In their comment on D. Trafimow, M. D. Lee and E. Wagenmakers argued that the requisite probabilities to use in Bayes's theorem can always be found. In the present reply, the author asserts that M. D. Lee and E. Wagenmakers use a problematic assumption and that finding the requisite probabilities is not straightforward. After describing the assumption and the conceptual problems it entails, the author presents some numerical examples to demonstrate that the conceptual problems cause important errors. The author explores some common ground between the original article and the comment by using the ratio form of Bayes's theorem but notes that this procedure is also not without problems.