In this talk, I begin by distinguishing three forms of overconfidence: overestimation, overplacement, and overprecision. Of these, overprecision (inflated confidence in the accuracy of our knowledge or predictions) is the most robust and least understood. I document its role in self-assessment, test performance, and macroeconomic forecasts. This leads to an epistemological exploration regarding what it means to be overprecise: the challenge of being both wrong and knowing it; that is, holding beliefs about which one is appropriately skeptical, and well-calibrated about the possibility that one’s beliefs are incorrect. I then offer a new theory to account for the evidence and test some of its novel predictions.