We consider an optimal control problem for a one-dimensional Itô diffusion and a stochastic game of optimal stopping associated with it. Their value functions satisfy ... and an optimal control defines a saddle point for the game. This extends earlier results to the case of bounded variation...