Even over the past 3 seasons? That's a combined ~180 games for the first 3 quarters and ~60 for the last quarter, per team.
You said yourself that not all teams revert to a PDO of 1. If it's likely just bad luck, how do we know that the Canucks' shooting luck will revert back to normal levels for sure? Wouldn't that imply that our goaltending would have to worsen?
I'm saying the Canucks seeing a drop in shooting percentage over a small sample size is irrelevant, yes. The last "quarter" of the season this year was like 10 games, too.
Look at it this way, for simplicity's sake. There are 30 teams in the league. Let's say you broke the season down into quarters, and say that each team performed above average shooting for two quarters and below average for two quarters. That's a gross simplification, but it should serve to illustrate my point.
Now, that would mean on average, 7-8 teams would shoot below average for the quarter of the season that fell last in the calendar. The next season, each of these 7-8 teams would have a 25% chance (by random distribution) of being one of the 7-8 clubs that had a bad shooting percentage in the last quarter again, so you'd expect probably two of the same teams to appear to "fall off the map" shooting for the second consecutive season.
In the third season, there would still be about a 50% chance that once of those two teams would replicate their poor shooting over the final quarter of the season again.
As I said, that's a really simplified way of looking at percentages, but it should highlight how a pretty random distribution can create a "pattern" that people will force narratives on to, even if it's pretty basic math. This sort of thing is bound to happen to some team. Even over 82 games, some teams will shoot at an unsustainable percentage, but are exceptionally unlikely to repeat it again. I'd say once you get to two seasons worth of hockey, most of those bumps in the statistical road will even out -- but that would be the last quarter of like 8 seasons of hockey.
Keep in mind I'm not saying there isn't some way to quantify shot quality, only that nobody has shown a way to do so that's convincing statistically. The Pens employ a firm that tracks shot location and they use that to make personnel decisions, but none of that data is public, so who knows how good it is. But using the shot metrics people have made public, we can say with relative certainty that on a large scale, shot attempts (your corsis and fenwicks) have a good correlative relationship with goal scoring/winning hockey games.
That's why I'd only find any of this interesting if for some reason there was some dramatic shift in who was taking shots for the Canucks. If we saw like a 30% uptick in shots from defensemen and a 30% downtick in shots from the Canucks 5 best shooting forwards, then maybe you could parse the data and see some kind of tactical shift that had occurred.
Edit: posted late. I meant to say that each team in the league shoots significantly below average for
one quarter of the season. If it was two above and two below, even more teams would fall into this category. I don't know which would be more accurate, but for the sake of showing how narratives get forced onto stats, we'll go with my original intent.