Manbearcat
Legend
I suspect that's entirely down to cause->effect being clearly mapped. Even if the character has no viable option beyond putting themselves in harms way, they can clearly see the outcome, and the outcome is consistent in response to the same input each time. Except possibly for that legendary action one, which is admittedly some 5e specific nonsense, and not the most elegant solution to the one monster vs. several PC action imbalance problem.
My experience in those games has generally been feeling like I'm waiting for the scenario to resolve enough that I can go play the game, and then realizing I don't have any tools to do that, and any time I try take an action, I'm risking the game unwinding further, and worse, I'm supposed to be okay with that and declare actions anyway. I keep waiting for the scenario to be clear, so I can try and solve it, and then realizing that will never be the case. It's like playing a roguelike that regenerates the level every time I take a step.
While I'm confident that this response is not representative of the entire "Success with Consequences Feels Bad" cohort of players, this is a component of it for sure and its likely the big piece for some players in particular (you for instance...given our exchanges in the past, this was how I figured you were oriented to the situation). So I'm curious about the bolded part in particular as it pertains to the below (if you'd care to unpack your thoughts):
In plenty of non TTRPG games (from PvP in CRPGs, some board games and parlor games like Pictionary, to ball sports to martial arts contests), engaging play involves a consequence-space that is both vital, dynamic, and evolving. The games are stochastic rather than deterministic. The prospective consequence-space is certainly neither unknown, unknowable, nor unbounded because its constrained by both (a) current gamestate and all the rules interactions that go with that and (b) the nature of relevant physical and intangible objects in the "play space" including both (i) their present relationships and (ii) their prospective future relationships post-interaction/collision. A huge part of the skill in game/sport is the ability to mentally process quickly and accurately build out a constellation of prospective outcomes and winnow those down to a working subset of most likely and gameplan (which includes adjustment in real time) around that (like the concept of "Fight IQ" or "Football/Basketball IQ" and all of the instantiations of the same concept). The best players are able to do this while under extreme durress...while the players who struggle in this (either the mental processing component or the "under duress" component) will be weeded out.
Clearly seeing the outcome takes on a different shape than what you've depicted above (it takes on the shape of what I've posted above). And while "the outcome spread" correlates to the inputs, I certainly wouldn't say the outcome is consistent in response to the same input each time.
So it seems to me that perhaps you're expressing a discomfort with stochastic models of play (and TTRPGs where the consequence space is such) and a related preference with/comfort for deterministic models of play? If so, does this express itself in the rest of your gaming interests (outside of TTRPGs...like maybe you don't like PvP in CRPGs or maybe you prefer certain ball sports to others)? If it doesn't express itself in the rests of your gaming interests, why do you think it is that it specifically expresses itself in TTRPGs (what is "the secret sauce" there)?