Plotting the distribution first; if the histogram shows a clear deviation from the expected entropy curve, then we’ll have a strong case for a genuine anomaly. Otherwise, I’ll chalk it up to a measurement error and keep my calculator on the desk.