Abstract: It is our lot to face decisions when we have no certainty about which option from among those available to us will lead to the best outcome. How are we to make such choices? Expected utility theory is the most well-known and widely used account, but there are many alternatives available. Which should you use? Here's a natural approach to this problem: a decision theory is an account of rational means-ends reasoning, so it's natural to assess it by asking how well it performs in the role of getting you the ends that you in fact have. The only problem with this approach is that, in order to assess a decision theory or anything else as a means to your ends, we need an account of which means to your ends it is rational to use. And without a decision theory, we don’t have that. Yet all is not lost, for this line of thinking nonetheless furnishes us with a test we can conduct on a theory of decision-making. We can ask of the theory: If I were to use you not only to make my normal day-to-day decisions, but also to make the higher-order decision about which decision theory to use, would you recommend yourself? If it would, we call it self-recommending. I'll argue it is a necessary but not sufficient condition on an adequate decision theory that it is self-recommending. I show that expected utility theory is self-recommending, but its most popular rivals are not.