Evaluating the Stability of Non-Adaptive Trading in Continuous Double Auctions: A Reinforcement Learning Approach