Or in the parlance of probability theory: "The expectation of the posterior probability, after viewing the evidence, must equal the prior probability."

http://lesswrong.com/lw/ii/conservation_of_expected_evidence...

(Notice the accompanying proof.)