clock menu more-arrow no yes mobile

Filed under:

Of Turtles, Killer Instinct and Blown Leads

The Capitals are terrible at holding leads, right? Well, actually...

NHL: Boston Bruins at Washington Capitals Geoff Burke-USA TODAY Sports

Let’s get this out of the way off the top - there’s no excuse for going 26 minutes without a shot on goal at any point within a hockey game. It’s hard to realistically envision a lead big enough to justify it, or injuries that can explain it away. And, no, “good teams” don’t surrender three-goal leads (except, of course, when they do).

Last night’s Caps win over Boston was an ugly one, to be sure (and the second time in the past month or so that Washington has gagged up a 3-0 lead which, despite what we know about score effects, isn’t exactly a hallmark of a “good team”).

So the Caps are bad at playing with the lead, a problem that has persisted since at least the dawn of the Ovechkin Era and probably long before it, right?


In fact, the Caps have been pretty good, relatively speaking, at holding leads (at which, it should be noted, they’ve had plenty of practice... because they’re generally a pretty good team).

Before we jump into the data that will surely contradict what your frustrated eyes and disappointed memories recall, a word or two on score effects. Put simply, we expect teams that are trailing to out-play teams that are leading, whether the result of urgency, complacency, shifts in risk tolerance, or some other factor (most likely, all of the above). Put less simply... well, here’s a link.

Assuming that the Caps and Bruins started on a roughly even playing sheet (they were, after all, the League’s top-two score-adjusted Corsi-For percentage teams heading into last night’s game), once the Caps took the lead, we’d expect the Bruins to have a bit of an edge in five-on-five shot attempts as long as that lead held. As that lead grew, Boston’s expected shot share went up. That’s score effects, in a nutshell, and last night’s game was an extreme example (and, to be sure, not the only driver of what we saw). Graphically, it looked like this, with the thin line showing the actual Corsi differential and the thicker line showing the score-adjusted differential):

via Natural Stat Trick

(Sorry, should’ve added a trigger warning.)

Anyway, last night was bad. Closed-door, players-only-meeting bad. But it doesn’t mean that the Caps are bad, or even bad at protecting leads. In fact, so far this season the Caps have the 12th-best five-on-five Corsi-For percentage in the League when leading by a single goal, fourth best when up two or more and sixth-best when leading overall. If we shift the focus from underlying numbers to goals scored (obviously much smaller samples), the Caps are second in goals-for percentage when up one and ninth when holding any lead. Since Barry Trotz took over behind the Caps bench, in fact, no team has scored more or a higher percentage of five-on-five goals when up one than the Caps (60%), and their Corsi-For percentage has them 12th (thanks, PDO!). (Their numbers holding leads of any size are even better). The result, in part, is a League-best .879 win percentage when leading after one period and League-high 83 wins when leading after two during Trotz’s tenure. For all of the talk of the team “turtling” with a lead under Trotz, both the underlying numbers and the results, in aggregate, have been pretty darn good (though it’d be interesting to see how many of those “leading after one” wins was tied at some point thereafter, like last night’s).

All of this isn’t to excuse last night’s game or recent performances. Frankly, if the players weren’t upset with themselves after last night’s win, that would’ve been a bigger cause for concern than what had actually played out on the ice. And that’s precisely the point - the Caps are a good team that has performed well with leads... last night notably excepted.