- P(E|H) is the likelihood of my current existence -- given the non-religious hypothesis of only one finite life, at most, for each potential specific self-awareness.
- Likelihood is not an after-the-fact computation. It's a before the fact computation. At the Big Bang, or the beginning of time, what was the probability that I would be here right now?
Jabba, let me fix this for you. Let's set aside complaints about the "pool of potential selves" existing or not, and just agree that as you've said you are talking about whether or not someone exists that would be really unlikely given all the other people that could have hypothetically existed instead.
I know this relies on the idea that there are WAY more hypothetical people that could have existed than ones that actually do exist. It would be nearly infinite, while our population is less than eight billion (extreme estimating because I can't be bothered).
So let's pick a target! You want to pick yourself, but that's getting into this whole big thing. So let's use a random number generator. Random.org is a good one, though they can't go higher than 1,000,000,000 which is a LONG way from "almost infinite". Still, that just skews the odds in your favor so let's do that. Let's say if I get a 1 it's a person who exists, and if I get anything else it's someone who doesn't. I'll run it 100 times.
Okay. Done. I could post the list here but it's big and ugly and I think people would understandably object so let's just cut to the chase: none of them exist. So then we plug them into your formula and it says...
1. The odds of this specific person existing are really really low under materialism.
2. And, in fact, they don't exist.
3. Therefore all is right with the world and consistent with materialism.
Like I said, I ran it a hundred times* (one click, but a hundred random numbers generated). This is what it looks like to pick a target fairly. So good news, reality is consistent with our expectations and we can be done here.
(*EDIT: I noticed this "again" button at the bottom and clicked it twenty times. So that's 2100 tests with odds WAY better than you were figuring.)