A Min-Max-Max-Min Approach to Solving a Stochastic Programming Problem with Simple Recourse