Abstract : A generalized pivotal quantity is developed that yields exact confidence intervals for the cumulative distribution function (CDF) at a specific value when the underlying distribution is assumed to be normal. This problem is similar to the development of a tolerance interval and unsurprisingly, its solution involves the non-central "t" distribution. Confidence bands for a normal CDF follow easily. Military applications include vulnerability and lethality assessment (e.g., interval estimation for the probability of non-perforation against homogeneous armored targets).