1.5M ratings
277k ratings

See, that’s what the app is perfect for.

Sounds perfect Wahhhh, I don’t wanna

Logarithmic Policing

@collapsedsquid

That’s basically just secret police
of a totally ordinary sort
that then indulge in kickbacks and protection rackets

Yeah, I realized I forgot to tag that #half shtpost, as I wasn’t being entirely serious.  It’s a bit of wishful thinking.

The thing is, I don’t think we adequately police the police.  We need cop cops.  Metacops, if you will.

The disutility of a crime can be modeled as the risk of getting caught times the penalty for getting caught.  Thus, a crime with a 5% chance of getting caught, times a 20 year sentence as penalty, is then modeled with an effective disutility of a one year sentence.

However, people tend to time discount and so on, so you might not get much more out of a 40 sentence than you would out of a 20 year sentence.  They’re both “a long time.”  This means that you can get more oomph out of increasing the capture & conviction rate.

Regular people, despite not being police, commit crimes for some reason.  Cops, who we send to increase the chances of catching them, also commit crimes.  However, since they’re the ones that we typically send to stop crimes, it’s harder to deal with this, especially when prosecutors have to work with the same police to prosecute normal cases, and thus can’t afford to anger the larger police force.

So we should have a dedicated force to investigate, police, and prosecute police misconduct.  “Sting operations all the way down,” so to speak. Of course, we’ll need to have someone keep an eye on them, too.

…but won’t that result in endless layers of bureaucracy?

Not necessarily.

Just as we only need some limited factor number of police for a given population size, we only need some limited number of metan-cops to police the cops.

We can model the required number of layers of metan-cops as logcop_ratio(num_cops). With a cop_ratio of 2, for instance, our total number of all metan-cops is roughly the same as our total number of non-meta cops. A more reasonable cop_ratio of 10-20 gets us a more affordable ~5-11% for a city with 10,000 cops.

The purpose of each layer is to increase the uncertainty of successfully getting away with a crime at the layer below through arrest and conviction through multiple means, including informants, patrols, sting operations, reports of suspicious activity, investigations, etc.

This would be in addition to other means, such as introducing randomness to make various forms of corruption more difficult, moving people around to prevent building up loyalty between layers, etc. We want any corrupt personnel to always have to act very carefully and in the face of a great deal of uncertainty, as anyone they are interacting with could be one of our meta-cops.

We’d have to trim some of the other laws before we enact this, though, or else be careful just what policies we’re enforcing. Some laws currently not being enforced should just not exist, and we don’t actually want them enforced.

politics policy flagpost