Hacker News Reader

Comment by ghostpepper | Hacker News Reader

This is my first thought as well. If police go to a place, there is more likely to be a recorded instance of a crime investigated in that place, and thus the algorithm will be more likely to send the police there in the future.

The crime map for Vancouver is interesting - I'm glad to see that the tool reported in the article is at least exposed to the public.

https://geodash.vpd.ca/

View on HN

In depends on the type of crime.

A break-and-enter, or a theft from/of a car, or vandalism against property, for example, are going to be recorded regardless of whether there's a police officer within 5 miles of where they occur. (Because people file a police report to get an insurance claim.)

The distribution of public nuisance arrests, or drug-related crimes, or people jumping the turnstiles at the subway, or parking tickets, or speeding tickets, or DUIs, on the other hand, are almost entirely going to be closely correlated with which areas you deploy police to.

In some sense, if you did not send police anywhere at all, you would have no crime to report.

The solution to end crime, then, is to never send the police anywhere.

I'm not sure if you're kidding. Of course police presence leads to escalation and more crime, and many places would be much safer without police.

This is not a good position. What you would get is vigilantism instead as people take matters into their own hands.

... What? This is not at all obvious to me. How exactly are you arriving at this conclusion?

Demonstrably false.

Ah, the Trump method!

Reply on news.ycombinator.com