3 ms·

sure. I'm not at all thinking my flaws are insurmountable, but I think the problem is you are unlikely to get enough data to draw many conclusions, at least in

sure. I'm not at all thinking my flaws are insurmountable, but I think the problem is you are unlikely to get enough data to draw many conclusions, at least in places that aren't incredibly dense and crime ridden. I'm thinking of the suburb outside of LA I live in. It's somewhat spread out, and some areas have small amounts of violent crime, while some neighborhoods are incredibly rich and most neighborhoods are gated or have private security etc. Having 3 departments be able to serve such a large land area to ensure that response types are evenly distributed seem challenging. But, maybe in a place like San Francisco (more dense, also more crime) this could work?

0xDEAFBEAD 3y ago
Sure, if an area is low crime, I'd say your policing strategy doesn't matter a ton, assuming your cops don't engage in police brutality or corruption. Probably in a low crime neighborhood, competing departments would focus on friendly interactions with the public and making really sure that none of their officers look bad in viral videos. I know I mentioned randomization, but I don't think uneven distribution is necessarily a problem. Ultimately, a voter who looks at 3 different logos on the ballot is going to think back to interactions they had with the cops, and what their friends tell them about the cops, in order to figure out which logo to vote for. Different voters are going to see different distributions of police presence, but that's probably fine. In the same way people tend to vote for the incumbent candidate, I would guess that voters will tend to vote for the logo they see the most in their neighborhood, assuming that logo seems to be doing an OK job. Another way to approach it is that each competing department could have a different service area, and the service areas grow or shrink every election based on election results. This could be done algorithmically by splitting the city in to precincts and finding an assignment of precincts to departments that approximately conforms to the election results, while also keeping service areas for any given logo contiguous through time and space as much as possible (minimize travel and switching costs). All this stuff could work for schools too btw. It achieves the "skin in the game" aspect of charter schools while reducing the ability/incentive to shuttle problem students elsewhere in order to juice your school's numbers (potentially, depending on implementation).

0xDEAFBEAD 3y ago

Sure, if an area is low crime, I'd say your policing strategy doesn't matter a ton, assuming your cops don't engage in police brutality or corruption. Probably in a low crime neighborhood, competing departments would focus on friendly interactions with the public and making really sure that none of their officers look bad in viral videos. I know I mentioned randomization, but I don't think uneven distribution is necessarily a problem. Ultimately, a voter who looks at 3 different logos on the ballot is going to think back to interactions they had with the cops, and what their friends tell them about the cops, in order to figure out which logo to vote for. Different voters are going to see different distributions of police presence, but that's probably fine. In the same way people tend to vote for the incumbent candidate, I would guess that voters will tend to vote for the logo they see the most in their neighborhood, assuming that logo seems to be doing an OK job. Another way to approach it is that each competing department could have a different service area, and the service areas grow or shrink every election based on election results. This could be done algorithmically by splitting the city in to precincts and finding an assignment of precincts to departments that approximately conforms to the election results, while also keeping service areas for any given logo contiguous through time and space as much as possible (minimize travel and switching costs). All this stuff could work for schools too btw. It achieves the "skin in the game" aspect of charter schools while reducing the ability/incentive to shuttle problem students elsewhere in order to juice your school's numbers (potentially, depending on implementation).