The Washington Capitals seem to have turned a corner. They’re not “back to good” yet, but they’re at least trending in the right direction. Though I shared a bunch of thoughts after watching Tuesday’s game in person, there’s one nit I’d like to pick in more detail.

I think the Capitals defensive pairs are all wrong.

This is new. For most of the season, I’ve really enjoyed Todd Reirden’s defensive choices. And I suspect that I like the team’s defensive roster more than other folks do. I just don’t like this particular combination of them:

Carlson – Dillon
Orlov – Jensen
Kempny – Gudas

That first pairing has been fun, but the second has had trouble transitioning to offense, and the third has been struggling all over.

So let’s figure out some better combinations. I’ll begin with an update of an exercise I did a while back. Where a pair of defenders intersects below, that’s the percentage of expected goals the Caps get when they’re on the ice together. Fifty percent is even, green is good, red is bad.

That’s a starting point for what we’ll do below, but I also want to provide that same information in a more straightforward way. Here’s each pairing with their expected goals percentage and time on ice. Please note the small sample sizes for Carlson-Dillon and Orlov-Kempny.

One last preamble: I think it’s helpful to visualize how pace plays into this assessment. The horizontal axis is how many expected goals opponents generate (left is bad, right is good). The vertical axis is how many expected goals the Cap generate (up is good, down is bad).

So, with all that information loaded up, here are my approved defensive pairings. I’m going to take a few stabs at this to explore some open questions — starting with Brenden Dillon. Here goes.

Ideal if Dillon is the real deal

Carlson – Dillon
Orlov – Gudas
Jensen – Kempny

We haven’t seen enough of Brenden Dillon yet. It’s been just 67 minutes, most of them with John Carlson and against some weaker teams. But the Caps have done darn well (58.1 percent of expected goals, even in goals, in some high-event hockey). If that holds, that’s awesome.

Beyond that, you’ve got an Orlov-Gudas pairing that is similarly high-event and also quite good (54.0 percent of expected goals), despite defensive deployments.

Finally, there’s Jensen-Kempny, who have played just 101 minutes together, but who have had 59.2 percent of expected goals in usage that I wouldn’t call sheltered. I think that’s interesting, and we’ll dig into it more in a moment.

This setup is also probably the most offensively potent of all the options. If your team needs a goal real bad, this is who I’d go with. But that all assumes what we’re seeing with Brenden Dillon is sustainable. Let’s go the other way.

Ideal if Dillon is a mirage

Carlson – Siegenthaler
Orlov – Gudas
Jensen- Kempny

But we’ve only had an hour and change of Dillon. The rest of this exercises assumes he’s not so special. Prior to the trade deadline, these are the pairings I would have gone with. I suspect these picks are a bit unconventional, so I’ll try to defend myself.

We’ve got Carlson-Siegenthaler together as a top pairing. They’ve controlled 54.9 percent of expected goals in a decent-sized sample (195 minutes) of offensively optimized deployments. More importantly, they slow game pace better than any other Carlson pairing, which I find encouraging for a player who plays a bit too high risk for my tastes lately.

Below them, we’ve got the same pairings: Orlov-Gudas and Jensen-Kempny. They’re both solid. We’ll talk about that more below.

These pairings are also my preferred setup if the Caps need to hard-counter an opponent, as they have the best shot-suppression numbers as a unit.

I had built a bunch more variations, but they all ended up similar. Orlov-Gudas and Jensen-Kempny really do seem like the best ways to use those four players. Carlson’s been very successful with both Siegenthaler and Dillon. And it’s been a happy coincidence that the combos who are good at shot-suppression also happen to be good at generating offense.

But there are some pairings I have not liked at all. Here are three, along with their expected-goals percentages together:

• Carlson-Kempny (48.8%)
• Gudas-Kemnpy (44.2%)
• Orlov-Kempny (42.9%)

If you’ve detected a theme, yeah. It sure does seem like Michal Kempny has trouble with anyone who isn’t Nick Jensen. (The Orlov-Kempny didn’t have much of a sample size though; just 48 minutes.)

I’m worried about Kempny overall, so here’s how I’d go if I had to build pairings without him:

Carlson – Dillon
Orlov – Gudas
Jensen – Siegenthaler

That’s . . . okay. I don’t like the third pairing of Jensen and Siegenthaler. They have been outplayed in their 176 minutes together (48.6 percent of expected goals), and I think there’s been some confusion around roles and assignments that may have played a part in Siegenthaler’s unfortunate benching.

When I advocated for resting Kempny last week, I hadn’t thought through the implications. Now that I have, I can see why the team might be hesitant to take him out of the lineup (if they’ve considered it at all).

I want to end where I began. I think the Caps have made noticeable progress in the last week. I think Brian MacLellan has been thoughtful and precise in staffing up the blue line in recent years. I think Todd Reirden is a smart coach with the right resources to be successful. There are definite flaws and weaknesses, but the success we saw in October and November could absolutely return — whether or not they use the pairings I like.

What defensive pairings would you like to see and why?

This story would not be possible without Natural Stat Trick. Please consider joining us in supporting NST on Patreon.