The feelings you get about this number are feelings that most any belief system would produce—you’d say the same thing with the same 5% if we replaced Christianity with Catholicism, for example.
So I take all beliefs systems as the reference class for that feeling. Smear that 5% over all belief systems—not just the ones that exist, but the possible ones that are as coherent and likely to arise as Christianity—and you should stop worrying somewhat.
After talking with 20 groups that have a very different worldview, I might think they are all are mistaken, but once in a while, maybe 5% of the time, it would actually be me.
As others have pointed out, probably 5% of the time you are mistaken, but that says nothing about the chance they are mistaken; unless you and them are mutually exclusive truth-tracking belief systems with all the evidence. Better assume you’re perfectly spherical and updating takes place in a vacuum...
My numbers? The measure of worlds where a belief system like Christianity meaningfully alters my behaviour (how I choose to act in lots of situations, not just situations involving Christians) is something like one in ten thousand. As you make more specific claims (e.g. Bible says I’m going to Hell, I should act in accordance with {subset of Bible} ) I penalise by one in ten thousand, times one in the complexity of the statement you made.
To provide a more concrete mechanism for what you suggest:
5% is clearly a non-normalized probability. You can add up all those 5%’s, as well as whatever probability you’d apply to atheism, and then divide 5% by that sum. You now have a normalized probability that fits into a proper distribution.
The feelings you get about this number are feelings that most any belief system would produce—you’d say the same thing with the same 5% if we replaced Christianity with Catholicism, for example.
So I take all beliefs systems as the reference class for that feeling. Smear that 5% over all belief systems—not just the ones that exist, but the possible ones that are as coherent and likely to arise as Christianity—and you should stop worrying somewhat.
As others have pointed out, probably 5% of the time you are mistaken, but that says nothing about the chance they are mistaken; unless you and them are mutually exclusive truth-tracking belief systems with all the evidence. Better assume you’re perfectly spherical and updating takes place in a vacuum...
My numbers? The measure of worlds where a belief system like Christianity meaningfully alters my behaviour (how I choose to act in lots of situations, not just situations involving Christians) is something like one in ten thousand. As you make more specific claims (e.g. Bible says I’m going to Hell, I should act in accordance with {subset of Bible} ) I penalise by one in ten thousand, times one in the complexity of the statement you made.
To provide a more concrete mechanism for what you suggest:
5% is clearly a non-normalized probability. You can add up all those 5%’s, as well as whatever probability you’d apply to atheism, and then divide 5% by that sum. You now have a normalized probability that fits into a proper distribution.
Yes, this is a much clearer and more applicable technique.