Responding to the OP:
First, if I'm reading GNS theory correctly (and I think I have a reasonably coherent, if far from "completionist" understanding of it), in my view simulationism and gamism aren't on opposite sides of the spectrum in terms of playstyle. Gamism, as a whole, is fairly adjunct to simulationism. A highly simulationist system can still support a "gamist" agenda, assuming that the gamism as an agenda is at least mildly tempered.
Gamism as an agenda is more directly opposed to narrativism.
The two agendas' goals ---
"Here's a challenge, let's step on up and win!" (Gamism)
"Here's an interesting moral, ethical, or psychological dilemma, let's play out the consequences of that premise!" (Narrativism)
--- will ultimately have to implement highly divergent mechanical underpinnings.
Simulationism supports gamist drift, to varying degrees, without generally completely destroying the simulationist leanings.
Don't get me wrong, there's always going to be a "game" in an RPG. The question isn't whether "gamism" exists in an RPG, it's always THERE, whether the system wants it to be there or not. "Skilled" understanding of a game's rules can always be a means to manipulating in-game outcomes.
The question is whether a game naturally intends, supports, or implicitly accords with gamist desires as a way of appealing players' sense of "challenge" achievement and its vicarious social currency.
Narrativist games, almost by definition, tend to devalue gamism as an agenda, because "winning" a scene as a form of real-life social achievement is different from "winning" or "losing" a scene to explore the underlying moral, ethical, and psychological "matter" that make the scene "interesting" in the first place. From a narrativist perspective, "winning" or "losing" a scene is often equally interesting; this is rarely the case for gamist agendas.
But this is mostly rambling thoughts of my own. Back to the salient point, a la D&D 5e ----
The problem D&D has always had, is it doesn't really know "What it's simulating." And because it doesn't really know what it's simulating, there's no way to determine how well
it's being simulated.
Part of the reason 4e was a breath of fresh air for GMs like @
pemerton is it finally gave up the pretense of "simulating real-world reality," and just accepted the fact that "Yes, if we're 'simulating' anything, it's heroic fantasy genre reality."
I personally am more of a fan of systems that casually simulate "reality" as much as possible, but make clear, token exceptions to genre-appropriate elements. 3e's problem generally is that it's very unclear about that division---what should be accepted as "simulation," and what should be viewed as "What the heck, it's fantasy."
If 5e takes a stance on "simulation" at all, at least make it a clear one.
But of course, as always,
>>> The Game <<<