Phrasing this in terms of two tribes each with precisely one candidate and geographically homogeneous support bakes in some assumptions that I think break the m

Phrasing this in terms of two tribes each with precisely one candidate and geographically homogeneous support bakes in some assumptions that I think break the model quite fundamentally. In the real world: * Rural vs. Urban produces a very strong polarization in party/policy support. So you simply cannot get the result described where you have one party winning all districts. * There are many candidates both within a party and between them that would receive approval from a given voter beyond their first choice. * Currently because of the structure of the system there can only be two serious parties. (A serious third party splits the vote and prevents either from being successful). But if you had more parties you’d probably get better representation. Basically, I think this model is both oversimplifying to the point where it is useless for making claims about what would happen in our current system if we eliminated gerrymandering, and also not interesting as a model for other voting systems. For what it’s worth I think Single Transferrable Vote and Approval Voting are both interesting options that get round most of the problems in our current system.