Hacker News Reader

That assumes that both your presence and the crime are equally distributed over area. If SF annexed a thousand square miles of crimeless, uninhabited wilderness, that would vastly decrease its crime per square mile, but would it affect how much crime each resident sees?

View on HN

The stat indicates there’s moderately high crime evenly distributed, or very high crime throughout a significant portion of the city, or extraordinarily high crime in smaller isolated areas. It’s the second of those (and OP is right, it’s hard to avoid crime if you live here and can’t confine yourself at all times to a few safe residential areas) but none of those outcomes are good.

Sure, but some of the comparison cities (e.g. Memphis and Albuquerque) have huge areas of undeveloped land within their city limits. There's no good reason to compute a per square mile statistic given the vastly different definitions of what constitutes each city.

Don't know why you're being downvoted -- there are parts of Chicago that are 2x as dense as San Francisco, and parts that are less than 1/10th as dense. The denser parts of Chicago hold about the same number of people as San Francisco.

If you wanted to use this metric, you'd need to compare the incidents per square mile of different neighborhoods or city centers. Just doing incidents/mi^2 for a whole city is nonsense

Reply on news.ycombinator.com