In some unfortunate scenarios, whatever credence one adopts generates evidence which undermines the rational adoption of that credence. These seem to create epistemic dilemmas. We consider the possibility that one’s credences are indeterminate and require for rationality that one’s indeterminate credences should follow any indeterminate evidence. We will show that rationality principles like these can always be satisfied. This allows the epistemic dilemma to be avoided, however it mandates indeterminate epistemic states and modifications of the rationality principles to apply to the indeterminate state.