Value Iteration with Guessing for Markov Chains and Markov Decision Processes