What is always tricky about any attempt to use statistical prediction on something rare like criminality is that you have two things that are both true: the vas

What is always tricky about any attempt to use statistical prediction on something rare like criminality is that you have two things that are both true: the vast majority of crimes are committed by men, AND the vast majority of men are not criminals. So what do you do with this information? If a crime is committed, you can guess with some accuracy it was likely committed by a man, but you can't arrest a random man and think he is likely the criminal. Any attempts to preemptively limit male behavior is going to affect a lot more innocent men than guilty men.