In scenario optimization, decisions are made in the light of past situations (or scenarios), and the “risk” associated to a scenario decision refers to the possibility that the decision does not perform as expected in a new case at hand. In the presentation, we will discuss a deep, and universal, link that relates the risk to the “complexity” of the decision, where complexity is a concept associated to the amount of information by which the decision can be constructed. This theoretical result provides a support to the use of inductive methods in decision making problems.