> So this is the problem with statistics That's not a problem with statistics, it's a problem with using particular statistics to talk about a non-particular g

> So this is the problem with statistics That's not a problem with statistics, it's a problem with using particular statistics to talk about a non-particular group. If you'd adjusted for people who don't break the law, people like you, who get shot on accident by a police officer, the statistics would be perfectly fine to compare to the threat of you dying at the hands of police vs terrorists. There's no problem with statistics in the way you describe them, just with bad statistics just like there's a problem with bad anything (math, engineering, policy etc)