I agree that they should have related their work to baseline RL performance metrics and problems. I do however appreciate their attempt at solving [somewhat] real world problems. The explanation of the Nash equilibrium problem could have been made more simple, and concrete. A plus about the paper is that it is not too ambitious.