Obviously there's some sampling bias in your response. That aside, we can assume that the ratio of crimes to reports remained constant - because there's no rea

Obviously there's some sampling bias in your response. That aside, we can assume that the ratio of crimes to reports remained constant - because there's no reason to think otherwise. So if the ratio of reports to daily riders is now lower, even if there are billions of un-reported crimes as long as the ratio improved it did become safer on a per-ride basis. Note that also you have a roughly 1% lifetime risk of dying in a car accident in the United States and your risk of dying on a train rounds to zero. So on balance, I strongly suspect you're safer even if there's a lot of unreported petty crime. My understanding is that per person-hour, the subway doesn't get much more crime than any other outdoor public space.