3/4/2018 – Optimizing the Defense of the Dallas Stars Using Player Type Data

Earlier this week I threw a post together trying to optimize the Dallas Stars lineup using Player Type data from tracking and analysis led by Ryan Stimson and Corey Sznajder and based off of this article.

You may have noticed that I didn’t touch the defensemen. If you have enough material to make two movies and double your profits, why wouldn’t you? I know I sat through both Kill Bill movies and I eagerly ate up Back to the Future III as a kid.

So why not dive into them now?

I used the lines from the Stars overtime loss to Tampa Bay for the previous post. We’ll use them again for this post.

The Types

def_xg_share

These four types aren’t all that surprising. Stimson found that only 11 fit the All Around type. We should have known defense-oriented guys don’t move the needle much, but there are still many holdouts praying at the Temple of the Derian Hatcher, God of Defensive Defensemen. that this isn’t the case.

It is. Let the healing begin.

Or don’t.

Whatever.

All Around

John Klingberg

Klingberg

I feel like this isn’t going to surprise anyone enterprising enough to stumble across this site because it should be glaringly obvious, but I don’t want to turn away potential new converts to The Church of Bae by being mean. Bienvenidos! Willkommen! Sign the guest book!

Volume Shooters

Stephen Johns

Johns

How many times have you seen Stephen Johns absolutely crush a puck this year and thought “shit, he crushed that puck?”

Again, this data is from BEFORE this season. I can only imagine how much more robust this looks now. Johns has been great all year. I keep thinking of the AHL game I watched shortly after the Stars picked him up from Chicago. He was an absolute beast and some of those gifs look a lot like the Johns we’re seeing more and more now.

Greg Pateryn

Hmm.

Pateryn

The Turner to Dan Hamhuis’ Hooch was surprisingly stout in the games tracked prior to this season when he worked his way into a more full time role.

Defense-Oriented

Dan Hamhuis

Hamhuis

Yeah, that looks about right. I wouldn’t be shocked if it were a little worse this year given that aging still happens, but as bad as last season was I don’t know. It could be a wash.

Esa Lindell

Lindell

I have some issues with the Lindell chart. It’s only 200 or so minutes worth of time on ice, and that old eye test of the 2018 season shows a significantly better player. Maybe he still is in the defensive category, but I can’t imagine he fills up so little of this chart now. He and Klingberg have been great together.

Marc Methot

Methot.jpg

Methot is pretty clearly the Stars 6th or 7th defenseman pending Julius Honka data. Honka has no data tracked.

I’m confident Honka isn’t falling into the Defense-Oriented group. If I had to guess I would guess he falls into the Puck Mover category, but with no data to use it is mostly a guessing game. He’s better overall than Methot though.

Optimized Defense Pairs

def_pairs

The optimized pairings present a bit of a problem for the Stars, but then again maybe not? The Stars have three guys that don’t fall into the Defensive-Oriented pile, and all three play on the right side which means each pair is inevitably going to have a Defensive-Oriented guy on it.

This limits the upside potential of the defense corps, but isn’t that the point anyway? Ken Hitchcock is about limiting mistakes and focusing on a sound defensive approach. Inevitably that is going to limit the top end potential for offensive output so, hell, maybe this is exactly what he wants.

A Klingberg/Lindell pairing gives an xG% of 51.3%, but optimally a Volume Shooter in the mode of a Johns would be with him. Then again, last year this pair had a 49% Corsi  % and 47.55% Scoring Chance % per Natural Stat Trick. This year? 53.47% and 54.83% respectively. Something is working. The team structure as a whole and Lindell specifically deserve a lot of credit.

The Methot/Johns pair and the Hamhuis/Pateryn pair both yield xG% of 49.9 so they break even.

This gets us to the wild card of Julius Honka. You don’t really want to take Pateryn out of the lineup and Johns certainly isn’t coming out. Do you play someone out of position and take Hamhuis or Methot out of the lineup, two defensive defensemen who do fill up their charts a little bit?

Honka is certainly a good player, but I don’t know how you reasonably get him into the lineup more often right now which makes not getting something to help the current roster (Pacioretty?) in a deal for him at the deadline all the more frustrating if you think the Stars can win the Stanley Cup. They can, Brad.

John Klingberg is a Legitimate Norris Candidate

This data isn’t even from the 2017-18 season when he has been otherworldly. The data collected identified eleven players that fit into the All Around category. One of those was Klingberg, and that data is from a time when a significant number of Stars fans continuously ripped the guy or wanted him booted out of town. Some of you still do and you should be run naked through Westeros with a Twitter chorus of Stars fans chanting shame at you.

Blueliners can be hard to gauge without more data available. At some point all of the available data trends to one outcome enough that either you choose to believe it or not. Accept that Klingberg is good and enjoy it. It’s ok to root for a good player who occasionally makes a mistake.

Defensive Defensemen

One of the bigger problems rising from the freely available access to data is a willingness of those interested in the data to draw conclusions that seem supported by the data, but in reality make little sense. Correlation is not causation. If data doesn’t support commonly held ideas the tendency seems to be to say “Eureka!” and jump out of the bathtub like a mad person, put on a bathrobe, and run into the tweets proclaiming the new discovery.

This information doesn’t prove that defensive defensemen are useless and the article doesn’t try to suggest they are. There are a lot of them though. I wonder how much these numbers drag from coaches hard matching guys they feel are their best defenders.

I also think these guys have some value as penalty killers. The skill set necessary to kill penalties doesn’t focus as much on offense, though those skills still have their place. Staying in lanes, reading play progressions, and puck movement are going to be the focus here.

Lindell figures in prominently here. A player like Dan Hamhuis is still going to be useful even though he falls in this category. Methot too, though not as much.

SCA

That’s just scoring chances against per 60 minutes using the raw data from Natural Stat Trick. Methot isn’t particularly good at it. They should really get Klingberg out there more. He should be a Norris candidate and he would be an even stronger one if he was given a chance to make an impact on the penalty kill.

How does that look league wide? Here are the top 25 in “scoring chance prevention” on the penalty kill. I arbitrarily cut off the list at the point where Methot slots in to get him involved.

SCAL

Yeah, that’s Lindell and Klingberg right outside the top ten. In the entire NHL. Defensively. Stephen Johns shows up at 24.

Patrik Nemeth

 Nemeth

The Stars could have saved themselves several million dollars and a second round pick had they not soured the relationship with Patrik Nemeth so badly that he requested a move. Nemeth isn’t much different profile-wise from Methot. He could always play, and he still can now. Miss you buddy.

Jamie Oleksiak

The wars are over, they said. It’s time to heal, they said. Let it go, they said.

Oleksiak

Never, he said in response as he tipped his cap and rode off into the sunset to return to the rewarding life he built for himself in the Post War Era.

Hey, click these:

3/2/18 – Optimizing the Dallas Stars Lineup Using Player Type Data
2/26/18 – Kick The Window Open
The Dallas Stars Need to Trade for Max Pacioretty, and Here’s Why
2/22/18 – David Freese, Brett Hull, and Arby’s
2/21/18 – Carmax, Priorities, and Ben Bishop

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s