In quicksort, one of the critical operations is choosing the pivot: the element around which the list is partitioned. The simplest pivot selection algorithm is to take the first or the last element of the list as the pivot, causing poor behavior for the case of sorted or nearly-sorted input. Niklaus Wirth's variant uses the middle element to prevent these occurrences, degenerating to O(n²) for contrived sequences. The median-of-3 pivot selection algorithm takes the median of the first, middle, and last elements of the list; however, even though this performs well on many real-world inputs, it is still possible to contrive a median-of-3 killer list that will cause dramatic slowdown of a quicksort based on this pivot selection technique. Such inputs could potentially be exploited by an aggressor, for example by sending such a list to an Internet server for sorting as a denial of service attack.
Musser reported that on a median-of-3 killer sequence of 100,000 elements, introsort's running time was 1/200 that of median-of-3 quicksort. Musser also considered the effect on caches of Sedgewick's delayed small sorting, where small ranges are sorted at the end in a single pass of insertion sort. He reported that it could double the number of cache misses, but that its performance with double-ended queues was significantly better and should be retained for template libraries, in part because the gain in other cases from doing the sorts immediately was not great.
Similarly, Musser also introduced a worst-case linear selection algorithm with time comparable to that of Hoare's algorithm, a simple adaptation of quicksort that is the most efficient selection algorithm used in practice. This is called introspection selection or Introselect.
The June 2000 SGI C++ Standard Template Library stl_algo.h implementation of unstable sort uses the Musser introsort approach with the recursion depth to switch to heapsort passed as a parameter, median-of-3 pivot selection and the Sedgewick final insertion sort pass. The element threshold for switching to the simple insertion sort was 16.