> people not suspected of a crime walking on a public street. But they are suspected of a crime, because of their demographic category. I should have explaine

> people not suspected of a crime walking on a public street. But they are suspected of a crime, because of their demographic category. I should have explained the thought experiment more: Suppose future sociologists identify some identifiable cluster of traits that corresponds with a 99.9% likelihood of committing a mugging on any given day--if you see someone with these traits, it's overwhelmingly likely that they've either just mugged someone, or they're just about to mug someone. With that level of certainty, you're either in favor of unreasonably searching them and seizing their mugging weapons (or some other intrusive measure like constant monitoring), or you're in favor of unnecessary muggings. The reason this is a thought experiment is that no real-world identifiable group is that likely to commit crimes. But there is a measurable likelihood ratio, even if treating everyone equally is more important than it.