This paper studies editing procedures based on similarity relations in an expected utility maximization context. It shows that these procedures are compatible both with a family of difference-correlated similarities on the prize space and with a set of families (one for each probability) of...