I assure you that in my example the problem is totally random - it is my example after all :). In the US it's probably less so. Institutional issues one can deal with, but what to do about "personal biases of the individuals involved (conscious or subconscious)"? We can try to minimize it, but one has to accept that it'll hit diminishing returns after some time.

But my point was about 'rrmm wrote - that "Getting everyone to admit a problem exists and getting our heads around how frequently it occurs is the first step towards resolving it.". It's true, but you have to be 100% totally damn sure that there is a problem in a first place and that you're not inventing it - because in both cases there will be a problem after media get involved.

>you have to be 100% totally damn sure that there is a problem in a first place and that you're not inventing it

Sounds a lot like a 'no true Scotsman' fallacy.

Everything from apartments to restaurants to taxi and semi drivers have ratings or a feedback system and it hasn't brought any of those systems to a grinding halt.

It's not about finding a true Scotsman - it's about allocating resources to fighting real problems instead of imaginary ones.

Ratings systems are not a silver bullet either - they're often easy to game (products) or measure wrong things (restaurants); users can be pressured into giving a good rating (employee interaction rating machines at border control, which you have to press while still in front of the officer) or bosses can use them to control employees (a known not-exactly-a-taxi, semi-illegal company I won't name). Particular failure modes depend on small details about the rating system - whether it's immediate or delayed, anonymous or not, relevant to the other side's career or not, etc. - but there are always some.

Ratings and feedback in general are a good concept, but you can't just take a "SELECT AVG(rate) FROM policeman WHERE rate < 3" an run to the nearest CNN office to talk about how things are very bad. You have to stop and consider all issues of a given implementation of a rating system, and there are quite many of those.

If you don't feel that the training and culture of LEO has some real problems then I'm not sure we have enough common ground to discuss a solution, to be honest.

I do agree that ratings can be gamed and are not a silver bullet though, that is true. It's difficult to get right.

Based on what I read, heard and saw I do totally feel that "the training and culture of LEO has some real problems". Especially in the US, but Europe also starts to go this way. I don't deny the need for addressing it. My original point was about something else - that we should be sure we don't overblow the problem.

The current situation is that citizens and police are stuck in a feedback loop of growing fear and resent towards each other. A key step to solving our problem is to contain this loop before it grows too strong. 'rrmm wrote about giving more visibility to the transgressions of policemen, but I argue that going for maximum publicity isn't always a best approach. Taking into the account how blindly people consume the media, I fear it may only exacerbate the vicious circle of fear and resent.

> We can try to minimize it, but one has to accept that it'll hit diminishing returns after some time.

That's true for everything, but let's not assume that we're at the point of diminished returns already.

Of course. We most likely aren't. But it's important to notice when you pass it (and expect a good solution to pass it at some point in the first place) - something society as a whole generally hasn't internalized yet.