In Protagoras 352a8-357e2, Socrates proposes to show that knowledge (ἐπιστήμη, 352c3; φρόνησις, 352c7) has the power to guide action even in the presence of passions such as anger, pleasure, pain, love, or fear (352b7-8). On this view, no agent in possession of knowledge of what is best for her in a given situation will fail to act on that knowledge, despite the pressure exerted by competing desires or emotions motivating her to choose an inferior course of action. Socrates’ view is usually taken by commentators to be grossly counterintuitive: humans are thought to act also on non-rational, rather than exclusively on rational desires. In this paper, I aim to show that criticism of Socrates’ argument along these lines is misguided, and to defend the view that, if we understand knowledge in the sense that Socrates proposes in this argument, i.e. as the ‘art of measurement,’ his view that knowledge is stronger than the passions is far more plausible than is generally assumed. In sections I-IV, I revisit the vexed question of the precise nature of the absurdity Socrates attributes to the view of most people that, while they know what is good for them, they fail to act on it because they are overcome by pleasure (355d1-3). Next, in sections V-VI, I discuss Socrates’ alternative explanation for the problem of poor choices, according to which errors in the choice of pleasures and pains are due to lack of knowledge of a specific kind (episteme metretike, 357d3-7). In Section VII, I consider briefly the conclusions Socrates draws from his argument, and finally, in section VIII, I defend his alternative explanation as a viable solution to the problem of poor choices. While Socrates does not develop this position in the Protagoras , I believe it can be shown that, when an agent possesses knowledge of the good irrespective of its temporal proximity to or distance from the moment of action, then it is at least plausible to expect that she will act in accordance with it, even if she has a strong desire to act contrary to her motives.