NS,
Probability is a methodoligical naturalistic concept. Using it to talk about supernatural 'evidence' is a category error and 'not even wrong'.
I agree with the logic, but isn’t there a practical problem here – namely that in the real world we must apply probability values to claims of the supernatural? When someone with a $5,000 suit and an ambitious haircut says, “Give me all your money and I will give you the keys to heaven” we have no choice but to work out whether or not to believe him. Is he more probably right or more probably wrong?
It’s not much use him saying, “Ah, but you’re making a category error there”. We have to make our minds up using the only tools we have don’t we?
I suppose the answer to that would be that making our minds up tells you nothing about the truth of the claim, just about whether or not to believe it. Absent any other method to assess the claim though, what else is there but ignosticism (or igleprechaunism)?