On rationally confident beliefs and rational overconfidence