We formulate the interaction between the communicating nodes and an adversary within a game-theoretic context. We show that earlier information-theoretic capacity results for a jammed channel correspond to a pure Nash Equilibrium (NE). However, when both players are allowed to randomize their actions (i.e., coding rate and jamming power) new mixed Nash equilibria appear with surprising properties. We show the existence of a threshold such that if the jammer average power exceeds, the channel capacity at the NE is the same as if the jammer was using its maximum allowable power all the time. This indicates that randomization significantly advantages powerful jammers. We also show how the NE strategies can be derived and we provide very simple approximations to the optimal communication and jamming strategies. Such strategies are very simple to implement in current hardware and software.