By the numbers: How the Jets skaters performed

[garret9]

I've gotten a hold of two new and very informative techniques... so of course I applied them to the Jets players:

http://www.arcticicehockey.com/2014...player-stats-evaluation-corsi-goals-posession

As always, please as questions or give your thoughts on what the numbers say.

Click to read more...

 

[garret9]

83+ reads (and according to clicks most went to the site) and no one is saying a thing?

Any shocks? Any surprises? Any thing as expected?

Anything you learnt?

 

[pucka lucka]

83+ reads (and according to clicks most went to the site) and no one is saying a thing?

Any shocks? Any surprises? Any thing as expected?

Anything you learnt?

I'd like to see Scheifele's numbers in intervals.

 

[garret9]

I'd like to see Scheifele's numbers in intervals.

That'll take a bit but I know with Kane his numbers are:

19.95 CF20

19.20 CA20

51.0 Corsi%

So you can take a peek at the tables with that.

Also:

 

[pucka lucka]

That'll take a bit but I know with Kane his numbers are:

19.95 CF20

19.20 CA20

51.0 Corsi%

So you can take a peek at the tables with that.

I think you know what I am getting at. Has he been getting better?

 

[garret9]

I think you know what I am getting at. Has he been getting better?

I just gave his raw cumulative Corsi% in the edit, which does make it seem like he's getting better, but the extent is hard to tell with that since usage, teammates, ozs, opponents aren't being taken into account.

 

[pucka lucka]

I just gave his raw cumulative Corsi% in the edit, which does make it seem like he's getting better, but the extent is hard to tell with that since usage, teammates, ozs, opponents aren't being taken into account.

We know playing with Thorburn didn't help. I just wanted to see how well it's matched my eyes. Looks like my eyes can see in Corsi

 

[Gm0ney]

It looks like the numbers are saying that Postma is better at suppressing shots attempts against that Bogo, Trouba and Stu? This doesn't seem to pass the eyeball test...

 

[theamazingchris]

I assume Postma' numbers are a sampling issue, he hasn't played much this year and what he did play was sheltered (I think).

 

[garret9]

It looks like the numbers are saying that Postma is better at suppressing shots attempts against that Bogo, Trouba and Stu? This doesn't seem to pass the eyeball test...

Sampling issue is most likely cause. His number of CA20 is WAY lower than last season and Postma only has 100 5v5 mins (everyone else is between 500 to 1000).

The regressions don't seem to work well on extreme usages either (see Pardy with super super soft sheltered minutes).

 

[theamazingchris]

I'm sorry for not commenting more, I know that's a pet peeve of article writers. But frankly, it's because your articles always make a lot of sense to me, haha. Everything lays out exactly as expected.

One thing that slightly surprised me was Setoguchi's numbers, though. They're lower than even I expected. That defensive number especially.

 

[garret9]

Ya Setoguchi does seem to be a bit of a surprise. I expected those trends with him, but not that severe.

I do find it interesting when the trends align to what I see.

I did some work recently on hockey-graphs re: Corsi and coaches giving out icetime, and there def is a correlation...

It does seem interesting how much what hockey people say already aligns with the numbers. The numbers just let you quantify the degrees.

What is even more interesting though is when the numbers and perception don't align.

 

[Puckatron 3000]

The first thing that jumped out at me was Wheeler's numbers. I just sort of expect any analysis of Wheeler to be very positive, and it isn't here. You address this:

Wheeler has never been a player that can push the play forward as well as the previously mentioned players; he does however score very well, especially when placed with possession players.

Which got me thinking about Corsi's use of shot attempts and puck possession. In the Wheeler case, is the Corsi gauge misleading? Certainly, puck possession is very, very important. But do we not also have to consider the quality of shots/puck possession as well? If Wheeler has less shots/possession, but he makes more of those shots/possession count (goals), won't his Corsi be deceptively low as a measure of his team contribution?

