This paper offers a new approach to process capability measurement based on techniques for estimating the fraction nonconforming in the tails of the observed process distribution. The methodology proposed can provide significantly more accurate results and applicability to a much larger class of process distributions than the standard approaches of calculating capability indices based on making parametric assumptions or using standard curve fitting techniques.The tail probability estimation approach to capability analysis can significantly enhance the practitioners understanding of the processes they are attempting to develop and/or control because it provides a more robust and accurate measure of capability. This clearer insight into process performance becomes increasingly important as the allowable fraction nonconforming levels decline.