I think the Condorcet method is best suited for, say, picking heads of state, where the focus should be on finding the most unifying figure. On the other hand, the purpose of parliamentary election is to represent the full diversity of society, so really there's no single-winner method that would ever satisfy such a goal. That's why we need proportional representation.
Yeah, this is mostly a question about executive positions rather than elections for legislative bodies which would be better served by some form of proportional system. I've been trying to do research on mixed proportional/ranked-choice systems, but that adds a layer of abstraction that makes it even more difficult to research into. That's part of why I wanted to start with the question around single-member executive positions and use that as a starting point to work outward from.
From what I can find, even under proportional systems, there is still room for the Condorcet method to find acceptable middle ground candidates and remove strategic voting. I'm still trying to figure out how to articulate exactly how it would function, but that's taking quite a bit of time and will have to be a longer-term project. I've mostly been leaning on Cambridge, Massachusetts' system here, as they actually used a mixed proportional/Ranked choice system. In this case, if a candidate receives surplus votes above the threshold needed, the balance is given to second place finishers (randomly, rather than weighted, which is its own problem). However, the question around electing inoffensive middle-ground candidates still exists within a mixed Condorcet/proportional systems. In fact, it brings up another secondary concern as well.
Here's the example: there is a highly polarized community with the same partisan balance as before: 45% Democrat, 10% centrist, and 45% Republican. Every Democrat will vote for the centrist over the Republican and vice versa.
One Winner: 1 centrist. The threshold to win is 50%+1. The centrist is elected under the Condorcet method 55% to 45% against either political party.
Two Winners: 1 Democrat;
1 Republican. The threshold to win is ~33%+1. Both parties hit that threshold with ~12% to spare.
Three Winners: 1 Democrat;
1 centrist;
1 Republican. The threshold to win is 25%+1. The balance for the third seat is 20%/10%/20%, with the centrist winning the Condorcet method 30% to 20% either direction.
Four Winners: 2 Democrat;
2 Republican. The threshold to win is 20%+1. Both parties hit that threshold with 5% to spare.
Five Winners: 2 Democrat;
1 centrist;
2 Republican. The threshold to win is ~17%+1. The balance for the fifth seat is ~12%/10%/~12% with the centrist winning the Condorcet method ~22% to ~12% either direction.
Six Winners: 3 Democrat;
3 Republican. The threshold to win is ~14%+1. Both parties hit that threshold with ~2% to spare.
Seven Winners: 3 Democrat;
1 centrist;
3 Republican. The threshold to win is ~13%+1. The balance for the seventh seat is ~8%/10%/~8% with the centrist winning the Condorcet method ~18% to ~8% either direction.
Eight Winners: 4 Democrat;
4 Republican. The threshold to win is ~11%+1. Both parties hit that threshold with ~1% to spare.
Nine Winners: 4 Democrat;
1 centrist;
4 Republican. The threshold to win is 10%+1. The balance for the ninth seat is 5%/10%/5% with the centrist winning the Condorcet method 15% to 5% either direction.
Ten Winners: 4 Democrat;
2 centrist;
4 Republican. The threshold to win is ~9%+1. The balance for the ninth seat is ~9%/10%/~9%. The centrists, finally hitting the thresholds, get the seat. The balance for the tenth seat is ~9%/~1%/~9% with the centrist winning the Condorcet method ~10% to ~9% either direction.
The secondary issue created by this method, at least in this simplified example, is that adding additional seats can lead to prior winners subsequently losing even if the vote breakdown doesn't change. The winning centrist candidate loses every time an even-numbered seat is added until you break double digits. In an ideal system, additional seats being added would not jeopardize a previously winning candidate. But, to me, this is still the system that results in the best outcome of representing the electorate, even if this hypothetical issue can occur.
Cambridge Municipal Elections - City of Cambridge Election Commission
Ranked Choice Voting - Albany City Clerk
