, I agree with Strombad's argument. You don't look to build a team with high Corsi. You look to build a team with skilled players. The Corsi will follow naturally.

[Street Pharmacist]

Originally Posted by strombad

I love it when people like V think that statistics predict the future, rather than represent historical data, which is all that they do. It shows a very poor grasp of what statistics is, and is supposed to accomplish.

Haha. Yes. History has nothing to do with the future. I think it may be you who has a poor grasp on the issue

[V]

I would agree that #fancystats are fairly primitive. We are still in an early exploratory phase, but compared to where we were 10 years ago I think there has been a pretty impressive progression. While Corsi is flawed, it does begin to shed light on what's important in the game, and I believe that continued work in this field will yield some pretty interesting answers moving forward.