Optimal QRAC strategies and an extension of the MUB condition