Amidst trade buzz, Caps forward Andre Burakovsky had not dressed for a game in nearly two weeks. But now that we’re into the holiday roster freeze and he cannot be moved, there’s less of a reason to keep Burakovsky out of the lineup. Still, considering that the Caps have won 12 of their last 15 games, it was probably not an easy decision for head coach Todd Reirden to bring him back. Tonight Burakovsky steps in as Dmitrij Jaskin and Travis Boyd will be scratches.

While it’s good to see Burakovsky back in the lineup, I wonder if the team’s anxiety over its mediocre penalty kill is sabotaging lineup decisions overall. Ideally, Burakovsky should really slot in for Devante Smith-Pelly.

Jaskin and Boyd are curious choices for scratches. The fourth line trio with them and Nic Dowd has been the best the Caps have seen there since Winnik-Beagle days of yore. But Coach Reirden said today that he wants more speed in the lineup, which at least explains Burakovsky’s return. Speed is something Smith-Pelly lacks, but it seems that he’s off-limits because of special-teams considerations. I’m not sure that’s the right evaluation.

On dashboard stats, there is not a big difference between Smith-Pelly and Burakovsky. Smith-Pelly has four goals and three assists in 33 games. Burakovsky has five goals and three assists in 29 games. (The real gap, I suppose, is how far below expectations Burakovsky has been, which I whined about at length last week.) Smith-Pelly and Burakovsky both play low minutes at 5-on-5, but there’s been a distinction in what happens during those minutes.

Hockeyviz has a helpful graph that understand the difference in shot-attempt percentages when teammates are playing without a player and vice versa. When the red box is nearer the top right and the blue box is nearer the bottom left, the implication is that teammates are brought down by the player. In the graph below, the player is Smith-Pelly:

With few exceptions, the effect is uniform: the Caps play more offense and less defense when away from Smith-Pelly. Only Eller, Connolly, and Carlson outshoot opponents when with Smith-Pelly. Burakovsky is not a stud in this respect either, but it’s a far more mixed result.

The numbers for how Smith-Pelly impacts the Caps’ ability to control play won’t surprise those who’ve watched him skate this season. Though he’s still got impressive stick skills when he’s on attack, he’s playing so far below the pace of the game that attacks rarely happen. With Reirden emphasizing speed in his media time on Friday, it’s curious how Smith-Pelly escapes criticism.

But even with speed Andre Burakovsky certainly hasn’t been setting the world on fire either, and he’s got the extra demerit of not playing on the penalty kill.

Then again, considering how bad the Caps PK has been, maybe that’s a good thing. Washington has killed 75.2 percent of their penalties, 27th in the league as of December 20, and it’s been pretty stable.

Except I made the case last month that the Caps PK isn’t as bad as it seems and is bound to improve with time. The Caps now have the worst PK goaltending in the league (80.7 percent), but they’re closer to league average at limiting opponent rate stats (good at overall volume, average at scoring chances, not great at high-danger chances). Aside from committing way too many penalties, Washington’s penalty kill will be fine in the end.

But Smith-Pelly does not bring stability to the PK. He’s probably the one sinking it.

Using Natural Stat Trick data, here’s a color-coded table of opponent rates during the Caps PK for various skaters:

• TOI – time on ice during penalty kill
• CA/60 – opponent shot attempt rate during penalty kill
• FA/60 – opponent unblocked shot attempt rate during penalty kill
• SA/60 – opponent on-goal shot attempt rate during penalty kill
• GA/60 – opponent goal rate during penalty kill
• SCA/60 – opponent scoring-chance attempt rate during penalty kill
• HDCA/60 – opponent high-danger chance rate during penalty kill
• Sv% – WSH save percentage during penalty kill

With 66 minutes of PK time, Smith-Pelly plays more than any forward except Lars Eller. During that time he allows opponents more shot attempts than any forward except Tom Wilson, whose sample is smaller due to suspension. The numbers continue as we narrow down for shot quality, with Smith-Pelly allowing high rates of scoring chances and high-danger chances (though not as bad as Evgeny Kuznetsov, who, woof, glad that experiment is over).

But the penalty kill is a team effort, and network effects can account for a lot of what we see in on-ice stats. To address that, I pulled opponent rates for four penalty-killing defensemen when they’re with or without Smith-Pelly.

With the exception of Dmitry Orlov, Caps defenseman have tougher jobs when they’re killing penalties with Smith-Pelly. It’s a feeding frenzy when he’s out there, and you can see in these heatmaps from hockeyviz. Purple means the opponents get more shots from that location compared to league average and green means fewer. Left is the Caps PK with Smith-Pelly, right is them without him.

With DSP	W/o DSP

While the Caps do a good job limiting peripheral shots (the green at top left) in both cases, the Caps are bleeding chances up close — which might explain why Holtby’s having a hard time keeping pucks out of his net.

Devante Smith-Pelly had a magical postseason, and I’m glad he’s back with the club, but it’s worrying to see him push interesting players like Burakovsky and Jaskin out of the lineup so he can struggle on the PK.

Headline hhoto: Cara Bahniuk