This paper deals with a stochastic programming model which complements long range market simulation models such as those currently developed by several gas producing firms in North America. These models generate scenarios concerning the evolution of demand and prices for gas in different market segments. A gas company has to negociate contracts with lengths going from one to twenty years. The stochastic programming model presented in this paper is designed for helping a Marketing research department to assess the risk associated with committing the gas production capacity of the company to these market segments. Different approaches are presented to overcome the difficulties associated with the very large size of the resulting optimization problem.