Tiered Reinforcement Learning: Pessimism in the Face of Uncertainty and Constant Regret