[EM] a simple, great ISDA-compliant method I've never seen mentioned

[Kristofer Munsterhjelm]

On 4/17/23 19:48, James Faran wrote:
> It is my impression that all elimination methods are not monotonic, 
> though I certainly can't prove that.

Just a nitpick: if the base method passes LIIA, then the elimination 
method is equal to the base method and is monotone if the base method is.

But I suspect that every elimination method based on a positional 
weighted method is nonmonotone, and it shouldn't be too hard to prove. 
I'm not sure if every non-LIIA elimination method is nonmonotone. 
(Possible proof strategy: suppose the base method fails LIIA, passes 
majority, and also has the property that raising A may change the social 
order from A>X>Y to A>Y>X; then if A, X, and Y are in a Condorcet cycle 
(A>X>Y>A) and the base method passes majority, raising A would lead to X 
being eliminated first, after which Y beats A pairwise and wins. Now 
"simply" prove that strong monotonicity - that raising A should never 
disturb X vs Y in the social outcome - is hard if not impossible to 
pass, so that every universal domain method usually considered leads to 
a nonmonotone elimination method.)

-km