framer
Well-Known Member
The object function of the algorithm would be to minimize the border distance. The map that minimizes the border distance wins. Bias is nonexistent and irrelevant.
If you want to engineer for urban vs rural, black vs white, lgbtq vs straight, whatever, have at it. I think it’s unnecessary election engineering that devolves into gerrymandering no matter how pure your intentions. I’d just like to elect the best candidate.
You are making the assumption that there is a "best" candidate. That is generally in the eye of the beholder. What I want is a candidate that represents the needs of a vast majority of the people in the district. If you don't take communities of interest into account when drawing up districts, you can effectively disenfranchise large swaths of the population.
I mean look at the middle East. After the World Wars, Europeans just arbitrarily drew lines that might have looked good on paper, but locked in conflict for the next 100 years. Too many large groups that were shut out from unified representation.
Probably the best option would be to have a computer spit out 5 models based on a certain formula and have a statewide referendum vote in choosing 1 of the 5 to go with.