Mechanics can help get players engaged with the system and can impose a degree of unified structure on the collaborative story you're telling; the fact that the rules sometimes explicitly let you invoke fiat-level actions doesn't make them any less useful for this. (Unless you inherently hate the idea of players using powers that invoke narrative fiat, of course, in which case it's not the right system for you.)
I don't hate the idea of players using powers that invoke narrative fiat
in general. I have a problem with powers that invoke narrative fiat
in this context, namely, an otherwise mechanistic, process-driven,
extremely rules heavy system with a multitude of exception-based powers, particularly given the in-setting stakes involved with the powers in question.
Such a system demands a high level of buy-in from gaming groups, particularly in the case of the GM, who is obligated to know how the thing works so he can construct everything else in the gameworld apart from the PCs' sheets. The unspoken, expected return for spending the time and effort to learn a complex, heavy system is that the system itself provides enjoyment through engagement on the game layer or setting-emulation resolution layer. Should a system not deliver either of these things, a question arises: why did our hypothetical gaming group bother learning this (and, y'know, paying the game company for a 600-page doorstopper tome plus extra supplements) if it has zero to negative utility? Clearly not worth it.
My view is that this
particular kind of system, with this kind of buy-in, does not mix well with the prospect of throwing around narrative plot fiat mechanics balanced by social contract, because the presence of such things is more likely to cause problems than in, say, Fate, which is a much friendlier context for them.
For charms specifically, they also serve to make general statements about the kinds of things the Exalted can do, to give players a general framework for how difficult these feats are, and to generally tie them into mechanics so you can easily adjudicate some of the most common interactions without having to resort to narrative fiat. You'll notice that the charms that rely heavily on fiat (or on the ST "using them right") tend to be high-tier charms with restrictions on their use; God-King's Shrike is 1 / season, while Dual Magnus Prana costs 30 wxp and only comes up when you die. I agree that something that resolved common attacks through fiat would somewhat defeat the purpose of having mechanics for combat; but I don't agree that having charms intended to represent rare, pivotal narrative events involve a degree of fiat necessarily invalidates the rest of the system.
Do you seriously think the one a season cooldown on God King Shrike or the (negligible if you build for it, apparently - taking this board's word for it since I still don't have the enthusiasm required to go out, find the leak and read it) XP cost of Dual Magnus Prana matters in the slightest? Do you think Creation-Slaying Oblivion Kick would have been fine if we put a once a year limit on its usage, and Zeal would be OK if we charged Masky McMask or Falafel 10XP every time he tried to use it? Because that's cold comfort in the aftermath of their uses. Creation is still, well, kicked to death. Your entire circle has still been Zealed. gg, no re. The fact that Ketchup can't Kick again for a year or Masky lost 10XP is kinda unimportant.
Even those narrative-fiat charms, though, still serve a dual purpose in giving players a general glimpse into what the Exalted can do and what the setting is like. You're not supposed to be able look at Dual Magnus Prana and say "all right this is a power Solars get at level 16, so I can anticipate that 10% of the Solars in the Usurpation must have had it"; the game's mechanics (including its charms) are designed to let you run games of wandering adventurers in the twilight of the Second Age, and provide only a very loose reflection of things outside that scope.
But you are supposed to be able to look at Dual Magnus Prana and say "all right, this is the kind of epic feat that highly-talented Solars can sometimes do, and this is a rough indicator of how difficult and costly it should be relative to other powers." I wouldn't be surprised if, during the Usurpation, the Dragon-Blooded found that one of the Solars they'd killed was actually a robot duplicate! Then they tracked down the real guy (maybe cutting through a few more robot duplicates in the way) and eventually killed him for real. A few Solars did survive for a while, after all, and the ability to go "no, wait, that wasn't me, I'm actually over here!" is much less useful when you're fighting massive armies of people who know you very very well and therefore know most of the places you might be hiding.
Yeah, I don't think "I can flawlessly destroy any city in existence on the face of Creation with no possible defense or counter from my work desk once a season, note
no possible defense or counter" and such is "the kind of epic feat that highly-talented Solars can sometimes do", is the thing. Or rather, it shouldn't be, and if the game is saying it is, we have a problem.
