Gains and Losses in Nonadditive Expected Utility

This paper provides a simple approach for deriving cumulative prospect theory. The key axiom is a cumulative dominance axiom which requires that a prospect be judged more attractive if in it greater gains are more likely and greater losses are less likely. In the presence of this cumulative dominance, once a model is satisfied on a "sufficiently rich" domain, then it holds everywhere. This leads to highly transparent results.