> How about something sensible -- take a census of which party each person supports and then you get proportionate representatives from each party above .15% of

> How about something sensible -- take a census of which party each person supports and then you get proportionate representatives from each party above .15% of the population. Elections then just determine who gets the seats allocated to each party. That doesn't solve the issue that there's different directions within a party, and that the choice within a party has significant influence. In the US there's a lot of daylight between e.g. Susan Collins, Rand Paul, and Tom Cotton, all republicans, and they're not along a one dimensional axis. Whether you have 52 Collinses or 52 Cottons, will yield drastically different policies. > That's roughly how it works in Germany. Poof -- suddenly there is no such thing as gerrymandering. (context for non-germans) In Germany, the issue I raise is partially addressed by also by having one party vote and one direct vote for an individual. And the difference between directly elected members, and the popular vote is then made up by creating new seats, allocating seats from a ordered list provided by each party. EDIT: To clarify, normally 50% is directly elected, 50% by share. If and only if there's a party share difference between directly elected and indirectly elected seats, then new seats are created to make up for the difference. But the problem with such a system is that the Party has a lot of directional power, by establishing the order on that list...