What is the correct semantics for indicative conditionals, and under what circumstances should agents accept a conditional claim? This paper presents a new case which has important implications for attempts to address these questions. The case involves an utterance of a certain indicative conditional in a particular context. It is shown that at least three prominent theories of conditionals (the material conditional view, the suppositional view, and Stalnaker’s view) predict that you ought to assign a high credence to the conditional in this case, but, it is argued, this prediction is incorrect. Finally, the paper discusses what conclusions we can draw from this case, both on the semantics of conditionals and on the epistemology of inference on the basis of suppositions more generally.