I understand what you mean. But it's not complexity as it's defined in the realm of design. That's where the term depth is used (in some fields of design). Complexity leads, amongst other things, to something being hard to understand, and that's almost never a good thing. Now, again semantics, some will say "No but I enjoy bashing my head against a hard puzzle to solve it, or I enjoy having a complex system to toy with and try and find an outcome". And that's where the separation is very important. The fun part is solving the puzzle, not understanding what the puzzle is. It's inevitable, every user will have a moment where they analyze what's in front of them. But a puzzle fails almost systematically if the user doesn't understand it well. Same thing with the complex system, you enjoy toying with the parts once you have understood them, understanding them is rarely the goal or the pleasant part.
And the reason why I'm drawing pretty sharp lines is that people will often confine talks about design by limiting it to game design. But design as a whole is a discipline, and designing a shoes, designing a car, designing an onboarding process for a business, designing a game; they're all not that different and just have some subtleties and other disciplines that come and mix with design.
So concepts like complexity being undesirable is a very well known and established principle in design, and has been for decades.
Now, I'm not saying there aren't some thought experiments, or indie games or niche exploration that are not turning the concept on its head. You might have a game that's so ludicrously complex that it becomes funny, or you have some puzzle games that are so obstuse in their complexity, and yet have a very simple solution once you do, and the contrast between the two can be really funny afterwards.
But if you sit at a table with other designers. And one of them says "Well the game is very fun, and I think we have an appropriate amount of depth for what our users want. But I think it should be more complex", you have a problem.
I don't want to agree to your jargon carte blanche here, but for the sake of this post let's let that pass.
I still fundamentally disagree that the complexity you speak of is generally undesirable. Learning to understand what the puzzle is can be the most fun part of the puzzle. The puzzle does not fail systematically because the user doesn't understand it. In fact, I would go as far as saying that a DM giving players a puzzle they don't understand can be one of the most important hallmarks of a TTRPG. Sometimes that puzzle is a literal puzzle in a dungeon. Sometimes that puzzle is an emergent narrative that no one at the table understands when the game starts. Sometimes that puzzle is part of the gameplay between players as they learn to interact with each other, and as they make choices leveling up. Sometimes that puzzle is figuring out character builds or strategic combat. And all of those complexities can be part of the mechanics of the RPG itself.
I can give examples if needed, but suffice to say I do not write these "sometimes" off as thought experiments or niche games. Going back to the OP, I believe Mearls' 20% figure was an off the cuff estimate rather than a hard value, but is much closer to accurate level of desirability than your assertion that players with who enjoy complexity can be handwaved away as outliers.






