Asymptotics are the heart of the complexity analysis of algorithms, but their usefulness for software development is limited since by their very nature they ignore constant factors. When constant factors are taken into consideration, some overall bad algorithms may be competitive over a certain range of input size. If this range happens to include all practical cases, the bad algorithm may turn out to be superior to an asymptotically much better one. How to determine this and how to apply this to practical situations is the goal of this chapter.