Should the true measure of a forward be some combination of goals/assists/points + Corsi? If so, what weight do we give to each factor?

 

[Hobby Bull]

83+ reads (and according to clicks most went to the site) and no one is saying a thing?

Any shocks? Any surprises? Any thing as expected?

Anything you learnt?

Firstly, thanks. It is interesting stuff.

I don't think that CORSI is a perfect tool for the analysis of player performance. As in most things, it only shows what it shows, and doesn't show what it doesn't show. People will make inferences from the data that aren't necessarily valid. As a tool, it generally works, but sometimes it just doesn't pass the eyeball test.

Kane vs. Wheeler in this case make the best examples I can give for player styles skewing CORSI numbers. Briefly stated, if Kane is in possession of the puck, he's going to shoot. If Wheeler is in possession of the puck, he might never shoot. In a world where, in fact, they "drove possession" equally, Kane would have a far superior CORSI.

I would not evaluate any player solely on CORSI data.

 

[garret9]

Just lil' but non-important FYI and history lesson, Corsi is a proper name not an acronym. Named after Jim Corsi who tracked it to weigh a goaltenders workload, since goalies react to all shot attempts, not just the ones that they save.

There is things that Corsi does not cover. You are right there and in being that you do not evaluate players solely on COrsi. However, the one you have reasoned out is false.

One: Wheeler actually tends to take a lot of shot attempts. He's been one of the highest Jets 2.0 after Kane, usually competing against Ladd and Frolik.
Two: There isn't really a correlation to guys who take lots of shot attempts and Corsi. Corsi% tends to correlate strongest to puck possession and a guy who does nothing but throw away the puck will not improve his numbers. As Eberle said: if you do the right things, like improve your breakouts and zone entries, your Corsi will improve, but if you work on trying to improve your Corsi, you will likely fail.

I never use Corsi on it's own; however, out chancing and out puck possessing is approximately 50% of what gets you wins, which means Corsi is about equally important as everything else combined... but NOT the whole picture.

As I said in the article:

There are good reasons to evaluate players with Corsi%:

1) Most people know who can and can't score intuitively quite easily, but it's much hard to tell who is assisting or harming puck possession

2) A player's affect on Corsi% is the input a player gives towards affecting the team's goal differential, and because of this we can extrapolate how much a player is affecting their team overall

 

[theamazingchris]

Just lil' but non-important FYI and history lesson, Corsi is a proper name not an acronym. Named after Jim Corsi who tracked it to weigh a goaltenders workload, since goalies react to all shot attempts, not just the ones that they save.

There is things that Corsi does not cover. You are right there and in being that you do not evaluate players solely on COrsi. However, the one you have reasoned out is false.

One: Wheeler actually tends to take a lot of shot attempts. He's been one of the highest Jets 2.0 after Kane, usually competing against Ladd and Frolik.
Two: There isn't really a correlation to guys who take lots of shot attempts and Corsi. Corsi% tends to correlate strongest to puck possession and a guy who does nothing but throw away the puck will not improve his numbers. As Eberle said: if you do the right things, like improve your breakouts and zone entries, your Corsi will improve, but if you work on trying to improve your Corsi, you will likely fail.

I never use Corsi on it's own; however, out chancing and out puck possessing is approximately 50% of what gets you wins, which means Corsi is about equally important as everything else combined... but NOT the whole picture.

Quick question, I remember reading this somewhere, but it escapes me. Do you know who calculated this, and what the other components/percentages are? A link would be much appreciated also if you have one.

 

[garret9]

Quick question, I remember reading this somewhere, but it escapes me. Do you know who calculated this, and what the other components/percentages are? A link would be much appreciated also if you have one.

I believe Tom Awad of Hockey Prospectus.

I believe the breakdown was:
~50% outchance/possess
~30% variance/luck
~20% goaltending talent, shot quality talent, special teams, faceoffs, etc.

 

[theamazingchris]

I believe Tom Awad of Hockey Prospectus.

