The Columbus Blue Jackets have an excellent PK. It’s truly amazing, fun, aggressive, and looks like a perfected version of what the Leafs do with Mitch Marner, Kasperi Kapanen and Ilya Mikheyev on the ice.

The Columbus system uses speedy forwards to disrupt the opposing power play setup and generate scoring chances, and they were near the top of the league in 2018-2019 by several measures of their own and the opposing shooting. They used the same system this year, generated less offence (just like they did at five-on-five), but kept the Expected Goals Against down to the league’s lowest number.  Their goalies were average (Elvis Merzlinkins) and below average (Joonas Korpisalo) shorthanded by Goals Saved Above Expected, just like they were at five-on-five. The combination of great defending and average goaltending produced the second lowest goals against, which is the main point of the PK, but all that aggression led to one goal for.

This what their PK looks like:

That’s an impressive level of defending, and keeps the Blue Jackets facing some of the lowest totals in Corsi Against on the PK of any team. Only the Flyers were significantly better this season.

But there is one PK that’s even better than Columbus’s this year, and it’s not the Flyers or even the Leafs.

It’s everyone else when Columbus is on the power play.

Teams facing Columbus shorthanded are even better at preventing net-front shots than Columbus is. And while they allow some shots from the circles, the slot is cut off, leaving only predictable shots from the outer edges of the ubiquitous 1-3-1 powerplay formations.

The whys and wherefores of the Columbus PK are pretty simple. The defenders are good at their job, and remain relatively stationary, while one of the two forwards forms the apex of a triangle with the defenders and the other harries the incoming power play. The two forwards swap places to very good effect, setting up opportunities for two-on-one rushes against.

Why, though, is their power play so bad? It gets to practice against a top PK, you’d think they’d be setup masters who can deliver the goods. The trouble seems to be the delivery men just aren’t really up to it. That and the inexplicable duality that is Oliver Bjorkstrand.

This is how Bjorkstrand shoots at five-on-five:

That’s very good, but not elite level individual shooting. The difference can be seen if you look at three Toronto shooters. Ilya Mikheyev is a little less selective than Bjorkstrand in his shooting locations, and has an Individual Expected Goals per 60 minutes by this model of .89. William Nylander is much more selective and has an ixG/60 of 1, and Auston Matthews is the best on the Leafs with 1.04, and he combines a heavy slot and net-front shooting pattern with sheer volume. Remember: shooting is a volume game, a location game and a skill game. If you do both volume and location, you’re very good. If you’re also skilled, you’re elite.  Mikheyev and Matthews shot at identical rates this past season, but one is a class or two above the other in location and skill.

Considering the volume of Bjorkstrand’s shooting — he’s in the top 30 in the NHL — he could stand to go to the same video sessions as Mikheyev to learn some discretion.

But put Bjorkstrand on the power play, and you get this:

That’s just objectively terrible. It’s a combination of Mitch Marner and Tyson Barrie, if you look at Leafs shooting patterns. (That’s not good, because both of those guys are there to pass not shoot.)

Bjorkstrand has played both the left and right wing on the power play this season, and is usually on the ice with Zach Werenski. Gustav Nyquist or Boone Jenner play the net-front, and the plan here really seems to be to use Bjorkstrand in a setup role. And that’s really weird, considering that Bjorkstrand is a shooter not a passer.  His “looks like a defenceman” shots come most heavily when Jenner is on the ice, so it could be a quest for tip-ins.

Making it weirder is that no one on the Blue Jackets shoots on the power play at even remotely the rate of Bjorkstrand. The only player on the Maple Leafs excellent power play who shoots more than Bjorkstrand is Auston Matthews.

So, what I want you to do is picture the Leafs power play with Matthews told to guard the points and defer to the net-front guy. This is the most, maybe the only, glaring example of a player on this team hopelessly miscast in usage.

Compounding the power play problems is that Werenski shoots at a rate of a top forward, well in excess of Barrie’s rate, and that means the bulk of their shots are coming from the points, almost like they have two shot-happy defencemen and some stationary, ineffective forwards...

