A speaker attempts to persuade a listener to accept a request by presenting evidence. A persuasion rule specifies what evidence is persuasive. This paper compares static and dynamic rules. We present a single linear program (i) whose solution corresponds to the listener's optimal dynamic rule and (ii) whose solution with additional integer constraints corresponds to the optimal static rule. We present a condition--foresight--under which the optimal persuasion problem reduces to the classical maximum flow problem. This has various qualitative consequences, including the coincidence of optimal dynamic and static persuasion rules, elimination of the need for randomization, and symmetry of optimal static rules.