Towards Improving Reward Design in RL: A Reward Alignment Metric for RL Practitioners