The (Many) Benefits Of Deploying Seth Jones And Zach Werenski On The Same Power Play Unit

Which seems to be by design, not some fault in Bjorkstrand’s own playing style. Look, who am I to tell the Blue Jackets that their heavy shooting from the worst-possible locations is a bad idea? Go for it. Go for it hard. The article shows 28 minutes of time with both defenders on the ice, and shows them shooting from locations normally occupied by forwards. So with all of these players healthy, using two actual shot-happy defenders might make the power play more unpredictable, or it might just dilute Bjorkstrand’s value as the lone skilled shooter in a whole new way!

From the article:


For years now, Tortorella hasn’t really referred to his two top blueliners as defensemen. He’s called them “rovers” and defined a position that isn’t like other defensemen or forwards, for that matter. Playing the role of a rover means Jones and Werenski have different rules, and it applies again to this possible power play configuration.

I think every Leafs fan’s eyebrows will go up at this claim of innovation from Tortorella. Imagine, if you will, a defenceman who roams around at or even behind the opposing net, shooting like a forward, passing like a forward, and leaving a better shooter to hold the blueline. Whatever would that look like? Speaking of Tyson Barrie, note that he actually shoots less on the power play, and feeds the puck to the best forward. He doesn’t move off the blueline either. But, of course, the Leafs have more than one good shooter.

The Columbus plan seems to be to use Jones more like a forward, like in the very short experiment the Leafs tried when Marner was injured and they put Barrie on the wing. It was workable as an emergency stopgap, but he was replacing a passer, not trying to be the main shooter.

Jenner is broadly as good as Zach Hyman at playing the net-front, with maybe not as good a shot. Nick Foligno also plays a lot of power play minutes, possibly because they just have no one else, and he moves the puck well. But that’s really the situation here. Even if Bjorkstrand was put in a position to succeed, the available personnel just isn’t there, which is the real reason for this “use the defencemen” concept. Only three teams in the NHL scored less on the power play this season: the Red Wings, the Ducks and the Islanders. So this trick might actually improve the Columbus power play, since it has nowhere to go but up.

Eye-testing the worst power play in the NHL

I watched a fairly random selection of Columbus games from this season, focusing on their power play, but including some other aspects of their game.

Zach Werenski is the single biggest reason their power play was so bad. He passes poorly, he shoots too much, and from low-percentage locations, and he’s not all that good at transitioning the puck to set up the power play. If he was a “rover” with Seth Jones on the ice, this year he hugs the left point.

Oliver Bjorkstrand is the second biggest reason their power play was so bad, and it was the role he was playing. He plays as if he’s the right-side defender most of the time, giving the formation a three forwards and two defenders look even when Jones was out with an injury. He also shoots a lot, and from low-percentage locations — the shot plots don’t lie about that. He visibly drifts back towards the blueline instead of moving into scoring position all the time.

On the penalty kill, the team is aggressive, dynamic and fast. They take their biggest five-on-five skill sets — picking off passes and poke-checking pucks — and turn them into the building blocks of rush chances in ways they never do at any other game state. It’s weird to watch them be totally inept at a five-on-three, failing even to control the puck, and then turn around and get a real scoring chance at four-on-five.

There’s something in all this about their coaching-induced attitudes to risk. If you go for it on the PK, you get rewarded, and the risk of a goal against isn’t any worse than it already was. It was already high, so you can get away with it without getting in trouble. On the power play, the forwards drift back towards the blue line and play like they’re terrified of giving up a rush against. At five-on-five, they get the puck and then cycle by keeping it against the boards in the offensive zone, passing from player to player, who protects it with his body instead of shooting it.

I made up a mean joke while watching them play: How do you know the Blue Jackets have the puck in the offensive zone? They’re skating it into a corner to protect it against the boards. I’m calling this the offensive zone trap, and it is their real innovation.

Micah McCurdy has said that a shot in hockey is the pivotal decision making play because it is the point at which you risk giving the puck away. The ultimate reward in hockey requires you to give up possession — the thing you’ve been spending all your time trying to acquire. The Columbus Blue Jackets only ever give up the puck willingly when their failure will be disguised by the circumstances of the game on the PK. And that’s why their power play is so bad. The very concept of a power play is antithetical to the team’s most sacred value: Never make a mistake.

To be fair, they do have a very low rate of shots against on the power play, which is not actually the entire point like it is on  the PK, but they are faithfully following their credo of fearing failure more than they desire success.