Re: Doomsday Example

Hal Finney (
Mon, 24 Aug 1998 13:56:02 -0700

I wonder if it would make sense to approach the question of the appropriate priors via the "Miraculous Universal Distribution", the title of an article in v19, n4 of The Mathematical Intelligencer by Kirchherr et al.

The universal prior sets the probability of an object at one over two to the power of its complexity (entropy).

Given two different universes of different size, you could compute the entropy on the basis of the number of states of various types. I'm not sure if you should include information on their locations, or if that should be ignored.

Ignoring locations, this would make the universe with an extra 100 rocks have greater entropy and therefore be a priori less likely initially. You would also count universes as less likely which had more diverse conditions than ones which were uniform (all rocks for example).