Follower of the Way
Incidentally, I was thinking of exactly what you describe in your second paragraph when I said, "sometimes even just one subtype" of G&S and "superior at delivering a specific kind of G&S-based play." (Edit: That is, thinking of the fact that there are these two distinct styles of G&S, lite vs crunchy.) I don't think I thought of you by name, but given our interactions over the years, I'd be lying if I said you had had no influence on that.This is an interesting perspective that gives me a little more insight into why this might be desirable. I really don't like the idea of trying to play a game that I don't know the rules to. In fact, I'd argue constraining the world down to a system of rules that every actually follows is the primary appeal of games as an activity for me. Most descriptions of FKR I've run across do read quite a lot like "I really wish my players didn't know the rules and would stop asking me about them," which honestly sounds less like a fun time and more like a No Exit style of hell purpose built for me. "Play the world, not the rules" is less a freeing motto, and more a reminder that the actual world is a frustrating place.
What I find interesting though is that I generally want to solve the same problems that I think FKR is aimed at. I'd like players to achieve, as @EzekielRaiden is putting it, "Groundedness & Simulation" where the events of the game more or less make naturalistic (or perhaps more precisely "mechanical") sense, follow from causes, and results are determined procedurally. The difference is that I've always viewed resolving that as a design question, something that should sit inside the rules, and from which failures should be hunted down and pruned away. What interests me is how much that ultimately seems to rely on design by accumulation; I'd rather like to see a bunch of rules documents worked up over years of playing games as starting places.
Because this really is an unusual dichotomy, where it seems like both sides fully agree on the reason for the action, yet there are two modes and most folks get pushed toward one or the other, ne'er the twain shall meet. Either extensive structure, iterated upon so it becomes ever more precise and correct, or minimalistic structure, reduced as much as possible so error becomes merely a matter of incorrect decisions. As you say, it's a matter of whether one treats the difficult spots as "this is a design problem, to be addressed by doing more design," or as "this is an evaluation problem, to be addressed by giving better evaluations." Ultimately, both require something to grow in complexity and intricacy; it's just whether that complexity is formal procedure or informal intuition.