The Washington Capitals have played 334 minutes of five-on-five hockey so far this season. That’s scarcely longer than The Lord of the Rings: The Return of the King (Extended Edition), which is to say not very long at all. And, glancing at the standings, it’s not a bad start. The Caps are an inch above .500 with a 4-3-0 record. That’s not so bad.

But a couple layers deeper, things are really, really scary. Is it small-sample weirdness — or the first impressions of a collapsing hockey team?

Controlling the flow of play, measured by proxy in shot attempts, is a core competency for any good hockey team. The Caps have been outshot 327 attempts to 299 (adjusted) during five-on-five play, ranking them 21st in the league with 47.8 percent (50 percent means even). If they kept that up for a full season, it’d be their worst performance since… uh… 2018. Okay, maybe never mind. Skip this one.

When you factor in the quality of their shot attempts, Washington looks much worse. They control just 42.5 percent of the expected goals, according to Natural Stat Trick’s reckoning. That’s bottom five in the entire league. Washington’s gap between quantity (i.e. attempt volume) and quality (expected goals) is huge, with only three teams bigger: Arizona, Montreal, and Colorado. Okay, that last one isn’t so bad, but they’re going from 56 percent of attempts to 49 percent of expected goals, which is far north than Washington’s territory.

Splitting offense and defense apart, we learn a lot more about the profile of the team’s play. The Caps are mediocre at limiting opponent shots (ranked 15th in opponent shot-attempt rate), but near the league’s bottom at generating their own shot attempts (22nd place). That offense is much worse when factoring in shot quality, as the Caps drop to 28th place in their offensive expected-goals rate. The likelihood of the average Caps shot attempt to become a goal is the second lowest in the league. (Colorado is worse, which may explain a bit the prior paragraph).

HockeyViz provides heatmaps for each team’s offensive and defensive performance. Blue means fewer attempts come from a location on the ice; red means more. You want blue on defense and red on offense, but for the Caps offense we get a big ocean of blue.

That’s not good..

For the record, the Caps have done a fine job at limiting opponent shot quality. They’re just outside the top ten in opponent expected-goals rate.

Shooting and saving percentages — what we sometimes call “finishing” stats — are what really drive results in small samples. The Caps have the league’s highest shooting percentage during five-on-five play at 12.3 percent, and a just-fine .910 save percentage — ranked 16th place in the league. The goalies are fine, but the shooting cannot last; the highest a full-season shooting percentage has been in the last decade is 10.5 percent, by the Leafs in the shortened 2012-13 season.

In effect, the Capitals have scored 5.5 more goals than expected during five-on-five play, while they’ve remained right about at-expected on defense. That’s enough extra goals to deliver maybe two extra wins so far.

On the power play the Capitals are generating a league-average volume of shots (18th place by shot-attempt rate, 14th place by expected-goal rate). Because current models don’t really account for the danger of Alex Ovechkin’s shot from the faceoff circle following a cross-ice pass, I try to consider these statistics as underrating the team. Ovechkin has taken 25 of the team’s 56 attempts during the power play, which is a stunningly high proportion.

But to that point, I don’t want to delve into individual player statistics. Partially because the samples are even smaller, but mostly because it’s too gruesome. The Caps offense is effectively dead when TJ Oshie’s on the ice — about half the league average expected-goals rate. Opponents are going on shooting sprees during John Carlson’s shifts– higher than any team-level opponent rate in shot attempts, without much offense to offset it. Ovechkin’s five-on-five attempt rate is down by almost a quarter.

I’m not using this space to discuss possible underlying causes for these problems. There is one nightmare scenario that could be coming true, but we are way, way too early to make conclusions. This space is just for describing the current situation.

Every one of these statistics, without exception, will improve as the season grows. The open question is how much they will improve. These are still bad performances, but they are made to look even worse by early-season sample distortion — except on the team level, where the 4-3-0 record suggests a team that is better than average. Right now, the Washington Capitals are not that.

This story would not be possible without Natural Stat Trick and HockeyViz. Please consider joining us in supporting those sites.

Headline photo: Alan Dobbins