2017-18 stats and underlying metrics thread [Mod: updated season]

[Whileee]

I am not sure why its difficult to believe that a lot of NHL teams are not competent when it comes to statistical analysis. Or many of them are liars.

... the Hawks were the referenced team earlier. If they are incompetent at advanced statistical analysis then how did they become possession monsters and SC champions? Or does it mean that you don't need to be competent at statistical analysis to be successful? I would buy the dissembling theory over the suggestion that they are incompetent. We should be a bit cautious in thinking that all of the talented analysts are outside the NHL.

 

[garret9]

... the Hawks were the referenced team earlier. If they are incompetent at advanced statistical analysis then how did they become possession monsters and SC champions? Or does it mean that you don't need to be competent at statistical analysis to be successful? I would buy the dissembling theory over the suggestion that they are incompetent. We should be a bit cautious in thinking that all of the talented analysts are outside the NHL.

Well...

A person can be a solid writer without spellcheck, thesaurus, editor, or dictionary.

Those are tools that help improve your writing. They help you make a good choice more often and a bad choice less often.


Now in regards to statistical evaluation, this is a world where primarily it hasn't been used extensively. If no one is using a comparative advantage, you don't *need* the advantage to be better. Then add in that there are other sources of comparative advantage...

 

[Whileee]

Possible, but suspicious.

From what I've heard from a guy I trust a LOT, there are maybe 2-4 teams that currently surpass what we can do. If I was to bet on teams, Chicago would be one I'd wager on... but still...

Ex: They signed an aging Khabibulin once and overpaid on Crawford due to what could essentially be called "clutch" sv%, which was then shown by Tyler Dellow to be guided by poor logic.
There are others but I'll leave it to that now.

Scoring chances though are subjective and not much more common than goals.

I'd also not be surprised if Chicago actually didn't focus on scoring chances but say they do.

Sure, I could buy the misdirection theory, but I think there is also a chance that some NHL teams have figured out more sophisticated ways of measuring shot attempts, shots and chances. I think that teams probably use it to fine-tune their systems and how their players play. This is different than just trying to decide whether a player is a good possession player or not. I would hazard a guess that player possession stats cluster substantially within teams, suggesting that it isn't just the player. I'm not sure that Bryan Bickell becomes a Corsi monster on the Avs or Leafs. Is Kris Versteeg a strong possession player? What about Johnny Oduya? From my perspective, we sometimes put too much emphasis on the individual players' stats, and ignore the effects of systems and how players fit into systems.

 

[Whileee]

Well...

A person can be a solid writer without spellcheck, thesaurus, editor, or dictionary.

Those are tools that help improve your writing. They help you make a good choice more often and a bad choice less often.


Now in regards to statistical evaluation, this is a world where primarily it hasn't been used extensively. If no one is using a comparative advantage, you don't *need* the advantage to be better. Then add in that there are other sources of comparative advantage...

I agree with this. I'm less convinced that because we don't know what the Hawks are doing that they couldn't have an analytical plan involving scoring chances that is effective in improving their team's performance.

 

[garret9]

Sure, I could buy the misdirection theory, but I think there is also a chance that some NHL teams have figured out more sophisticated ways of measuring shot attempts, shots and chances. I think that teams probably use it to fine-tune their systems and how their players play. This is different than just trying to decide whether a player is a good possession player or not. I would hazard a guess that player possession stats cluster substantially within teams, suggesting that it isn't just the player. I'm not sure that Bryan Bickell becomes a Corsi monster on the Avs or Leafs. Is Kris Versteeg a strong possession player? What about Johnny Oduya? From my perspective, we sometimes put too much emphasis on the individual players' stats, and ignore the effects of systems and how players fit into systems.

The effects aren't ignored though...

For example, you are listing examples of team effects and systems effects.

Contextual nuances have been the main source of study in shot metrics for the past 4 or so years.

I don't believe you can correct for team or system factors by adding in scoring chances.
The reasons why team factors exist is a combination of QoT and OZS. That will persist for both metrics.
Then with systems, we see that teams and coaches regress heavily to NHL average in scoring chances per shot volume both offensively and defensively. Systems affect shot volume.

If you want to get more look into coaching effects, Ben Wendorf has been running a series on Hockey-Graphs.com, using 2pS% that allows him to go even further historically than the BTN era.

 

[Whileee]

The effects aren't ignored though...

For example, you are listing examples of team effects and systems effects.

Contextual nuances have been the main source of study in shot metrics for the past 4 or so years.

I don't believe you can correct for team or system factors by adding in scoring chances.
The reasons why team factors exist is a combination of QoT and OZS. That will persist for both metrics.
Then with systems, we see that teams and coaches regress heavily to NHL average in scoring chances per shot volume both offensively and defensively. Systems affect shot volume.

