The Dallas Stars power play isn't just fixed. It's some kind of monster.

Why the new top unit might be the best power play unit hockey has seen in years.

Special teams is fascinating because it’s only 20 percent of any given game, and yet within that 20 percent is an entire universe of new roles, systems, tactics, and philosophy. Milquetoast broadcasts will drone on endlessly about the momentum shift of a game by pointing out a big hit, or one of hockey’s many glorified wrestling matches, but nothing truly changes the momentum of a game like a violation committed by one team, and the opposing team capitalizing (or not) on that violation.

And then there’s all that informs the modern scoring boom. You have better shooters aided by skills coaches, which is a phenomenon that is itself aided by a regressive size bias in goalie scouting, limiting the pool of elite goalies. More skill at positions once considered ‘marginal’ like the fourth line or the blueline’s bottom pair. Rule changes. The technology players have access to, quite literally between shifts. Increased defensemen scoring. The list goes on. And that list includes increasingly dangerous power plays. Teams have gone from a 15 to 18 percent conversion rate to 20 and 23 percent.

For a hot minute, Dallas was not shifting momentum. We’ve been over this here at the Stars Stack, so I don’t want to talk too much about this. In late October, the message was ‘whoa wait, they’ve been really good under Steve Spott. Calm down people.’ In late November it was ‘it’s okay to be frustrated, but even-strength scoring is a problem too.’ We got to late December, and it was “this is where I think moving Jamie Benn and Miro Heiskanen to the second unit could pay dividends.” Yea, quoting myself. By late January, it was: enough is enough.

Now that Heiskanen and Benn are off the top unit with Thomas Harley and Matt Duchene in their place, I can write obnoxiously about how I told you so, right? Nope.

It’s important to distinguish bad from underperforming. Dallas was underperforming, and due regression. Before Heiskanen went down to injury negligence, the Stars were 7th in minute-adjusted expected goals on the PP. That’s an underperforming power play, not a bad one; arguably even an unlucky one more than an underperforming one. As noted, however, I don’t believe it was just about shot quality. Expected goals at even-strength are different than expected goals on the PP. What’s happening now is that the Stars are regressing at the same time that they’re dressing an optimal top unit.

But what this top unit is doing, nobody could have predicted.