Since the problem of energy forecasting in SHPPs without storage capacity is little explored, we apply simple schemes to investigate: (a) the essential model structure and information; (b) the advantages of first using a flow forecasting model, instead of a direct energy forecasting; (c) the importance of rainfall data, as a proxy of the catchment state; (d) the training procedure and performance measure; and (e) the representation of predictive uncertainty and its practical interpretation.