I’ve been beating the drum for keeping Jamie Benn, Alexander Radulov, and Tyler Seguin apart for a while now to follow in the Pittsburgh Penguins mold. Spread your stars out and complement them with useful players to maximize the offense. It’s an idea that has been out there for a little while now, and I remembered hearing a lot about roster construction optimization when the Passing Project information really started getting out there in force.
Corey Sznajder (@ShutdownLine) has tracked a ton of data: zone entries, zone exists, defenseman touches, and a lot of other important microstats. Ryan Stimson (@RK_Stimp) led the Passing Project to get a better idea of how passing impacts scoring. Both sets of data were put together to give a more complete picture of players across the league.
One of the resulting conclusions drawn from the data was the existence of four “Player Types” at forward. Stimson put this excellent piece together at hockey-graphs.com detailing what he found.
Teams are incentivized to spread their best players throughout their lineup. This is due to a fact that a team can only have so much success with a stacked line (looking at you, Boston). Furthermore, with hockey being a strong link game, ensuring that the best players are on the ice as much as possible at different times gives you an advantage, or at least doesn’t put you at more of a disadvantage. The final piece is identifying which players can complement those elite forwards the best.
He found that the four player types are Playmakers, Shooters, Balanced, and Dependent. Dependent is a nice way to say plugs. The other three showed to be useful, with playmakers being the ones who drive scoring the most which makes sense.
So as we watch the Dallas Stars struggle through a tough stretch I thought “hey, I wonder how optimized lines for the Stars based off of the most up-to-date data available would look.” Let’s find out.
For future reference, this is what the Stars rolled out last night against the Tampa Bay Lightning:
(All data is from 2015-16 and 2016-17 which poses a problem: players age or get worse for whatever other reason. I’ll note where I think this could show itself as a problem.)
First lets see what roles are currently on the Stars roster.
This group is the guys that move the needle the most. The names mostly aren’t going to surprise you. All the visualizations are from this awesome Tableau.
I mean, yeah. Obviously.
I said they weren’t going to surprise you.
This dude was such a bargain.
Spezza has had a long career of being a quality player. We’ve seen him struggle to fit into the system under Ken Hitchcock at times, but he’s still Spezza. He still drives the play, but he’s been hampered by an aggressively low PDO driven by a really low on-ice save percentage.
Is it more likely that his numbers have gone down because he’s terrible, or because there has been an inordinate amount of puck luck going against him? I’m willing to let him stay in the playmaker category until data exists to prove it wrong.
I shit you not.
He may not be in the Benn, Seguin, and Spezza category, but over the sample period Hanzal has been really good. You have to ask yourself if he has regressed this year, or if the role he is being asked to fill has killed his numbers off. Like Spezza he is the victim of an aggressively low PDO, but his is driven by both low on-ice shooting and save percentages. He legitimately is a bottom six anchor, but, like, a good anchor.
One thing Ritchie does do is pump shots on net. I’m not going to try to argue that he’s elite or anything, but he isn’t as useless as many would like to think.
The Balanced Guys
These guys don’t fit into the top two categories, but they make useful contributions.
I think I would be willing to argue that Faksa is more of a playmaker this year, but without data to support it I want to keep this as fair and realistic as possible. You could make a case for he and Hanzal flipping types here based on a 2018 season eye test and I wouldn’t try to argue against it.
After missing a year Janmark is getting jobbed by PDO like Hanzal and Spezza, but he has been visibly impressive all season.
I’m not going to lie to you, this one surprised me. Roussel is just a good all around depth player. You can see him making solid plays all over the ice, but he doesn’t jump out as this decent. Math is fun.
Shore has been nothing special, but he isn’t a plug either.
There is currently no data for Tyler Pitlick, Gemel Smith, Jason Dickinson, or Remi Elie. I want to be as fair as possible and assume there are some plugs somewhere on the roster up front, but none of these guys strike me as drains. I don’t think it’s out of the question to suggest that every forward in the Stars lineup contributes something useful.
So, how best do the Stars increase their odds of scoring?
Stimson included this expected goal chart in his story. Let’s use it.
The Stars have the ability to use a playmaker on every line, and a second one on one line. Down the middle it makes a lot of sense to go with Seguin, Benn, Spezza, and Hanzal.
If you split Seguin and Benn apart you have to figure out where Radulov goes. I have to wonder if some of the consternation people feel about Benn this year has to do with Radulov. When they’re on the ice together Radulov’s production lags. Seguin and Radulov hum along just fine.
Seguin-Radulov-BalancedF for your first line is the second best combination available at 57.2 xG%.
A playmaker, a shooter, and a balanced forward work well together too. Jamie Benn, Brett Ritchie, and a balanced forward of your choice would slot in at 55 xG%.
The final two lines would each feature a playmaker and at least two balanced guys, with the chance of Janmark or Faksa being in a higher “type cluster”.
Your third line would still be all balanced players which is still effective. A playmaker with two balanced lines goes for 52.2 xG%.
Those Stars lines from the Tampa game? They’re ok, but with all three of the alpha scorers lumped together the Stars minimize their offense to a degree.
Should the Stars ever use Benn, Radulov, and Seguin together?
Yes. Late in games when the bottom of the lineup is generally used less frequently the Stars would increase their expected goal output by throwing those three together. A line of three playmakers could expect an xG% of 58.9%. This is the most deadly line possible, and late in the game you aren’t hurt as much by throwing your top guys together.
Late in the game they could also throw Jason Spezza, Martin Hanzal, and Brett Ritchie together to optimize a second unit. That line would, based just on player roles, look at a 55.1 xG%..
How do you most appropriately use Hanzal?
I think there is a good argument that the Stars already are doing that, but the results aren’t there due to some bad luck. What he is doing isn’t showing up on the standard or advanced stat sheet without more data available, but I’ve seen enough to think he’s more useful than he appears at times.
Now he just needs to stay healthy.
Are the Stars getting the most out of their powerplay?
When Stimson put his story together he pointed out that there were about four playmakers and shooters per NHL club. The Stars have six, and two of them park themselves in front of the net on the powerplay.
This may make sense for Ritchie conceptually since he’s a shooter, but it doesn’t for Hanzal. It minimizes his ability to make plays by reducing him to a glorified crash test dummy hoping to get drilled with a puck. Maybe try to get him more space to make some plays on the second unit, or make him net front on the first unit.
For the record, this is Max Pacioretty.
For those of you who want the Stars to have a shooter, but don’t want Brett Ritchie near the top six, there’s your guy. If only he had been available and attainable at the trade deadline.
Hey, click these:
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
2/20/18 – Heiskanen Scored a Goal