"Not dying" is not equal to "winning".
I think, to first approximation, they are equivalent. You are correct that death isn't the only way to not win, but really, if the PCs are going into the hazardous scenario to stop the bad buy, they live and the bad guy dies, generally, that means the PCs won. It is a broad generalization, I admit, and applies more strongly to the "beer and pretzel" gamer than to the highfalutin' types we get in these parts, but we are perhaps in the minority.
When more interesting situations and goals are introduced, it becomes perfectly possible (and very fun in play) to lose a fight and live, or to win, dying.
Can we substitute "complicated" or "subtle" for "interesting"? I expect it preserves the meaning you intend, without speaking to what others find interesting.
If the character dies because the player decided to put everything at stake during a climactic conflict - than they are exactly the emotions we want to have at the table.
Well, the emotions I'm thinking of you don't ever want at your table. The scenario you're talking about can avoid the emotions I'm talking about, yes, but that scenario is only one way of dying among many. You're not just talking about the general frequency/ease of death, but of *how* the character may die.
It depends on both. You won't have a consistent game if the system you use does not fit the setting.
I don't think we are disagreeing in essence here. Surely, not all settings can be well-reproduced with all systems.
Let me take D&D as an example. In D&D, there is magic that revives characters from the dead - those powers are in the rules. How many clerics the world has to use those powers, and their willingness to do so, are setting conceits, not given by the rules. How the world reacts to those powers (or if the world even knows they exist) is likewise a setting issue, not a rules-issue.
What players do not know - and what neither superhero comic books nor LotR help them with - is how people think and act in a world where real death is not present.
I think you've jumped to a particular setting conceit - that "real death is not present".
You don't need to have resurrection available for everybody. It's enough to have it happen from time to time to put any typical setting on its head.
I think maybe your version of "from time to time" may be different from mine, and maybe you have a different vision of how well real information travels through a world.
In the *real* world, where as far as we can tell nobody ever comes back from the dead, virtually every culture has myths and legends that it happens. I think that, for far-pre-information age cultures, where documenting an event is mostly an issue of hearsay, I don't feel the issue is as big as you paint it.