If you want to get more look into coaching effects, Ben Wendorf has been running a series on Hockey-Graphs.com, using 2pS% that allows him to go even further historically than the BTN era.

I expect that there are various ways of classifying "scoring chances". After all, Corsi is basically a "scoring chance" in the broadest sense, and Fenwick is a slight refinement. I think it's conceivable that some further refinement can be done.

In any case, I think it is a very positive development that teams are starting to utilize advanced stats. I think it will be very interesting to see how this will be incorporated into actual personnel and coaching decisions with NHL teams. It might be a lot of fun to see how advanced stats guys assess and critique the Oilers from the perspective of analytics now that they have Dellow on board. Or the Devils with the poker guru. Will he really be able to influence decisions? To what extent? In which areas?

Also, how long before we see an article where analytics guys are credited with improving teams that had nowhere to go but up?

 

[garret9]

I expect that there are various ways of classifying "scoring chances". After all, Corsi is basically a "scoring chance" in the broadest sense, and Fenwick is a slight refinement. I think it's conceivable that some further refinement can be done.

In any case, I think it is a very positive development that teams are starting to utilize advanced stats. I think it will be very interesting to see how this will be incorporated into actual personnel and coaching decisions with NHL teams. It might be a lot of fun to see how advanced stats guys assess and critique the Oilers from the perspective of analytics now that they have Dellow on board. Or the Devils with the poker guru. Will he really be able to influence decisions? To what extent? In which areas?

Also, how long before we see an article where analytics guys are credited with improving teams that had nowhere to go but up?

The refinement though doesn't seem to bring any added information unless you are going 3+ seasons or using team level samples. The issue is due to sampling. And, as I pointed out, the consensus in the NHL for nomenclature of a scoring chance has to do with a shot in the slot with either minor or no contesting.


To the article point... I did one even better than that:

http://morethan140charactershockey.blogspot.ca/2014/08/corsi-didnt-help-tyler-dellow-and-kyle.html

 

[Whileee]

The refinement though doesn't seem to bring any added information unless you are going 3+ seasons or using team level samples. The issue is due to sampling. And, as I pointed out, the consensus in the NHL for nomenclature of a scoring chance has to do with a shot in the slot with either minor or no contesting.


To the article point... I did one even better than that:

http://morethan140charactershockey.blogspot.ca/2014/08/corsi-didnt-help-tyler-dellow-and-kyle.html

re: article.

I expect that there are different ways of classifying scoring chances... the one above seems pretty narrow. My point is that some teams might use different ways to classify and use "scoring chances", and they might relate to how they organize their system. By the way, a shot from Wright in the slot should never be classified as a "scoring chance".

 

[garret9]

Well, maybe that's because how often he gets those?
He's only had ~23 of those in his entire career. (estimated shots from the slot)

Besides, that's not the point. How many team's have James Wright's taking a large percentage of their shots.

 

[Whileee]

Well, maybe that's because how often he gets those?
He's only had ~23 of those in his entire career. (estimated shots from the slot)

Besides, that's not the point. How many team's have James Wright's taking a large percentage of their shots.

....how many teams play a guy like Wright making sure that nobody on your team is getting very many shot attempts?

 

[garret9]

Lots of guys have scored like James Wright:
http://www.behindthenet.ca/nhl_stat...29+30+32+33+34+45+46+63+67+17+18+19+20#snip=f
About 560 to play at least a game.


But anyways, adding scoring chances only would refine if there is something added in the "quality" argument.
But, as has been shown, no matter how you cut "quality" there always seems to be no sustainability in quality per quantity.

So really, even if Blackhawks are smart and good, I'm incredibly skeptical that they could use scoring chances more than a red flag for video review on certain individuals to see what's going on.

 

[garret9]

... the Hawks were the referenced team earlier. If they are incompetent at advanced statistical analysis then how did they become possession monsters and SC champions? Or does it mean that you don't need to be competent at statistical analysis to be successful? I would buy the dissembling theory over the suggestion that they are incompetent. We should be a bit cautious in thinking that all of the talented analysts are outside the NHL.

All yes, most? Maybe not a few years ago.

Look at Nonis' comments now (6:09 PM CT) on Hope_Smoke's (@Hope_Smoke) timeline.

Basically Nonis says the stuff Charron and Dubas are bringing is miles ahead of what they were using previously.

He said people came with similar stuff but they ignored it because it was presented in a way that didn't make sense (and I quote) to "old school guys".

