[EM] RE : Randomized MCA, new weird voting method idea

Warren Smith wds at math.temple.edu
Thu Feb 15 15:46:46 PST 2007


>Venzke:  agree with your conclusion, that 3-slot median rating would not see much
advantage over Approval. I don't think I understand your reasoning though.

I don't think the value of X is especially useful information to the
strategic voter. I can see a value of 50% as possibly useful

--WDS:
it isn't that the particular value of X is useful info.  It is that with any FIXED
value of X, it will be 99.9999% certain that (with some very large #voters who behave
in some probabilsitically-selected way) the election will end in round R,
where R is 1 or R is 2, I don't care which, all I care about is R is the same
every time.  This is just the law of large numbers.

Now if everbody knows the first round is the only one that matters, then everbody
will vote approval-style.  

BUT if the value of X is unpredictable that may no longer be true.

Frankly, though, I do not understand how to analyse this.
Or how the voters would then behave. Or what.



More information about the Election-Methods mailing list