Why obsessing over Corsi will lead to failure

[King Mapes]

Watch the move MoneyBall and research Billy Beane. One of the biggest and brightest minds in the game came up some of this stuff and has been very successful.

I believe hockey ones are a lot more questionable. In baseball it's not how you hit it but where you hit it. In hockey we've seen poor corsi teams can continually be at the top of the league.

 

[Pookie]

Watch the move MoneyBall and research Billy Beane. One of the biggest and brightest minds in the game came up some of this stuff and has been very successful.

One of my favs.

The thing I like about Moneyball is the focus on doing things differently. Sure, they used stats to justify their decisions but also challenged (out of necessity) the notion that you needed star players to have a winning team. A's do it. Tampa does it. Kansas was sure fun to watch weren't they?

Buying based on a need is sort of like how I view Kessel. He will be 30 by the time the team is even remotely in the playoff hunt again. Why do we need a "star" player?

Challenged the same thing with TFC and Jermain Defoe last year. Great player. Injured a lot in the time leading up to his contract. MLS doesn't give relief for injured Designated Players which, IMO, made that a bad contract. TFC needed a goal scorer but there were better choices. (Of course it was MLSE so they also needed to sell tickets so there you go).

When you buy a star, you need to be filling a specific reason. 28 year old "stars" on the Leafs make no sense right now. We need to be buying for the future.

Anyways, great movie. Remember though that Beane had some rough years where the team didn't perform well. You could say it's a philosophy... not quite an infallible science.

 

[leafstilldeath*]

I find it extremely amusing that all the 30 teams in the NHL have embraced "advanced" stats. And now NHL.com too has come out with "Enhanced Stats" and the naysayers still keep on calling out stats (which is evidence based: i.e. FACTS) and want to force that opinions are the way to go (i.e. No Facts: I just feel like it).

riiiggghhhttttttttt (Austin Powers accent and voice)

 

[Eb]

I find it extremely amusing that all the 30 teams in the NHL have embraced "advanced" stats. And now NHL.com too has come out with "Enhanced Stats" and the naysayers still keep on calling out stats (which is evidence based: i.e. FACTS) and want to force that opinions are the way to go (i.e. No Facts: I just feel like it).

riiiggghhhttttttttt (Austin Powers accent and voice)

I feel like you contradicted yourself here. Do you not feel like that if the NHL themselves are not calling them advanced statistics, then they really aren't advanced at all?

 

[Pookie]

NHL.com publishes a lot of things. I'm not going to go all ******* over +/- or leading/trailing after the first period.

 

[King Mapes]

NHL.com publishes a lot of things. I'm not going to go all ******* over +/- or leading/trailing after the first period.

I was coming to post those examples. It doesn't make it more accurate it just means more fans know and understand them so they tossed them on the site.

 

[Menzinger]

People get caught way too up on terminology. The 'advanced' prefix isn't important - what matters is that they help you understand the game better than before.

Corsi is incredibly simple to understand - it's not 'advanced' yet that doesn't mean it's not incredibly useful.

 

[leafstilldeath*]

I feel like you contradicted yourself here. Do you not feel like that if the NHL themselves are not calling them advanced statistics, then they really aren't advanced at all?

ummmm.... they weren't "advanced" to begin with. they were pure observations in a way that hockey world never used to observe (except: Hawks, Kings. they had these stats before they were out for everyone to see).

Second, advanced stats as they are available in baseball (think WAR type stats) are still in their infancy in hockey. They are still being worked on and tested.

Finally, the stats have been pretty good with predictions.

 

[Caesium]

Watch the move MoneyBall and research Billy Beane. One of the biggest and brightest minds in the game came up some of this stuff and has been very successful.

I believe hockey ones are a lot more questionable. In baseball it's not how you hit it but where you hit it. In hockey we've seen poor corsi teams can continually be at the top of the league.

My problem with corsi is that people point to baseball and say that stats are a good way of building teams. They're pointing to the first fancy stat they can find and declaring it the holy grail. Hockey stats aren't even in their infancy yet, the stats that will actually lead to being able to construct a winning team are not tracked. What evidence is there that corsi is the same as on base percentage? If corsi is analogous to on base percentage, then why does a player's corsi vary drastically from year to year and team to team? Can a stat that varies so much have any value at all when it comes to constructing a team?

David Clarkson's corsi was better than any Leaf the year before we signed him, at 23. The next highest Leaf had a corsi of -1. In fact, Clarkson's corsi rating puts him at 10th overall in the league that year. Surely, SURELY, if corsi was a good measure like baseball's on base percentage, he must be a really good NHL player and would vastly improve any team he were added to.

 

[King Mapes]

I won't argue that, it is useful when used right, much like stats like plus minus. My problem is some live or die with it and once a team loses a couple it's "I told you so". But yes it can be useful I won't argue that. Evidence in the fact teams do use it.

 

[The Winter Soldier]

I find it extremely amusing that all the 30 teams in the NHL have embraced "advanced" stats. And now NHL.com too has come out with "Enhanced Stats" and the naysayers still keep on calling out stats (which is evidence based: i.e. FACTS) and want to force that opinions are the way to go (i.e. No Facts: I just feel like it).

riiiggghhhttttttttt (Austin Powers accent and voice)

This is actually Not true, the Anaheim Ducks do not employ an advanced stats department nor does their coach put too much emphasis on them. They are doing pretty well this year.

http://blogs.windsorstar.com/sports/duff-value-of-hockey-analytics-under-debate

 

[Pookie]

Jets vs Leafs last night was one of the more entertaining games of the year. Played completely wrong by the Corsi enthusiasts as the Jets outshot the Leafs by 1.

