Bayesian comments on The Allais Paradox