In this paper we present a new rolling horizon approach for solving stochastic mixed complementarity problems (MCPs). Such a scheme allows for decision-dependent probabilities, endogenous learning and closer realism to energy market MCPs. We also introduce a new concept, the Value of the Rolling Horizon (VoRH) to measure the closeness of different rolling horizon schemes to a perfect foresight benchmark. Lastly, theoretical and numerical results are presented with an application in natural gas markets to demonstrate the value of the proposed approach.