This article nicely explains the problem, but is a bit defeatist about the possibility of reforms that could mitigate it. To offer one reform, allow me to present the following:
Suppose in the hypothetical country, the total number of votes cast for Blue is some number "B", and the total for Violet is "V". Further suppose that the number of seats won by Blue is "C" and the seats for Violet is "W". A non-gerrymandered map would be one where B:V is approximately C:W (or at least there is no possible change to the seats result that would make C:W any closer to B:V).
My proposed reform, therefore, is that after an election, if those ratios are not as close as possible, the electoral officers would go down the list of seats that were lost by the under-represented party (from best performance to worst) and flip the result so that the under-represented party gains representation, until the ratio can't be improved any more.
Obviously this would feel hugely unfair to the voters in the districts where the result was flipped, but they should blame the people who drew the unfair map in the first place. Indeed, the incentive would be on both parties to draw districts which were most likely to produce a balanced result that doesn't need any flipping. This method has the added bonus that it would encourage parties to appeal to voters in "safe seats" just as much as "swing seats", since everyone's vote would count towards the B:V ratio.
The maths gets a little more complicated when dealing with more than 2 parties, and with independent candidates who don't belong to any party, but I don't believe these are insurmountable obstacles. Solving the problem of gerrymandering, unrepresentative legislatures, and voter apathy with this one weird trick (that doesn't involve changing the ballot papers or counting process) seems like such a big win that the cost of a few flipped seats has to be an acceptable price to pay.
>My proposed reform, therefore, is that after an election, if those ratios are not as close as possible, the electoral officers would go down the list of seats that were lost by the under-represented party (from best performance to worst) and flip the result so that the under-represented party gains representation, until the ratio can't be improved any more.
This seems like an implementation that would harm the trust in the system. If you're going to bother having representation by geography in a democratic system, it should be based purely on geography, not based on demographics within the geography.
I agree with the author of the post that the geographic element itself is the source of the issue and that there are perhaps other better ways of dealing with the "tyranny of the majority" issue that the founding fathers intended to handle via republicanism.
Or you could just do IRV or multi-member electorates or any number of other things done in other more democratic countries that are at least harder to characterise as stealing votes.
Like I say, my proposal allows the electorate to keep the same ballot papers, the same counting process, and the same number of seats in the legislature.
You're right, though, that if voters were prepared to sacrifice some of that familiarity, they could do much better than the system I proposed (which was somewhat inspired by the multi-member system I mentioned in another comment).
Suppose in the hypothetical country, the total number of votes cast for Blue is some number "B", and the total for Violet is "V". Further suppose that the number of seats won by Blue is "C" and the seats for Violet is "W". A non-gerrymandered map would be one where B:V is approximately C:W (or at least there is no possible change to the seats result that would make C:W any closer to B:V).
My proposed reform, therefore, is that after an election, if those ratios are not as close as possible, the electoral officers would go down the list of seats that were lost by the under-represented party (from best performance to worst) and flip the result so that the under-represented party gains representation, until the ratio can't be improved any more.
Obviously this would feel hugely unfair to the voters in the districts where the result was flipped, but they should blame the people who drew the unfair map in the first place. Indeed, the incentive would be on both parties to draw districts which were most likely to produce a balanced result that doesn't need any flipping. This method has the added bonus that it would encourage parties to appeal to voters in "safe seats" just as much as "swing seats", since everyone's vote would count towards the B:V ratio.
The maths gets a little more complicated when dealing with more than 2 parties, and with independent candidates who don't belong to any party, but I don't believe these are insurmountable obstacles. Solving the problem of gerrymandering, unrepresentative legislatures, and voter apathy with this one weird trick (that doesn't involve changing the ballot papers or counting process) seems like such a big win that the cost of a few flipped seats has to be an acceptable price to pay.