[EM] Where the best Condorcet methods (Schulze, Ranked Pairs, River etc.) differ (Toby Pereira)
tdp201b at yahoo.co.uk
Fri May 29 12:46:42 PDT 2020
Thank you for this response, Forest. I was reminded of this subject again when I re-encountered Jameson Quinn's work on Voter Satisfaction Efficiency the other day. According to his simulations, to be found here https://electionscience.github.io/vse-sim/VSE/ ranked pairs performs quite a bit better than the Schulze method. This surprises me since I wouldn't expect much difference in practice (as I put in the original post of this discussion). I'm not sure if Jameson still reads the stuff on this mailing list, but it would be interesting to know what caused the difference.
On Wednesday, 4 March 2020, 17:45:08 GMT, Forest Simmons <fsimmons at pcc.edu> wrote:
If the Smith set is a cycle of three, then the methods you mention give the same result as long as the defeat strength is measured the same way. (You knew that)
Not all of these satisfy Independence from Pareto Dominated Alternatives. I doubt that would make a difference in any known public election from the past, but all else being equal, it is a difference that could make a difference.
Simplicity of explanation and implementation, along with heuristic appeal, and other selling points may be more important than any other distinction among the methods you mention.
For me the easiest to sell formulation of Schulze is in the form of "beat-path." But that is probably just the mathematician in me appreciating an elegant way of creating a transitive relation with minimal violence to the intransitive relation on which it is based.
Thanks for asking this question!
Election-Methods mailing list - see https://electorama.com/em for list info
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Election-Methods