Skip to content

Comment on A numerical analysis of Quicksort: How many cases are bad cases? [pdf]

Comments

An attempt to solve this problem has been randomization as shown already in Hoare’s first articles on Quicksort [1]. On the other hand, this does not change the statistical probability for bad cases.

I really like the randomization solution. Worst case? What worst case?

Well, the worst case is still there. You're just going to have a very hard time crafting an input that triggers it.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.