If this is the kind of epic feat that highly-talented Solars can sometimes do, then, well, if my PCs' Circle has made highly-talented Solar enemies and has infrastructure that can be blown up, why is it not blown up? One can't simply go "because that would be a bad story", because we have established by the existence of this effect in the book (by your own words) that this (and things like it) is a thing Solars are supposed to do. So I have a choice now: are my Solar antagonists all retarded and unable to figure out that blowing up your enemy's base with a flawlessly auto-successful attack is a good idea (which would not be a good story), or is this
not a thing Solars do (and the book happens to be huffing paint)? Not fun, that.
Or, let's look at this from the other angle, and posit my Circle of highly talented Solars has this capability and we want to snipe our enemies' capital cities off the face of existence one at a time, once per season, as this would be a very useful military advantage. I can simply use my mighty GM fiat authority to declare that this isn't acceptable, but hey, apparently wiping cities off the face of existence is a thing you're supposed to do, so I'm kind of being a dick here. Not fun? And a bit of a waste of space of that Charm given that neither side is allowed to use it.
Or, let's make it worse and say I
do posit that my Circle's Solar enemies are in fact retarded, while not stopping my own PC Circle's city-razing efforts. This now looks like a farce, the sort of thing that should be accompanied by Benny Hill. Not fun. Why does this thing exist?
This partially answers your other questions; for one thing, the Creation-Slaying Oblivion Kick, Zeal, and OSOI are bad because, even taken as general statements of a capabilities, they make bad statements about the people who have them -- "kill everyone in the world in one kick" or "murder other Celestials automatically, bypassing all defenses with no warning" or "completely ignore all virtually all attacks for an entire fight" are not supposed to be epic feats available to SMA or Solars at all. When the books end up saying something that is obviously silly or which doesn't fit the design of the setting, then that's totally a problem.
The other reason is because Zeal, OSoI, and mote-attrition do, generally, undermine the ability of mechanics to impose any sort of useful unified structure or to get players engaged with the system. They break the game in the sense of making it fail to work at emulating the wandering adventurers in the twilight of the Second Age it was designed for.
And "I can build robo-Solar duplicates of myself which require no commitment or upkeep that apparently are so good in terms of emulation of my personality and Solar magical power they can make the greatest Solar detective on the face of Creation look like a rube with no possible defense or counter" and "I can destroy any city on the face of Creation with no possible defense or counter from my room (hurr watch me nuke the Imperial Manse)" are good statements to make about the people who have them?
TBH, not seeing a lot of difference here. City-Killing Oblivion Punch (with a 1 Season Cooldown!) is certainly
smaller in scope than Creation-Slaying Oblivion Kick, but it's just as impossible to stop and almost as irritating, as an example. If either of these things would have been released in Dreams of the First Age, would either of us have tried to defend them as not that bad?
But I don't understand why you see Dual Magnus Prana as a problem akin to those. The ST doesn't have to "fix" it the way they had to fix the other things you mentioned. They have to decide whether they want an NPC to have it and use it, yeah, but that's the kind of decision an ST has to make constantly anyway, in the same way that they use fiat to decide how strong NPCs are, how many of them you fight, what happens to stuff offscreen and so on. Bad decisions for any of those things can cause just as many problems as an ST spamming Dual Magnus Prana or God-King's Shrike in a careless manner.
(Unless you misunderstand what I'm saying? I'm not saying the ST should use fiat to prevent players from using Dual Magnus Prana; I'm saying the ST should use their own judgment in having NPCs use Dual Magnus Prana -- which they have to do anyway, since they're not actually tracking wxp for NPCs or anything like that.)
Like I said in my previous post, this logic can be used to excuse just about any problem in the game that we have ever seen in its publication history. From the perspective of someone who's spent
a lot of time talking about mechanical problems in the game which look remarkably like the sort of Charms people are talking about in this thread, this kind of argument makes me, let us say,
negatively predisposed to this product without even having read it.
Or is your concern that a player who has the charm could accumulate a huge amount of wxp and become nearly impossible to kill? I can answer that in another post if you want, but it's a somewhat different argument than the stuff about the way the charm operates on narrative fiat, so I should save it for another post rather than this already-long one.
Feel free to do so, it's useful data.