But I'll tell you that edge in that game was the reason that I didn't turn the game off like I have done so many times before.

Let's find a way to measure and duplicate that.

 

[leafstilldeath*]

Jets vs Leafs last night was one of the more entertaining games of the year. Played completely wrong by the Corsi enthusiasts as the Jets outshot the Leafs by 1.

But I'll tell you that edge in that game was the reason that I didn't turn the game off like I have done so many times before.

Let's find a way to measure and duplicate that.

Learn how to read stats or for that matter what is Corsi

http://hockeystats.ca/game/2014020875

Leafs Corsi>Jets Corsi

FWIW Jake Gardiner once AGAIN with 50+ CF% at 68% post the toaster of a coach Carlyle era.

#StatsMatter!

 

[Menzinger]

My problem with corsi is that people point to baseball and say that stats are a good way of building teams. They're pointing to the first fancy stat they can find and declaring it the holy grail. Hockey stats aren't even in their infancy yet, the stats that will actually lead to being able to construct a winning team are not tracked. What evidence is there that corsi is the same as on base percentage? If corsi is analogous to on base percentage, then why does a player's corsi vary drastically from year to year and team to team? Can a stat that varies so much have any value at all when it comes to constructing a team?

David Clarkson's corsi was better than any Leaf the year before we signed him, at 23. The next highest Leaf had a corsi of -1. In fact, Clarkson's corsi rating puts him at 10th overall in the league that year. Surely, SURELY, if corsi was a good measure like baseball's on base percentage, he must be a really good NHL player and would vastly improve any team he were added to.

No they aren't. Read any serious proponent of so called 'advanced' stats, and they continuously say that it is useful but not the only thing you should pay attention to. They aren't a substitute for watching the game, they're meant to help enhance your ability to analyze the play that's it.

As for Clarkson, there are statistical anomalies with everything - not just in terms of sports data. He had one really good year in terms of corsi - that's it. And again, a good example that you obviously still need to watch the games. to get the full picture.

Here's an interesting (and useful) article that breaks down Clarkson's advanced stats
http://corsihockeyleague.com/2014/09/15/david-clarkson-what-happened-why-and-what-to-expect-next/

 

[leafs in five]

couple pages back he disproved the 13-14 Kings CF% superiority by counting SOGs as well

 

[Pookie]

Learn how to read stats or for that matter what is Corsi

http://hockeystats.ca/game/2014020875

Leafs Corsi>Jets Corsi

FWIW Jake Gardiner once AGAIN with 50+ CF% at 68% post the toaster of a coach Carlyle era.

#StatsMatter!

Really? 61 to 60 with a dominance in OT is the way Corsi enthusiasts are recommending the way to play?

 

[leafstilldeath*]

Really? 61 to 60 with a dominance in OT is the way Corsi enthusiasts are recommending the way to play?

Do you understand that the teams that were within top 3 places in the NHL w.r.t. Corsi in the last 5 to 6 seasons have gone on to win the cup? (exception: Penguins).

Don't take my word for it just go ahead and do the research yourself.

Just FYI if these stats weren't so good at predicting (ofcourse not 100% but still fairly high) people wouldn't even give a crap about these stats.

Instead people care about these stats because these stats provide "Value" which the so called opinions "Do Not"

 

[achtungbaby]

Pinkie is jerking some chains.

 

[Pyromaniac3]

David Clarkson's corsi was better than any Leaf the year before we signed him, at 23. The next highest Leaf had a corsi of -1. In fact, Clarkson's corsi rating puts him at 10th overall in the league that year. Surely, SURELY, if corsi was a good measure like baseball's on base percentage, he must be a really good NHL player and would vastly improve any team he were added to.

As always, you need to use context behind every stat. Blindly taking a single stat and using it to evaluate a player makes no sense at all. And no one said Corsi is a perfect stat to rate a player.

 

[Pookie]

Do you understand that the teams that were within top 3 places in the NHL w.r.t. Corsi in the last 5 to 6 seasons have gone on to win the cup? (exception: Penguins).

Don't take my word for it just go ahead and do the research yourself.

Just FYI if these stats weren't so good at predicting (ofcourse not 100% but still fairly high) people wouldn't even give a crap about these stats.

Instead people care about these stats because these stats provide "Value" which the so called opinions "Do Not"

You mean Cup favourite teams tend to outshoot their opponents, score more and tend to win close games? And these cup favourite teams actually sometimes go on to win the cup.

Good teams have good stats? And bad teams, if they had the stats of good teams, then they would be good?

Ground breaking and riveting all at the same time.

 

[leafstilldeath*]

You mean Cup favourite teams tend to outshoot their opponents, score more and tend to win close games? And these cup favourite teams actually sometimes go on to win the cup.

Good teams have good stats? And bad teams, if they had the stats of good teams, then they would be good?

Ground breaking and riveting all at the same time.

and if the stats say the same story you have a problem with that?

I don't get it

 

[King Mapes]

This seems like a never ending circle with the same arguments back and forth. I think its run its course.

Closed