102

Your belief makes it real (sh.itjust.works)

submitted 1 month ago* (last edited 1 month ago) by emergencyfood@sh.itjust.works to c/memes@lemmy.ml

26 comments fedilink hide all child comments

Context: https://en.m.wikipedia.org/wiki/2024_Sri_Lankan_presidential_election

you are viewing a single comment's thread
view the rest of the comments

[-] mindbleach@sh.itjust.works 8 points 1 month ago

And what do you think happens next?

Those two parties won't keep slapfighting for an uneven split of 60% of the vote, while this third party takes power with 40% of the vote. They will merge. Or the smaller one will simply vanish, if its voters prefer the bigger loser to the plurality winner.

Even in the US, where plurality has a hideous repeated bottleneck on which two parties can meaningfully exist, we did not always have these two parties.

[-] mindbleach@sh.itjust.works 8 points 1 month ago

... wait, Sri Lanka has ranked ballots! What the fuck? They're not even using Plurality, they're doing RCV!

Ranked Choice is a hot mess on its own, but-- oh for fuck's sake I'll just use the example I always use. Say an election goes like this:

40% vote A > B > C.
35% vote C > B > A.
25% vote B > C > A.

Plurality says A wins, because Plurality sucks. You don't even need a bare majority. You just need everybody else to split up.

RCV says C wins: B has the fewest top votes, so they're eliminated. The race becomes 40% A > C versus 60% C > A. Better... but still wrong, because 65% of people would prefer B > C.

Condorcet methods like Ranked Pairs get that right. They model every runoff: A vs B is 40-60, A vs C is 40-60, B vs C is 65-35. B wins every 1v1 and is obviously the best candidate according to these voters. The supermajority prefers B.

And of course Approval Voting is just letting people check multiple names, and it somehow matches Condorcet results when enough people vote, because you are unique, just like everybody else. Genuinely there is no good reason we're not doing Approval by default.