“Doesn’t exist, or doesn’t give a fuck about suffering” is the answer that matches the data
I agree with you. (Though I might rephrase the second as ‘doesn’t care about suffering the way we do’. Either way, your point is valid.)
My point wasn’t to say ‘doesn’t exist’ is wrong, but that there is more than one possibility. If you or anyone has taken the time to evaluate the possibilities and come to the conclusion that ‘doesn’t exist’ is the more likely / simple / predictive model, then I commend you. That is what rationality is about.
All I ask is the same courtesy as I might be exploring a different set of models than you are.
Ha! Well done. I spent a week making sure my math was right and never thought of this. I agree that updating the truth probability is a better model of this situation, and I can confirm your numbers.
I suppose we could also update each day’s success chance, with some kind of prior balancing updating truth probability vs. success probability. Though by that point we are likely no longer “simplifying”.