Hmm. I still think you need rigor in how you know what differences are important and which are ignored. There is no such thing as actually identical universe-states across branches, so you're somehow collapsing different physical things into "same" experience.

If you asked how many movies does a person know on that, they'd likely say 1, because they're collapsing all the copies, which vary only trivially. If you ask "what's the probability that a randomly chosen disc in a used DVD store is Avatar", then the ratio of individual copies matters.

The purpose and type of comparison matters a lot in determining whether distinct things are fungible.

To be clear, I totally get that two experiences can be indistinguishable as anticipation or memory for an entity. I'm just not sure why it matters when counting universe-states for the purposes of a probability estimate.