 

[Aavco Cup]

... the Hawks were the referenced team earlier. If they are incompetent at advanced statistical analysis then how did they become possession monsters and SC champions? Or does it mean that you don't need to be competent at statistical analysis to be successful?

The Hawks have better players that perform very well in shot metrics.

Do they have those players because they looked for something specificaly related to metrics when drafting, trading or signing FA's? Or is their metrics performance from just being lucky that they happened to acquire good players?

 

[garret9]

The Hawks have better players that perform very well in shot metrics.

Do they have those players because they looked for something specificaly related to metrics when drafting, trading or signing FA's? Or is their metrics performance from just being lucky that they happened to acquire good players?

Or just being good at acquiring and seeing good players.
Or mix of both.

 

[Aavco Cup]

Here's an interesting story about MacKinnon's home vs. road Corsi. (Setoguchi made his chart)

http://www.tsn.ca/nhl/story/?id=462502

 

[jetkarma*]

The Hawks have better players that perform very well in shot metrics.

Do they have those players because they looked for something specificaly related to metrics when drafting, trading or signing FA's? Or is their metrics performance from just being lucky that they happened to acquire good players?

Honestly I think the Hawks utilize advanced stats very well but I doubt it forms the core of their opinion on players they acquire.

Defining in a statistical way is obviously becoming more accepted and focused but the premise behind why a player or a team is successful , or more so than other teams or players isn't imo .

Going YEARS back , I was watching the summit series with one of my best friends who happened to be a really good hockey player , and he elaborated at a young age why dump and chase wasn't the best way to play and that most coaches then always wanted him as a C to do that.

He grew up to be a pretty good hockey player at the University level and become an Economics professor and has utilized both strengths to be involved in hockey. The philosophy isn't new , and while there certainly are reasons to embrace this wave , I think the core of what analytics "tells" us has been understood for a long time.

 

[Whileee]

Honestly I think the Hawks utilize advanced stats very well but I doubt it forms the core of their opinion on players they acquire.

Defining in a statistical way is obviously becoming more accepted and focused but the premise behind why a player or a team is successful , or more so than other teams or players isn't imo .

Going YEARS back , I was watching the summit series with one of my best friends who happened to be a really good hockey player , and he elaborated at a young age why dump and chase wasn't the best way to play and that most coaches then always wanted him as a C to do that.

He grew up to be a pretty good hockey player at the University level and become an Economics professor and has utilized both strengths to be involved in hockey. The philosophy isn't new , and while there certainly are reasons to embrace this wave , I think the core of what analytics "tells" us has been understood for a long time.

Is Versteeg a good "corsi" player? What about Oduya? Handzus? Pierre Marc Bouchard?

What about Frolik? Ladd? Byfuglien?

It doesn't seem that the Hawks make player decisions specifically to acquire or retain strong Corsi players.

They have a few fantastic players that carry others along. They probably also employ a very effective system and get players to work into that system.

When it comes to scoring, I expect that a lot of it is designed around limiting scoring chances in the defensive zone and then on a range of possession-related systems around zone exit, neutral zone play, zone entry, etc. I expect that some of the "analytics" used by teams are designed to look at the specific performance of certain players in certain aspects, and others are more system related.

 

[garret9]

Is Versteeg a good "corsi" player? What about Oduya? Handzus? Pierre Marc Bouchard?

What about Frolik? Ladd? Byfuglien?

It doesn't seem that the Hawks make player decisions specifically to acquire or retain strong Corsi players.

They have a few fantastic players that carry others along. They probably also employ a very effective system and get players to work into that system.

When it comes to scoring, I expect that a lot of it is designed around limiting scoring chances in the defensive zone and then on a range of possession-related systems around zone exit, neutral zone play, zone entry, etc. I expect that some of the "analytics" used by teams are designed to look at the specific performance of certain players in certain aspects, and others are more system related.

Versteeg: he was pre-lockout shortened year, but has struggled since, but not terribly so
Oduya: he's decent, but not exceptional
Handzus: no, not at all... terrible
PMB: can dominate soft minutes, yes; toughs, not at all
Frolik: yes
Ladd: hell, yes
Byfuglien: as d-man, yes; as forward, no

However, I don't think this would really support that their movements are for or against possession players.
There is the cap involved and also only so many players are wanted and not wanted on the market. Frolik, Ladd, Buff may all have been kept if not for the cap. Same with Campbell.
Plus, there is still human error in assessing talent, as we've seen Chicago screw up with decisions too.

Also, good/bad Corsi (or anything really) is completely relative. Good or bad is simply a standard compared to something else, so it depends on what you are comparing to.

