Emphasis mine.
You're trying to draw a distinction which doesn't actually exist--it's not actually the case that FK and Braunsteins have no elements of subjective judgment to them. Rather, in both cases, the referee's subjective judgment is all that ultimately matters. You don't get to tell your referee that your battleship didn't really blow up; you don't get to tell your referee that nunchucks are actually awesome weapons; you don't get to tell your ref that the French cities being invaded have not actually been evacuated yet; you also don't get to tell your GM that his Spartan-flavored society actually produces weak and dysfunctional armies. The fact that the FK ref is running a slightly-fictional world and the GM is running a more-highly-fictional world is not in fact relevant, especially if they are both trying to perform the same job: good-faith extrapolation.
In both cases, the quality of the extrapolation (and overall simulation) depends on the quality of the GM/ref.
There's another role, scenario/adventure builder, but that isn't universally considered part of the ref's/runtime GM's job.