Ars does A/B testing on its headlines: Authors submit two versions, and for the first few hours visitors will randomly see either one. Whichever one draws more clicks becomes the permanent headline.

Interestingly, the URL remains constant, which is why you'll often see a disparity between it and the headline.