I don't know how to answer that.

To me, the hypothesis "X" where X is an event/behavior is simpler than "R(X)" where R is an overarching something that executes the rules of X. I should then prefer a model of the universe where there is no overarching thinghy that runs behaviors, just the behaviors behaving, over the other way around.

This sounds like there's just some confusion somewhere. If the behavior is a fundamental "Do(X)" and nothing else, then why does there need to be anything above or around that executes or hits run on the behaviors outside the simulation?

The ability to simulate X using R(X) is only very weak evidence that X requires R() (or any R' ) to function.