This paper proposes a measure of explanatory demand formulated in terms of inaccuracy. The concept of explanation plays a significant role in epistemology and philosophy of science, but it is also known for controversies. In this paper I take a step back and investigate the notion of explanatory demand by way of surprise. The concept of surprise itself resists an easy analysis, but we have a better grip on it because surprise is part of our everyday life. The paper identifies unexpected inaccuracy as the reason for surprise, and relates it to explanation; viz. a surprising outcome calls for an explanation, while a good explanation puts an end to surprise. The measure of explanatory demand is neutral on the substance of explanation, such as uncovering causal mechanism or unifying theories, but this does not make it an empty exercise. Among other things, it points to a solution to the puzzle of asymmetry in explanation. The analysis is also part of a broader shift in focus in epistemic evaluation from probability to accuracy, where a theory receives a high mark for its (estimated) accuracy, or closeness to the truth, even if the probability of its truth is low.