FanPost

Why won't hockey execs pay attention to analytics?

How the Oilers Analytics Blogs are missing half the game

As someone who finds advanced stats interesting, I’m ashamed to say that I only recently saw "Moneyball", which provided some insight into how they started to take over baseball. In many ways, the situation there mirrors the current situation in hockey, with the stat-heads coming in with new ideas and methods of evaluation, while old school "hockey guys" resist the influence.

I’ll start by saying that a Moneyball approach is underutilized by hockey GMs; it’s a great way to build a team to ensure maximum value for contracts, and ignoring advanced stats is likely to result in a GM making poor decisions about who to sign to the team, for how much, and for how long. However, part of the reason this exists is that there is no push from coaching staffs to incorporate advanced stats, and there’s a very good reason for that.

You’ll note in Moneyball, there are plenty of scenes of Brad Pitt talking with Jonah Hill about Player X or player Y, and how good they are, and I can imagine many such conversations being possible between a hockey GM and a strong analytics person. However, there are also scenes of Jonah Hill talking to players; explaining how to adjust their play to increase their chances of success (taking more pitches, ignoring stealing, getting rid of sacrificing, not swinging at the first pitch, etc). That’s where hockey is missing out.

In hockey advanced stats, there is currently little reason for players or coaches to buy in.

This is best explained by taking a simple example. Every day, I pick my kid up at school, and what do you know, Nick Schultz’s kid goes to the same school. Maybe we exchange a few hellos now and then, and one day, because I want the Oilers to do well, I decide "I’m going to help Nick Schultz and the Oilers, and explain Corsi, fenwick, shooting %, PDO, and usage to him."

So far, in 2013-2014, at the time I reviewed this, Nick had played 17 games. His Corsi was 44.3%, poor even for the Oilers (his Corsi relative is -2.5%). He can’t claim to be unlucky, as his PDO was 100.7, and team save % when he’s on the ice is at 0.936. By most advanced stats methods, the conclusion is pretty simple: Nick Schultz is not capable of playing top 4 minutes, and is borderline at best to play top 6.

So, after a few days, he understands it all. Every time he gets the puck on his stick, it’s either directed at the opposition net if in the OZ, or down the ice if in the DZ (to prevent shots against); his on ice stats start improving. On offense, because he’s rushing so many shots from questionable locations, his shooting percentage drops even further below league average for his position. On defense, because he’s icing the puck so much, he’s getting more than his fair share of defensive zone faceoffs (and likely facing tougher opposition as they line change top lines in). As a result, he is maybe holding his own Corsi-wise, but can legitimately say he’s facing top opposition with difficult zone starts, and maybe now he looks like a legitimate top 4 defenseman based on his on ice stats.

But is this player helping the team win? No. Because his actions, while improving the metric, are not actually improving possession.

Advanced stats will not break through in hockey until such time as they can result in a coach and/or player saying "in order for us to be successful, I must do X". Let me give an example of the kind of thing that would be necessary.

Once again, I’m chatting up Nick on the playground, and I offer to help him as best as an out of shape, old, never played serious hockey guy can do. I work from a hypothesis: many of Nick’s problems stem from an inability to efficiently exit the defensive zone after gaining control. Some good posts have been made on this subject already:

http://blogs.edmontonjournal.com/2011/10/10/magnus-paajarvi-packhorse/

or

http://www.coppernblue.com/2011/11/11/2554850/jeff-petry-theo-peckham-zone-exits

but the question again remains…what do I tell ole Nick?

What’s necessary to do is to break things down even further.

Each time a player gets clear possession of the puck in his defensive zone, there is one of three game situations he can be in:

1) no pressure (no opposing players inside the blue line),

2) forecheck pressure (1-2 opposing players inside the blueline), or

3) zone occupied (3+ opposing players inside the blueline).

At that moment, the defensive player has generally three choices:

1) Clear the puck out of the zone (uncontrolled clear),

2) Pass the puck (attempt to pass to a teammate, inside or outside the zone), or

3) Skate the puck (attempt to cross the blueline with the puck).

These choices can then lead to one of several results within the next 5-10 seconds:

1) Zone exit with possession (team still has possession after the puck crosses the blue line)

2) Zone exit without possession (puck exits zone, but is now in control of other team)

3) Turnover (puck does not exit zone)

What you might do then is to create a simple evaluation tool; say zone exit with possession is worth 8 points, zone exit without possession is worth 5 points, turnover is worth 1 point (these numbers are just made up for the example; the correct ratio between these numbers may very well be completely different). You then evaluate each player and see what they do well. From there, you have something both a coach and player can take away. Let’s take an imaginary chart for Nick that I just made up:

20 imaginary defensive zone touches:

Situation

Action

Result

Score

2

3

1

8

1

2

2

5

2

3

1

8

3

3

2

5

2

2

1

8

2

2

2

5

3

3

3

1

1

2

3

1

2

2

1

8

3

1

1

8

2

2

3

1

1

2

2

5

2

1

2

5

3

1

1

8

2

3

3

1

3

2

1

8

3

1

2

5

2

2

2

5

1

1

3

1

1

2

2

5

From there, you then extract a performance matrix:

Action

Situation

1 (clear)

2 (pass)

3 (skate)

1 (no forecheck)

13%/1

50%/4

n/a

2 (mild forecheck)

63%/1

68%/5

71%/3

3 (occupied zone)

88%/3

100%/1

38%/2

And now, I can say something to Nick. When you are facing a no forecheck pressure situation (situation 1), you are much better off attempting a pass out (action 2) than an uncontrolled clear (action 1). When you are facing an occupied zone situation (situation 3), and you attempt to skate the puck (action 3), you are seeing much worse results than if you immediately pass (action 2) or uncontrolled clear the zone (action 1). From there, our player can actually take advanced stats and make adjustments to his play to improve the team results. Note as well – this must be personalized. No one on earth should claim that Nick’s best decision in these spots is the same as, say Jeff Petry’s, or Justin Schultz’s.

The other benefit to going "bottom up" (trying to sell the players on advanced stats instead of GMs) is that a player has the ability to change his play without restriction; when in an occupied zone situation, Nick can start chipping the puck out instead of passing immediately. A GM seeing "oh lord, player X is a disaster" is highly restricted from making a change due to cap constraints, roster spots, contract lengths, and so on; alternatively, a GM might say "player Y is a perfect fit for what we need", but have no reasonable method by which to acquire him.

So why are the Oilers Analytics Blogs hurting the team progress? Because the focus on the wrong analytics (player evaluation rather than player improvement) overlooks what is of primary use to the team, and because the team doesn’t see use in the information, it either ignores or discounts the lessons that can be learned. It does absolutely nothing right now to compile and report on daily Corsi or Fenwick stats, other than to say "player X sucks/is great"; basically, a report card with no suggestions that isn’t going to be viewed favorably by anyone.

Analytics has done a great job in proving that possession correlates strongly with victories, to the point that denying that is foolish. However, they have not really yet taken the next step of going deeper into how to improve possession.

Analytics should answer not just what, but how. I would note that some work on this is already happening; information such as average shot distance, zone exits/entrances, and such, but it needs to go further. Until it does so, it will never achieve the mainstream standing it should have. Focus on the "why" and players and coaches will listen.

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

Join The Copper & Blue

You must be a member of The Copper & Blue to participate.

We have our own Community Guidelines at The Copper & Blue. You should read them.

Join The Copper & Blue

You must be a member of The Copper & Blue to participate.

We have our own Community Guidelines at The Copper & Blue. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9355_tracker