In that case, the fact that I believe something different than wedrifid is evidence against his point.

Wedrifid's preferences aren't literally zero percent evidence for his view of the universe, but they're probably less than one in a googol. My overall point is pretty clear. The nitpicking is annoying, and it seems to me like it's being done because people don't want to change their beliefs.

Your argument doesn't really make sense. You say that although agents might not know how to evaluate the probability of an action's occurrence, the evidence is still unlikely to be perfectly balanced. But probabilities are subjective, and so the fact that you don't know how to evaluate a piece of evidence indicates that for all functional purposes the evidence is perfectly balanced.