[EM] Improvement to Ranked Robin method

[Kristofer Munsterhjelm]

On 02.04.2022 03:24, Forest Simmons wrote:
> Besides being prone to ties, Copeland also suffers from Clone
> Dependence, which makes it inferior to IRV in that regard.
> 
> Here are two versions of decloned Copeland that are not prone to ties:
> 
> 1.Elect the candidate that, on the greatest number of ballots, pairwise
> defeats the candidate designated "worst."

If I'm not mistaken, Stensholt's three-candidate method BPW can be
phrased this way, and I don't think it's monotone. It elects the
candidate who beats the Plurality winner when there's a cycle.

-km