The problem is that you keep considering the self to be a separate entity...
Which he is quite entitled to do when he is talking about P(E|~H). When reckoning that, he can assume souls exist. He can assume there are an infinite number of them waiting in heaven to come down and inhabit the little baby bodies that righteous mommies and daddies make for them, and that when the body dies the soul goes back up to heaven. He can assume whatever mystical or religious claptrap he pleases.
But when reckoning P(E|H), he can't assume anything H doesn't give him. It doesn't matter whether he strongly disbelieves H. It doesn't matter whether he can cajole his critics into "agreeing" that H may not be true. It doesn't matter even whether H
is true. When reckoning the conditional probabilities that forum the likelihood ratio that -- in a statistical inference -- takes you from the priors to the posteriors, he
must reckon each element of the ratio as if its corresponding hypothesis were true.
Not only does H not supply any sort of ineffable "self,"
it explicitly precludes it. P(E|H) cannot be properly evaluated if it has to accept the existence of anything immaterial. A lot of us are trying to point this out, but hitting the outside of the bullseye ring in the clamor to point out that H doesn't include this. The point they're missing, in my opinion, is that Jabba is trying to sneak his notion of a soul into E, not into H.
Via his typical equivocations and handwaving, Jabba remains ambiguous on what E actually is in his formulation. He wants E to be not only the experience of self-awareness, but the notion that it's an entity that has some form of existence separate from the body. He has loaded E with begged elements of his preferred theory in hopes of fooling his critics into accepting it as observed data, not speculated theory, and to force materialism to explain it as data. Materialism does not explain it and cannot, so by loading the E dice as he has, Jabba wants to make H seem so very improbable.
...despite the fact that this would make the self far less likely than the materialistic model which is what H is supposed to be.
That too. Jabba always evades this discussion. He tells us it's "complicated" or that "there's more to it than that," but we never descend into the alleged complication (after five years we should have) and we never get the explanation of what there allegedly is more of. The fact remains that Jabba's hypothesis, requiring both a suitable body and a lucky spirit, can automatically be no more probable than a hypothesis requiring only a suitable body.