Should "Bad Luck" Last Longer in the Playoffs?

Mitch Marner is famously on an 18 game goalless streak during the NHL playoffs. Additionally, the Vegas Golden Knights' have chronically underperformed their expected goals for 3 postseasons in a row (34 games as I begin writing this). In both these cases, we have begun hitting that point where people are generally unwilling to accept this as simply luck. It has been 3 years in both cases, so this is not unreasonable. But, I am, apparently, far less certain than everyone else that 3 season-long playoff trends cannot be bad luck.

I don't think either of these trends would be league-wide stories in the regular season. We would all look and say that is simply the wrong side of variance. After all, we all know nothing can make up for a lack of data. And it is worth stating these samples are extremely small for any real analysis. We are talking about less than half a season for Vegas and fewer than 75 shots for Marner! But, something changes in the playoffs, people are unwilling to accept these poor outcomes as a result of luck.

The problem is, I think, if anything, we should be more willing to accept the luck/variance argument in the playoffs than the regular season. I believe the sustained periods of bad luck are more likely to persist for long periods of time in the playoffs, not less. The reason why comes from a concept familiar to any poker player that has ever moved up in stakes.

How Poker May Help Us Understand Playoff Variance
Poker is a way to gamble while playing cards. Who ends up winning can mostly be a function of luck in the short run, but will mostly be a function of skill in the long run. (The same is true for hockey.) In poker, there are various stakes (I.e. how much money you are playing for) that determine the quality of players you will be up against, on average. When you move up, beginning to play higher stakes with more money on the line, you will begin playing more difficult opponents. 

When doing this, any good player who has risen through the various stakes will tell you something very important you need to be ready for. When you start playing more difficult opponents, you need to be ready for your "downswings" to last longer. Downswings in poker represent a period of sustained bad luck where winning players lose because of bad luck in spite of a skill edge. You can think of this as being PDO'd by an inferior team in the NHL. What happens in poker, is that as a winning player moves up, they begin to play against opponents who are closer to their skill level than their previous opponents were. So based on sheer will and skill, the winning player cannot play their way out of bad variance as easily. Why? Because as their opponent's skill gets closer to their own, they need more things to go right to continue winning. As a result, a run of cards that might have led to a 2000 hand downswing at the lowest stakes may become a 3000 hand downswing at higher stakes.
​
Basically, the downswings begin to last longer as the skill gap between you and your opponents gets smaller and smaller, and I suspect the same happens in hockey. But first, let's see what I mean from my poker database. I looked at my data from 3 different levels, we will call them low (the least money on the line), medium and high stakes. Here are my results in terms of how many hands my longest downswing lasted.

As expected, the higher the limits, the better the competition, the longer the downswings tended to last for me. And as mentioned above, this is closer to the exception than the rule. It is something most good players tell you to prepare for when moving up in stakes. While no analogy is perfect, I suspect something similar probably happens in hockey. At least in theory, it should. Difference being moving up in stakes becomes playing in the playoffs. In the NHL playoffs, any good team's skill edge should decrease, on average. Especially when looking at the value of a teams shooters compared to the opposing goalie. Because not only are the opposing goalies going to have generally performed above average, but there will be no backups, barring injury. The resulting increase in competition should likely produce a similar effect right? 

As a result, if I am right, we should expect the opposite of general wisdom among smart hockey fanalysts. Sustained runs of bad luck should be more likely to occur during the playoffs, not less because as skill edges shrink variance becomes even more important, and even more punishing to those who are on the wrong side of it. Meaning, if a trend would not worry you during the regular season, say an individual like Marner struggling over like 60 shots, or a team like Vegas struggling over 35 games, then you should be even more willing to excuse it as luck in the playoffs right? I know the emotional value we place on playoff results will make this almost impossible, but I believe it is the logical thing to do?

All the question marks are there because I am not aware of a way to test this given the differences in sample sizes in the playoffs. So I am open for ways to test this theory, I would love to hear ideas, and theories counter to mine, testable or not. That being said I think the logic holds up. My prior is to trust the poker community's understanding of variance more than the hockey community, but I'm curious to hear people's thoughts about that too. And if I am right, that would make it a huge mistake to trade away great players (Marner), or make fundamental changes to a your elite team (Vegas) based on poor playoff luck (when defined as an uncharacteristically bad stretch of finishing or point production kind of thing).

If I am right and people are finding false signals within even more noise than this very random sport usually produces, it may even mean that picking up players who are labelled as "playoff chokers" is a market inefficiency. Because these players are simply being punished by variance more harshly than we tend to give them credit for. Remember somebody actually is that unlucky, no matter how annoying it is to watch. Of course, some people probably do chronically underperform in the playoffs because of nerves or whatever rather than luck. Maybe one day we will find out certain players truly have the clutch gene. The problem is as purely an outside observer, I will probably never have enough data to be confident someone truly is incapable of playing playoff hockey. And hey let's be honest, without the benefit of hindsight, you don't either. ​