I believe the breakdown was:
~50% outchance/possess
~30% variance/luck
~20% goaltending talent, shot quality talent, special teams, faceoffs, etc.

Got it. And I don't.

What does this mean exactly? Is it to say a team that outpossesses every other team consistently will win 50% of its games? Or is that wrong?

Also, has there been any work on parsing down that last 20%? I'm sorry to abuse you like this, but I'm not in the right circles.

 

[Puckatron 3000]

That's an interesting Eberle quote. I haven't heard players talk about Corsi much.

I'm still trying to get my head around exactly what Corsi should be telling me, so these conversations are helpful. I feel like I need a handle on that before I can talk about the actual numbers.

 

[garret9]

Got it. And I don't.

What does this mean exactly? Is it to say a team that outpossesses every other team consistently will win 50% of its games? Or is that wrong?

Also, has there been any work on parsing down that last 20%? I'm sorry to abuse you like this, but I'm not in the right circles.

To be honest I haven't even read the original article, just Desjardins quoting it.

I believe it was "what makes a win" although I'm not sure of what method Awad used to come to this conclusion, so take with grain of salt... It is however well established the fact that out chancing/possession helps drive wins more so than anything else that has been formulated thus far.

And no, no one has succeeded with the last 20%. There is too much variance/luck factors in shooting talent, defensive talent, goaltending and special teams that it becomes very muddied. Maybe one day.

 

[garret9]

That's an interesting Eberle quote. I haven't heard players talk about Corsi much.

I'm still trying to get my head around exactly what Corsi should be telling me, so these conversations are helpful. I feel like I need a handle on that before I can talk about the actual numbers.

Way I look at it:
Corsi% is ratio of team's shot attempts and closely mirrors effective puck possession when you are on the ice. When you get a larger sample size though score-close situations does better. It does okay for estimating scoring chances but Fenwick tends to improve this.

 

[Sweech]

83+ reads (and according to clicks most went to the site) and no one is saying a thing?

Any shocks? Any surprises? Any thing as expected?

Anything you learnt?

84 and I ain't sayin nothin'

 

[garret9]

84 and I ain't sayin nothin'

Haha double negative! so you are saying something

Also, I think this thread has a few more than that now haha.

 

[Hobby Bull]

To be honest I haven't even read the original article, just Desjardins quoting it.

I believe it was "what makes a win" although I'm not sure of what method Awad used to come to this conclusion, so take with grain of salt... It is however well established the fact that out chancing/possession helps drive wins more so than anything else that has been formulated thus far.

And no, no one has succeeded with the last 20%. There is too much variance/luck factors in shooting talent, defensive talent, goaltending and special teams that it becomes very muddied. Maybe one day.

There was a short interview clip somewhere here on HF. I imagine it was on the Oilers' board.

 

[Hobby Bull]

Just lil' but non-important FYI and history lesson, Corsi is a proper name not an acronym. Named after Jim Corsi who tracked it to weigh a goaltenders workload, since goalies react to all shot attempts, not just the ones that they save.

There is things that Corsi does not cover. You are right there and in being that you do not evaluate players solely on COrsi. However, the one you have reasoned out is false.

One: Wheeler actually tends to take a lot of shot attempts. He's been one of the highest Jets 2.0 after Kane, usually competing against Ladd and Frolik.
Two: There isn't really a correlation to guys who take lots of shot attempts and Corsi. Corsi% tends to correlate strongest to puck possession and a guy who does nothing but throw away the puck will not improve his numbers. As Eberle said: if you do the right things, like improve your breakouts and zone entries, your Corsi will improve, but if you work on trying to improve your Corsi, you will likely fail.

I never use Corsi on it's own; however, out chancing and out puck possessing is approximately 50% of what gets you wins, which means Corsi is about equally important as everything else combined... but NOT the whole picture.

As I said in the article:

Thanks. I was always trying to figure out what it was an acronym for. I swear I was spending time every day trying to weave words together that would make sense.

Now, what to do with that extra time...