If we take any finite interval [0,a], there's a unique uninformed prior of the uniform distribution, and we get a (a-b)/a probability of gaining by switching un

If we take any finite interval [0,a], there's a unique uninformed prior of the uniform distribution, and we get a (a-b)/a probability of gaining by switching under that distribution. If we drag a to infinity, we get with probability 1 that switching is better. Numbers are big yo