We are concerned with the average case runtime complexity analysis of a prototypical imperative language pWhile in the spirit of Dijkstra’s Guarded Command Language. This language is endowed with primitives for sampling and probabilistic choice so that randomized algorithms can be expressed. Complexity analysis in this setting is particularly appealing as efficiency is one striking reason why randomized algorithms have been introduced and studied: in many cases, the most efficient algorithm for the problem at hand is randomized.