https://electionscience.org/research-hub/the-limits-of-ranked-choice-voting
https://en.wikipedia.org/wiki/Negative_responsiveness_paradox#Specific_examples
We’ve seen it happen in actual elections, as shown in the Wikipedia link.
RCV is just a flawed system, which is expected for something created by a couple of guys 150+ years ago.
The first article is from someone who wants to save RCV, despite that one flaw that they’ve drilled into.
The problem is that it’s a known attack vector, the Wikipedia article talks about how it was used intentionally by a political party in 2005 in Germany to effectively steal an additional seat in their parliament.
My second link is a deeper dive into more of RCV’s many flaws. Because why stop at monotonicity? Seriously, the fact that increasing support can cause a candidate to lose, and not just lose but elect the worst choice, is insane.
That fact that there are more flaws, just as game breaking, means we should all follow the example of the Marquis de Condorcet, the guy who invented RCV, abandoned it because he saw how broken it was.
Then you have the lying liars at FairVote saying that the Condorcet criterion doesn’t matter in elections.
The Condorcet criterion is that if you were to hold a series of one on one elections between all candidates, the winner of those should be the same winner of your election system. RCV fails this in most elections, which is why Condorcet abandoned it.
It wasn’t until about 30 years after Condorcet’s death that an Englishman revived the voting method, but added a proportional twist. It still had all the flaws that Condorcet wrote about, but Condorcet was French, and lost the political games of the French Revolution, so he was mostly ignored.
As a side note, the political writings of Condorcet should be required reading. The guy wrote this in 1790