Perhaps the best way to district is more about maximum contention in a race? Voroni partitioning so that the maximum number of "purple" districts are achieved makes sense to me.
Hmm, okay, so what would be fair? Perhaps we want to give each party a fighting chance in each election? Isn't that fair?
2 districts each hold 100,000 blues
8 districts hold 50,000 blues and a 50,000 violets
So now the violets have a 50% chance of winning 8 districts. The blues are granted two districts automatically, but that is fair, since they are more popular than the violets. Each party has some chance of winning.
However, in this model, if 1 blue voter in each district changes their mind, then the violets end up with 80% of the legislature. We’ve engineered a system that is very unstable! If Katechon is like the nation of Hungary, where a 2/3rds majority can amend the constitution, then Katechon is now vulnerable to an authoritarian takeover, much like what happened in Hungary after 2010. And that's because 8 people out of a 1,000,000 changed their vote.
You actually remove the possibility of accountability if you are drawing the maps so that a party that behaves badly still has a 50% chance of winning.
But then you're maximizing the number of "loser" voters in each district, who don't feel that their representative in government accurately represents them. Is that the goal of a representative democracy?
Only looks like a problem on a micro scale. If elections are indeed 50:50 on a macro scale, then yes, you would want all purple districts, because the "loser" voters are indeed close to 50%.
If the voters are 70:30 on a macro scale, it wouldn't matter if you create a bunch of purple districts, because they all can't be purple when you have exhausted the 30% of that ratio.
Conversely, if you were to create fixed color districts, you could hide a larger majority as "loser" voters on a macro scale by stuffing them to a few of that color on a micro scale.
But to be clear, I don't think artificially creating purple districts for that express reason is a good idea, I just think it would be less of a problem than the current gerrymandering, when optimizing for democracy.
The problem is that you've fallen into the trap of modern political thinking: that all politics is national, and all that matters is the big decisions that get voted on that affect everyone equally.
And similarly, the false notion that all members of Party A are fungible, and it doesn't matter if I, personally, am represented by someone I hate, so long as there's a member of my party somewhere in congress.
But that's not really the point of a representative republic. Our specific, local needs and desires are supposed to actually be represented by by our members of Congress.
So in terms of actual representation, it would actually be better to have a district that voted nearly unanimously for candidate A, and another district that voted nearly unanimously for candidate B, because then the vast majority of people would be getting the representative that they voted for, and represents them.
In the scenario where every district is purple, fully half the people have no representation of their interests in congress.
Macro scale doesn't have to mean national. There are many precincts per district, and many districts per county, etc.
> So in terms of actual representation, it would actually be better to have a district that voted nearly unanimously for candidate A, and another district that voted nearly unanimously for candidate B, because then the vast majority of people would be getting the representative that they voted for, and represents them.
No, only the throwaway districts in gerrymandered maps are represented, all other districts swallow up many more unrepresented voters.
> In the scenario where every district is purple, fully half the people have no representation of their interests in congress.
But you missed the point, not every district can be purple. In absolute terms, more people would be represented than the status quo. It's better than gerrymandering.
