start with common concerns and then respond to individual reviewer comments as space permits: 2 Common: There should be a baseline using MCTS and assuming access to simulator / common random numbers