Bad cases for shaker-sort (Q1114410)

It is shown, by example, that shaker-sort is quadratic in the worst case for the specific increments suggested by its authors, Incerpi and Sedgewick, and that simple variations of these increment sequences do not do significantly better. At the same time, shaker-sort does not seem to be significantly faster than shell-sort for any file size, especially when good increment sequences are used for shell-sort.