Expand a little on a previous quick demo of Bayesian learning to play with different ways of applying the analysis to a gambling scenario. This shows the benefit of moderating the amount you gamble based on confidence estimates, and gives an excuse to make the comment that maybe "breaking even" is underrated.