When the Canucks dispatched the Minnesota Wild in four games during the NHL’s return-to-play play-in series, it was largely on the backs of stars like Quinn Hughes, Elias Pettersson and (game four aside) Jacob Markstrom.

That said, the Canucks’ depth pieces pulled their weight, and in a match-up in which their bottom six was clearly outclassed on paper, the Canucks depth players held their own in terms of controlling play, even as injuries to Tyler Toffoli and Micheal Ferland forced extra players into the lineup.

When it comes to possession numbers, none demonstrated this quite so much as Zack MacEwen, who rocked at 83% expected goals for ratio at 5-on-5 in limited minutes in the two games he played against Minnesota, according to Evolving-Hockey.com. That led all Canucks skaters in the qualifying round by ratio, as well as rate both for (4.09 expected goals per hour) and against (0.80 expected goals per hour). None of those numbers are what one would consider sustainable, but that’s beyond the point. The NHL post-season is results-oriented, and when MacEwen was on the ice, the results were good.

So how did he do it?

Expected goals are a tricky stat. They’re intimidating perhaps to some, likely because they’re generated mathematically and don’t exactly correlate to actual events on the ice like shot attempts or scoring chances do, and yet they’re more predictive than any other publicly available measure of play control.

Shot attempts, or Corsi, are occasionally maligned because quality is neglected in its analysis. Scoring chances, on the other hand, count only those attempts on net that come from dangerous areas, or from naturally dangerous circumstances like rebounds and rushes. The downside of the latter metric is that it doesn’t account for high volumes of poorer shots — 100 shots above average over a large sample is still a calculable advantage even if the shooting percentage from those areas are low.

Expected goals are the best of both worlds, assigning to each shot a likelihood of that shot becoming a goal, taking in as much context as is feasible given the data the NHL provides.

MacEwen was in the black in most (but not all) types of shot metrics against Minnesota, but it’s the dynamite expected goals share that should provide the most optimism.

As the table above demonstrates, with MacEwen on the ice the Canucks out-shot Minnesota 8-3, out-attempting them 13-9 and, most importantly, out-scored them 2-0. They also generated 0.83 expected goals and allowed just 0.17. The only metric that wasn’t in the Canucks favour was scoring chances, which seems odd considering the xGF dominance.

There’s an important lesson in both how the different types of shot metrics are intertwined, but also how they can diverge in small samples. Let’s take a closer look at what happened when MacEwen was on the ice.

In a little under 13 minutes at 5-on-5, MacEwen was on the ice for a total of 22 shot attempts, with 13 directed at Minnesota’s net and 9 directed at Jacob Markstrom.

According to Evolving Hockey’s xG calculations, the two best chances that occurred in this time frame were a MacEwen slapshot 10 feet from the net and Bo Horvat’s tying goal, both in game four. Both chances were worth nearly a third of a goal (with 33% and 32% chances of going in, respectively). As you can see in the videos below, there’s no doubt that these are bona fide scoring chances.

According to Natural Stat Trick (since Evolving Hockey doesn’t track scoring chances), the Canucks had only two other scoring chances with MacEwen on the ice: a MacEwen net-front shot from game three and a Brandon Sutter high-slot miss from game four.

Meanwhile, the Wild had five scoring chances against the Canucks with MacEwen on the ice, but according to Evolving Hockey, their unblocked shot attempts amounted to just 17% of a goal in the aggregate.

If you’re a bit skeptical at the fact that these clips amount to legitimate scoring chances (especially that flub by Joel Eriksson Ek at the end), you’re not alone.

During the pause, Canucks coach Travis Green spoke with The Athletic’s Harman Dayal about how his team has used and continues to use analytics. Dayal subsequently published a phenomenal feature based on that conversation, and one of the areas that Green touched on was the occasional deceptiveness of location-based scoring chances:

In fairness to the numbers, while the Wild had five scoring chances against MacEwen, none of them were labelled as “high danger scoring chances” by Natural Stat Trick’s formula (and rightfully so). In fact, the Canucks haven’t allowed any high danger scoring chances with MacEwen on the ice, though they’ve generated three of their own.

All of this is to make two points. First, we always have to be careful when it comes to interpreting shot data in small samples, especially when it’s purely location-based. Expected goals are a more reliable measure, with Green even admitting to Dayal that the team has its own proprietary version of the metric.

Second, Zack MacEwen got off to a fine start. While it would be premature to suggest that this type of territorial dominance is going to continue, it’s certainly correct to say that he did his part to push the Canucks in the right direction as they dispatched Minnesota – he even drew the penalty that led to the game winning goal in game three. Not bad at all for a guy who some people in the town had written off before he’d even seen a minute of pro hockey (ahem).

Unfortunately, the Canucks are now facing a much stiffer test, and MacEwen and the rest of the fourth line got scorched in game one against the St. Louis Blues. If the Canucks are going to have continued success against the defending champs, they’re going to need contributions from up and down their lineup. Certainly we should expect the top six, and the Lotto Line in particular, to account for the bulk of the offence, but if the bottom six can follow MacEwen’s small-sample lead from the qualifying round and at least control play and draw penalties, it could give this underdog Canucks team a fighting chance.