Even risk-averters may love risk (Q1611616)

If a player is an expected-utility maximizer, his/her choices will depend on the nature of the game (whether he/she can observe the previous outcomes before making the next decision), his/her utility function, the prior distribution on the bias. In the case of coin tossing, the authors prove that even a risk averter might optimally choose a riskier coin when learning is allowed.