Suppose you have some evidence E, which strongly supports proposition P, but you come to have high confidence in ~P. Let J be the proposition that you have high confidence in ~P in response to E. This paper addresses the question whether J counts as evidence for you against P. A positive answer to this question implies that beliefs are self-confirming. Intuitively we find this quite implausible, presumably because it involves a kind of bootstrapping. However, I have argued elsewhere for a principle I call Parity, that if a proposition concerning another agent's belief is evidence for you in support of the content of that belief, then so is a proposition concerning your own belief. In other words, accepting any kind of conciliatory position about epistemic disagreement commits us to the view that beliefs are self-confirming. I conjecture that most epistemologists would rather accept that beliefs are self-confirming than reject conciliationism; bootstrapping seems more acceptable than dogged steadfastness in the face of disagreement. This paper argues that the consequence of endorsing self-confirming beliefs is far worse than bootstrapping: it involves an illicit form of double-counting. Given the Parity principle that I establish in other work, the moral of the present paper is that we must give up conciliationism.