As AI tools become increasingly integrated into everyday life, there is a need to expand resources for K-12 AI education. In this work, we introduce a game-based learning activity focusing on teaching reinforcement learning concepts to middle school students. We present the iterative design of the activity, discussing its initial implementation and subsequent enhancements to improve student learning outcomes. The game-based learning activity was implemented during two summer camps with middle school students. We provide an overview of the two different versions of the activity and analyze key insights drawn from student feedback, collected through pre and post-activity surveys.