That makes it entirely academic, though, doesn't it? The point of an RPG is not to have a mathematically "better" RPG, it's to play it and have fun. What's the point of calling an RPG "better" if said definition does not mean it's any better at doing what it's designed to do?
What I'm saying is, I reject any definition of a "better" RPG that does not take into consideration what an RPG is for.
Well, that brings us back to Umbran's point, from the other angle. Is 2d10 or 3d6 as a resolution mechanic better than 1d20? Impossible to say. Does 2d10 or 3d6 produce a bell curve of probable options while d20 is linear? Yes, most definitely. Ergo, if we want a curve in the main resolution, 2d10 or 3d6 or some other such variant are highly likely to be better than d20. If we want linear, the opposite.
That doesn't say anything about perfection, either. Just because we want linear, doesn't mean that 1d20 trumps percentage dice or even d12 or something, at least not on that basis alone. It does pretty much trump 2d10 or 3d6.

A is better than B doesn't say anything about C, D, etc.
Moreover, and this is where the point becomes really relevant, one of the multiple dice curves are likely to be much better than d20 + (bunch of stuff to fake a curve). And for that matter, d20 by itself is likely to be a lot better than d20 + (bunch of stuff to fake a curve), even with the curve being now missing. So the only way this won't be true is if the "bunch of stuff" add something important to the game, that you couldn't get more cleanly by using one of the more straight-forward options. Thus, once we know if we want, generally, a curve or linear, and what else we want on top of it, we can examine the main die resolution
and the supporting materials objectively to see if they do that.
Sometimes when you do that examination you find out that the analysis of the desires is somewhat incomplete. (This is what happens someti9mes when people, for example, disparage hit point or Armor as AC without reallying considering the full range of what they are intended to do.)
It's impossible to be in any way objective about pieces of game design without clear analysis of the design. Once having arrived at the clear analysis, the critique of the design is likely to be less controversial, because it will be couched in terms of the analysis. Or on the flip side, it's not so much that people are arguing with bad logic, as that they are assuming premises that aren't shared or even articulated.
