Validation and extension of the reward-mountain model