The single-agent smash and grab problem take the form Max_xU(x).h(x). The agent selects a target value (how much to grab). The probability that he will receive (get away with) x is h(x), the cumulative probability distribution of the maximum achievable level of x. Generalizations are developed...