Jan Christian Refsgaard comments on a visual explanation of Bayesian updating