> worst case, not the average Yup, I can't remember which book but I'm pretty sure I remember reading some stuff from Michael Abrash about how some of the algo

> worst case, not the average Yup, I can't remember which book but I'm pretty sure I remember reading some stuff from Michael Abrash about how some of the algorithms in quake were changed for worse average performance but better worst case performance. It's fairly intuitive when you think about the context... and that is the critical point - abstract problems are interesting, but it's important to evaluate them in their full context.