 

[Whileee]

Versteeg: he was pre-lockout shortened year, but has struggled since, but not terribly so
Oduya: he's decent, but not exceptional
Handzus: no, not at all... terrible
PMB: can dominate soft minutes, yes; toughs, not at all
Frolik: yes
Ladd: hell, yes
Byfuglien: as d-man, yes; as forward, no

However, I don't think this would really support that their movements are for or against possession players.
There is the cap involved and also only so many players are wanted and not wanted on the market. Frolik, Ladd, Buff may all have been kept if not for the cap. Same with Campbell.
Plus, there is still human error in assessing talent, as we've seen Chicago screw up with decisions too.

Also, good/bad Corsi (or anything really) is completely relative. Good or bad is simply a standard compared to something else, so it depends on what you are comparing to.

Sure, I understand the cap problems. My point is that they don't seem to be focusing acquisitions and cap space on corsi monsters, but most players get a bump in their corsi when playing with the Hawks in their system. I think a lot of it has to do with: a) the quality of the top end of a team's roster, which allows lesser players to face easier opposition in better situations and, b) the use of an effective puck possession system that gets refined over time. I guess that sounds simple - great players and good system, but it does seem to put a premium on getting the top-end players. In that case, what role does "money puck" play when trying to build a top echelon team in comparison to getting a few top end players? It's a complex question, I suppose.

 

[garret9]

I I'm not sure how much of a point that is though...

The things you cite are merely the contextual nuances that are fairly well known: OZS, DZS, QoT, QoC, TOI, TmCorsi%off, OpCorsi%off, etc.

We are pretty good getting at understanding an established player's true Corsi value and also who drives more than others.

In the role of "money puck" comment, the answer is simple. More good decisions and less bad ones. Chicago is far from optimal or perfect in their decision making and I truly that even some of the simplest underlying metrics would have improved their decision making process over the past while.

It's in the margins for hockey. As Gabe Desjardins said, teams have been tracking data and reviewing video for a long while. Because of this, hockey teams are not quite at the same place baseball was when Sabremetrics arrived. They know a lot already that shot metrics say. I believe there is already well over 80% in agreement when looking at usage optimizing and TOI usage.

Hockey is a game littered with chance. This is why I believe hockey analytics should be about always increasing that chance to win.

To the original point, I don't view Chicago (or shot metrics) as perfect, but I think that even the simple metrics and analysis available through the hockey blogosphere could have bettered Chicago.

 

[garret9]

Corey Sznajder* just got hired by an "unknown team"

*better known as the guy who tracked all the neutral zone data I use

 

[garret9]

Also, Dimitri Filipovic has been hired.

For those that do not know, I bartend at a bar that usually has the following guys hang out together:
Myself
Rhys Jessop (That's Offside! guy)
Dimitri Filipovic (Unknown NHL team; well unknown to most )
Cam Charron (Toronto Maple Leafs)
Harrison Mooney

 

[Aavco Cup]

Lack of NHL stats in an information age is mind-boggling

http://nypost.com/2014/09/27/lack-of-nhl-stats-in-an-information-age-is-mind-boggling/

 

[garret9]

Lack of NHL stats in an information age is mind-boggling

http://nypost.com/2014/09/27/lack-of-nhl-stats-in-an-information-age-is-mind-boggling/

I agree with the title.
I hate that article bulk of the article...
There's a lot of issues with it.

Some of the stats listed "missing" are not actually missing.
Some of the stats listed "missing" are not actually needed as the sample size would be so small it would be meaningless.

 

[garret9]

Yesterday Gabriel Desjardins -founder of Behind the Net & Arctic Ice Hockey, and former consultant for multiple NHL teams- commented on that NY Post article. I talked to him after about it.

Essentially his point goes like this:

Hockey stats "revolution" in the NHL is different than Baseball... and it has nothing to do with the style of the game.

Baseball had players that were vastly under appreciated and other vastly over appreciated. Analytics was able to turn the game upside down.

In hockey, events have been tracked by teams for decades. Teams have been analyzing events and video for years. There isn't any low hanging fruit 4th line player who is unappreciated and actually a top line player.

With hockey it's not the debate of analysis vs no analysis, as analytics have been going on in hockey for a while. The issue is good analysis vs far better analysis. For the most part team's are extraordinarily internal, and not just in terms of hiding their data. Teams essentially track and test metrics only with themselves and on themselves, and the testing part is very minimal.

This relates to the discussion earlier on this thread between Avco, Whilee, and I. Teams already make good decisions, like Chicago. This doesn't mean that the current movement online that has been vigorously testing things league wide couldn't improve those decisions. They also make bad ones, like the example I gave about situational save percentage they use actually performing worse than regular save percentage.

He adds that you won't be seeing the teams that hired guys turning over and becoming successes instantaneously for that reason.