In Brazil, the short-term energy price is determined by means of a hydrothermal SDDP-based dispatch model. Such model assumes a series of hypotheses and scenarios that can change over time for reasons like unexpected delays in the construction of new power plants, fuel unavailability, and so on. Thus, it becomes very difficult to obtain a realistic estimation of the future short-term price to be used in optimal contracting models. This work aims to develop a new methodology to assess the willing-to-supply curve of a hydro generation company (or other Gencos with intermittent generation, e.g., wind farms). The model uses a hybrid approach between stochastic and robust optimization in which the physical production is regarded as a random variable, with known distribution, and the prices as an endogenous-stress scenario, function of the optimal contracting strategy and simulated generation scenarios. A case study is presented based on realistic data from the Brazilian power system.