Maybe I've lost the context, but I don't follow."I don't know if I can do it, or even if it's theoretically possible but if it is possible then I know I can do it."
Yeah, no. Not buying it. Of the various possible ways to run this, that one seems less probable than most alternatives. More likely candidates include, "I don't know if it's possible, but even if it is possible, I don't know how to do it," and "I am fairly confident that it is possible, and if my understanding is correct, then I can probably perform that task."
99% of new business owners - "I don't know if it's possible for me to start and operate successful business, but I'll give it a go". Within 5 years around 80% of them find out that it either wasn't possible, or that if it was possible they didn't know how to do it. Many of that 80% won't know which of those two possibilities was true of them.
Me, sitting down to write a paper - "I don't know if I can find an argument that supports XYZ, but I'll give it a go". Sometimes I don't find an argument, but remain committed to XYZ, and so use rhetorical devices to minimise the appearance of gaps in my reasoning. Sometimes I change my mind about XYZ - rarely completely, but at least in nuance. Sometimes I find the argument I wanted.
A chaos mage knows that the world was created from chaos, knows that s/he can create magical effects by manipulating chaotic forces, knows that the magic items are a type of magical effect that exist in the world, and knows that ambient chaos energy is present. Can s/he harness that energy to create a magic item, much as the Primordials harnessed that energy to create the world, and much as s/he harnesses that energy to create spell effects? Perhaps - it's not obviously impossible, but the ways of chaos are (naturally) chaotic. Let's give it a go! (Which, in the real world of playing a game, means - Let's set a DC and then roll the dice!)
You seem to be saying that there is no such thing as a purely metagame rule.In a wider sense, I would take the view that there must be an in-world reason for PCs to have more hit points (if that's what the game says). There really is no such thing as "rules without in-game reasons"; game rules, by definition in my view, have in-game reasons. They might not be dictated to you in the rulebook, it might instead be left to the players' imaginations, but there must be reasons; that's just fundamental to what reasons are.
I don't think I agree.
In BW, a PC's abilities can advance with practice or training. They also advance from making checks in the game, but there are two sources of constraint on this: (1) because of "say yes or roll the dice", checks are only made when there are genuine dramatic stakes; (2) the rules are very strict that advancement can't occur unless checks are made at a range of difficulties, including (near-)auto-failure, so players have an incentive to manipulate their dice-pool resources and the situations they find themselves in to generate that range of tests, which itself helps achieve dramatic pacing and outcomes.
Those are rules, but I don't think they reflect anything about the gameworld. The gameworld I am using for my BW game is Greyhawk, but GH hasn't suddenly become a disc-world style parody driven by the logic of drama.