We had a recent thread about an article from Tim Palmer who has a good take on this:

https://news.ycombinator.com/item?id=33223391

The idea is that the natural laws are fractal, which when mapped out trace many state space configurations. But with a fractal there are always gaps in state space that remain no matter how much you zoom in or out; fractals preserve certain structural invariants at all scales.

So the idea is that counterfactual reasoning sometimes fails in a fractal universe, because statistical independence no longer works as a general rule (your counterfactual could lie in one of those gaps). He first published this idea in 2009 under the name "invariant set postulate":

https://en.wikipedia.org/wiki/Invariant_set_postulate

His paper with Hossenfelder is also a good overview:

https://www.frontiersin.org/articles/10.3389/fphy.2020.00139...

Superdeterminism gets an unnecessarily bad rap.