Joyce and others have claimed that degrees of belief are estimates of truth-values and that the probability axioms are conditions of admissibility for these estimates with respect to a scoring rule penalising inaccuracy. In this paper I argue that the claim that the rules of probability are truth-directed in this way depends on an assumption which is both implausible and lacks any supporting evidence, strongly suggesting that the probability axioms have nothing intrinsically to do with truth-directedness.