Back in early November we checked in on a Caps power-play unit that was struggling to convert its opportunities at the rate to which the hockey world had grown accustomed. And while it was still quite early in the 2016-17 season, the slide we identified dated back to the middle of last season - a 48-game stretch that was enough to raise some eyebrows, if not red flags.
What we’ve seen in the nearly two months since is a bit of an uptick, as a 7-for-16 five-game span in mid-December propelled the 25-game stretch to a respectable 15-for-78 (19.2%) overall. Of course, that means that over the remaining 20 games, the Caps’ power play went just 8-for-62 (12.9%), including an 0-for-17 run that ended last night against Toronto.
Add these 25 games to the 48 regular season matches that preceded it and you’ve got a Caps power play - for years the most feared unit in hockey (a distinction that Columbus no doubt now holds, at least temporarily) - that has converted on a very pedestrian 16.3 percent of its chances (37-for-227) in 73 games. (As a point of reference, that would rank them as the 21st in the League today, a tick below their actual 17.2%/17th-ranked unit.)
But as we noted in that November post, goals are rare events, so conversion rate isn’t necessarily a great way to judge a power play’s efficacy if your concern is predicting future success. To that end, per Corsica Hockey, the Caps’ shot-attempt rate per sixty minutes of five-on-four ice time (CF60) over the 73 games in question is 93.4, which would’ve ranked 10th in the League last season; solid, if unspectacular.
But here’s the rub - the Caps’ five-on-four goals-for per sixty (GF60) over these 73 games is just 4.7, a mark that would’ve bested only Tampa (4.3) and Vancouver (4.1) a season ago. Obviously there’s a disconnect between shot generation and shot success, a relationship more commonly expressed as shooting percentage. Sure enough...
The Caps’ overall shooting percentage and unblocked shooting percentage (Fenwick) are both at the lowest levels we’ve seen in “the BtN Era” of advanced stats. We’re talking lower-than-Hunter, and well below half as efficient as the unit was during the lockout-shortened season in which Adam Oates brought “1-3-1” into common parlance.
Again, there’s some variance at play here, and we’d reasonably expect these numbers to regress upwards some. But how much? First, let’s take a look at the Caps’ actual five-on-four goals versus their expected (and for more on xGF, head back over to Corsica):
For the first time since Dale Hunter was behind the Caps’ bench, the Caps are underperforming their expected goals while a man up. Why might that be? Bad shooting luck is a contributor, for sure. Here are the individual shooting percentages for the Caps’ top unit over the past three seasons:
There are some numbers that stand out here, but we’re into really small samples now - give Alex Ovechkin one more goal and he’s at 15%, take one away from Johansson and he’s at 20%, two more seeing-eye shots from Carlson in either of the past two seasons and he’s at 9.7% and so on (and if you think John Carlson has been missing the net a lot this year, it’s because John Carlson has been missing the net a lot this year). Those breaks add up, of course.
But here’s the deal: when you have the single most reliably dangerous weapon in the NHL - Alex Ovechkin - on the power play, you get him the puck as much as possible (deep insight, for sure). That’s been the Caps’ bread and butter for years, of course - Ovechkin is a volume shooter and a deadly finisher, converting on 17.2 percent of his five-on-four shots since the start of the 2013 season.
Among skaters 23 skaters (14 forwards) with over 500 minutes of five-on-four ice time and 400 shot attempts over the past five seasons, only Joe Pavelski has converted a higher percentage of his shot attempts (10.3), i.e. Corsi Shooting Percentage, than Ovechkin (8.7), and Ovechkin’s 975 attempts are 500 more than Pavelski has - not bad for a guy who isn’t parked on top of the crease. You get the point and probably see where this is going, which is here:
The Caps’ overall power-play shot rate has dwindled and Ovechkin’s share of those shots has also declined, while the share of shots coming from the top of the zone (i.e. the defensman “1” in the 1-3-1) has increased. In other words, the Caps have swapped out (either by design or, more likely, a combination of necessity and execution) dangerous shots for less dangerous ones, to predictable results (and this may go a long way towards explaining that xGF/actual GF discrepancy, as Ovechkin likely outperforms what’s in the xGF model for shots from, well, anywhere). Comparing this year’s first unit to last year’s paints an even more stark picture:
Ovechkin took half of the Caps’ top unit’s shots last year and a third this year; Carlson’s share has more than doubled (and while it’s great to see an uptick in T.J. Oshie’s shot attempts, the two Swedes aren’t shooting much at all, which no doubt makes the other side of the ice easier to defend).
To put a finer numerical point on it, Alex Ovechkin attempted 51.4 shots per hour of 5v4 ice time in 2014-15 and is at 40.1 so far this season - the overall reduction in shot rate and the reduction in his individual share of those shots has resulted in 22% fewer Ovechkin shots per hour at 5v4. That’s... not ideal.
So the ball’s in Barry Trotz’s court. Tweak the power play, adjust to the adjustments, or settle for a middle-of-the-pack extra man unit. With the talent the Caps have - both on the ice and behind the bench - the former is really the only option.