Kukla's Korner

The Malik Report

More From Mike Babcock On Analytics

from Dan Rosen of NHL.com,

A basic element of the analytics movement happening in the NHL is possession-based stats featuring shot attempts-for versus shot attempts-against in 5-on-5 situations. Understanding the positive effects of possession and shooting the puck is not new to the hockey community, but stats found through possession metrics such as Corsi and Fenwick are relatively new and still are being examined for their usefulness and effectiveness.

Babcock said a key to using stats-based analytics, whether it's Corsi, Fenwick or another form, is making sure the numbers are accurate.

"The best analytic groups are getting the best information, so you have to make sure the stats you're using are correct," Babcock said. "When a guy tells you how good this player is stat-wise and you and your general manager don't agree with him, well, we better go back and assess it. Just because the analytics like the player and we don't, well, then we don't. We'll make the decision. But if you're wrong, go back and take a look at it so you know why."

more

Filed in: | The Malik Report | Permalink
 

Comments

RWBill's avatar

Shots for and against 5x5 is inherently and by its very nature biased against certain types of players.

A grind line or 4th line player doesn’t see a lot of minutes in the 1st and 2nd periods, however if his team enters the third period up 2-1 or 3-1 he will possibly play half or more of his minutes in that period.  This is a time when they are content to lay back and swarm defensively, forcing a low percentage shot, picking up the rebound, transitioning to center ice or across the other blue line long enough to hold on to the puck to kill time and ultimately slapping it behind the other net so they can reset their defensive posture, or change lines.

This player will have tons of <typically low percentage> shots against him with few shots generated by him and his line, by design.  Some stat nerds will fixate on some numbers and cry out loud about what a drag or how ineffective certain players are.  Meanwhile, the teammates and coach, i.e., people who actually watch and understand the game, know what contributions players bring, or not.

There are numerous examples in all sports of fans over emphasizing stats.  After all, they are not in the dugout or on the bench, have a limited view of the ice or the whole field of play, and the thing they have left is to gaze at an ever increasing mountain of stats, some of which are of dubious value.

Babcock implied this, coaches and GMs have an opinion of a player’s value, if some stat doesn’t seem to jibe with observed performance the stat is probably wrong.  I’m glad to read that, an over reliance on numbers over personal observation and evaluation is not the way for a professional team to run.

Posted by RWBill from cruising Brush Street with creepy Rob Lowe. on 07/24/14 at 02:15 PM ET

RWBill's avatar

I should say that I myself have posted numerous comments referencing different statistics. But I am also aware of their limitations when I do post those.

Posted by RWBill from cruising Brush Street with creepy Rob Lowe. on 07/24/14 at 02:42 PM ET

Avatar

I believe that the “Moneyball” philosophy might work in Baseball and possibly football.

Of the major 5 “team” sports in the US (Baseball, Basketball, Football, Soccer, Hockey) I’d imagine that team cohesion matters the most in hockey.  Advanced stats mean little when they are team based (such as corsi or fenwick) and not player based (moneyball). 


Additionally, not all SOG are equal.  Samuellson missing the net 15 times in a game shouldn’t increase his value.  Shots from center ice mean little to nothing unless you are Lidstrom or get the goalie to hold the puck.

Posted by neffernin on 07/24/14 at 05:52 PM ET

RWBill's avatar

Samuellson missing the net 15 times in a game shouldn’t increase his value.

William Tell !

Posted by RWBill from cruising Brush Street with creepy Rob Lowe. on 07/24/14 at 06:40 PM ET

Avatar

One can only fathom what the stats must have told Babcock and Holland on Cleary,  Quincey,  Glendening,  Andersson,  Kindl…  Stats don’t do you much if half the players you have signed are over the hill or mediocre.  Those are fundamental core stats that are irrefutable.

Posted by karlander on 07/24/14 at 06:48 PM ET

Nathan's avatar

A grind line or 4th line player doesn’t see a lot of minutes in the 1st and 2nd periods, however if his team enters the third period up 2-1 or 3-1 he will possibly play half or more of his minutes in that period.  This is a time when they are content to lay back and swarm defensively, forcing a low percentage shot, picking up the rebound, transitioning to center ice or across the other blue line long enough to hold on to the puck to kill time and ultimately slapping it behind the other net so they can reset their defensive posture, or change lines.

