Uniqueness of convex-ranged probabilities and applications to risk measures and games

We revisit Marinacci's uniqueness theorem for convex -ranged probabilities and its applications. Our approach does away with both the countable additivity and the positivity of the charges involved. In the process, we uncover several new equivalent conditions, which lead to a novel set of applications. These include extensions of the classic Fréchet-Hoeffding bounds as well as of the automatic Fatou property of law-invariant functionals. We also generalize existing results of the "collapse to the mean"-type concerning capacities and alpha-MEU preferences.