0 and 1 as probabilities do make sense under the UDT probability-as-measure-of-caring interpretation (as opposed to the Bayesian probability-as-subjective-degree-of-confidence interpretation). In UDT you don’t do Bayesian updating so you don’t run into the divide-by-zero problem with probability 0 events, so you can rule it out as a valid probability on purely mathematical grounds. (EDIT: Obviously I meant to write “can’t” in the last sentence.)
Good point. Also, we can push steven0461′s analogy a bit further by saying indifference is what happens when UDT finds itself in a universe that has prior probability 0 :-)
I wonder if it’s fair to say that indifference isn’t a preference in the same sense that zero and one aren’t probabilities.
0 and 1 as probabilities do make sense under the UDT probability-as-measure-of-caring interpretation (as opposed to the Bayesian probability-as-subjective-degree-of-confidence interpretation). In UDT you don’t do Bayesian updating so you don’t run into the divide-by-zero problem with probability 0 events, so you can rule it out as a valid probability on purely mathematical grounds. (EDIT: Obviously I meant to write “can’t” in the last sentence.)
Good point. Also, we can push steven0461′s analogy a bit further by saying indifference is what happens when UDT finds itself in a universe that has prior probability 0 :-)