zlacker

Police patrols should be entirely proactive otherwise they are responses instead of patrols. The only point of police patrols is to have a moving and visible presence in the community.

>>austin+(OP)
There are many cases of reactive police patrols. For example, at a protest, as a response to a people feeling that it could help their community, and so on.

I assure you mathematicians aren't optimizing the "visbile presence in the community". That's not something you need a mathematician for. Mathematicians are optimizing the ability of police to maximize the number of arrests by deciding which areas they should patrol in order to maximize crime.

>>austin+(OP)
Count me as one people would rather the cops just respond when called than cruise around looking for crime. If it's that important someone will call it in. If they want a better response time they can do what EMTs do and forward deploy.

>>sudosy+c1
> I assure you mathematicians aren't optimizing the "visbile presence in the community".

How, is this your line of work?

>>austin+z2
My guess is that it is hard to measure "moving and visual presence in the community" and input it into a machine. Not that it wouldn't be nightmarish to communities that have already been overpoliced if that's what was being maximized. I don't want armed guards exclusively patrolling the streets of poor and minority neighborhoods.

>>throwa+e1
You might be underestimating the prejudice of the people calling things in to the police. Listening to police radio for just a few hours I was astounded at the things people report to the police: 1) A man and young girl walking down a country road... the man happened to be a minority, the girl was white. Police decided to check it out. 2) A man wearing camoflague and carrying a long rifle.. in a country area, during duck hunting season. Police decided to check it out.

>>pessim+36
Not at all. You can measure average density of patrols in various areas at various times, the mean and average distance between patrols, the average response times for different patrols to converge onto a response. You can also poll the community on perceptions of local police, visibility, community satisfaction, and so on. To say this sort of thing is too hard to measure really only means you don’t want the data.

>>austin+z2
This is addressed at the very start of the article. It's about predictive policing.

>>austin+z2
Maybe I'm going out on a limb here, but it doesn't seem to me that you'd need fancy predictive mathematics to get reasonable solutions for community visibility.