<p>According to success semantics, a belief is true just in case all actions based on it are guaranteed to be successful. A compelling objection has it that the truth of a belief can never guarantee success, since there may always be an unknown impediment or a chance accident. I argue that we must distinguish the <i>completion</i> (i.e. its full and proper performance) of an action from its <i>success</i> (i.e. it achieving its goal). The proper formulation of success semantics is then that a belief is true just in case all actions engendered by the belief are successful <i>if completed</i>. I conclude that this qualification can help to construct a tenable version of the pragmatist theory of truth.</p>