With a continuous random variable, X, it is evident that the sum of the usual upper-tail probability, P(X≥x), and the lower-tail probability, P(X≤x), is 1. However, with a discrete random variable this sum is equal to 1+P(X=x). This suggests that, for example, the calculation of the upper-tail probability should then use mid-P, defined as![mid-P](Images/oree/doc/10.1093/acref/9780199679188.001.0001/acref-9780199679188-math-0414-full.gif)