...

Babcock implied this, coaches and GMs have an opinion of a player’s value, if some stat doesn’t seem to jibe with observed performance the stat is probably wrong.  I’m glad to read that, an over reliance on numbers over personal observation and evaluation is not the way for a professional team to run.

Posted by RWBill from the open bar on The Hasek. on 07/24/14 at 02:15 PM ET

Generally, I think you have the causation backwards. The grinders play less and play in the situations you describe because they aren’t as good and capable at possessing the puck and putting shots on the opponent’s net. If they were, they’d be top six forwards.

Don’t mistake me as saying those players don’t have value and that a scenario like you describe isn’t an effective way for a GM and coach to build and run their team. It is very sensible. But as much as we all love the Helms and Drapers of the world, they play(ed) the roles they play(ed) as much because of their limitations as their skill sets. Pavel Datsyuk and Henrik Zetterberg have fewer limitations and broader skill sets with higher ceilings, so they play more hockey.

I definitely agree with your second part, and while I don’t think you’re doing this in your comment, it touches on what the Steve Simmonses of the world love to do. They craft pathetic strawman arguments that paint proponents of these statistical techniques as being completely blind/oblivious to watching the games with their two eyes and making judgments based on the old fashioned eyeball test. And certainly and similarly, there are fringe stat geeks that do have blinders on and craft arguments about “old school” guys that paint them out to be strawmen they aren’t (e.g. incapable of understanding the numbers, disingenuous, etc.).

The way I look at it is that when Babcock says he loves the data, and also loves the eyeball test, and will be critical/skeptical when the data don’t match what he sees… that is either a really really great sign, or a really bad sign. It’s a bad sign if it leads to you putting too much stock in Dan Cleary because of those “intangibles” or putting too much stock in Samuelsson because the numbers show he directs a lot of shots toward net for his role. It’s a great sign if it leads you to reconsidering both how you’re judging what you see and how your data is being analyzed.

In my opinion, the way of the future is that the Hollands and Babcocks of the world need to employ a couple geeks on their staffs, and they need to give the geeks a certain level of autonomy. For the part of the GM and coach, when they see a player they want to bring in or think should have his role changed, they should ask the geeks what the numbers say. For the part of the geeks, when the GM and coach are insistent that a player the geeks are rating highly isn’t what he seems, they need to take the judgment to heart and try to determine if that is due to a flaw in their model.

This attitude among baseball SABR-types is what has made that world so productive and influential the last decade. Sure, you can still find some guys in the BBWAA that are stuck in the stone age, and you can still find some SABR dweebs that only exist in the dark corners of the internet to troll people and haven’t seen an actual baseball game in 12 years. But it was the majority of SABR-types in the middle that have won out, and they won out because rather than solely fighting against the old school guys, they work(ed) to take the intuition and knowledge of old school scouts and managers, and use them to question their own statistical models, which led to tremendous improvements in those models.

Posted by Nathan from the scoresheet! on 07/25/14 at 09:48 AM ET

thethirdcoast's avatar

Hey, what’s Cleary’s WAR number?

I mean, it must be pretty high for him to snag the contract he just got, right?

Posted by thethirdcoast from Algiers, DZ on 07/25/14 at 12:26 PM ET

Add a Comment

Please limit embedded image or media size to 575 pixels wide.

Add your own avatar by joining Kukla's Korner, or logging in and uploading one in your member control panel.

Captchas bug you? Join KK or log in and you won't have to bother.

Smileys

Notify me of follow-up comments?

Feed

Most Recent Blog Posts

About The Malik Report

The Malik Report is a destination for all things Red Wings-related. I offer biased, perhaps unprofessional-at-times and verbose coverage of my favorite team, their prospects and developmental affiliates. I've joined the Kukla's Korner family with five years of blogging under my belt, and I hope you'll find almost everything you need to follow your Red Wings at a place where all opinions are created equal and we're all friends, talking about hockey and the team we love to follow.