[EM] CWO may be worth fighting for

[James Green-Armytage]

Hi Jobst,

	I understand your point. I have been aware of it already, and I think
that it is valid, but I suggest that you are overstating its severity.
	First, I agree that CWO may produce a counterintuitive result in some
situations, as you describe. 
	However, the winner in Smith-efficient methods will always still be a
member of the Smith set. Even in non Smith-efficient methods such as IRV,
the result will tend strongly toward the Smith set as a result of the CWO
mechanism.
	Does CWO add an extra layer of strategic protection to only-ordinal
Condorcet methods? I have suggested that it does. See this posting from
2/12/05:
http://lists.electorama.com/pipermail/election-methods-electorama.com/2005-February/014703.html
	I regard Smith efficiency and resistance to burying as two of the most
important properties of a single winner method for public elections.
Hence, although I don't regard CWO to be ideal, I think that its merits
are likely to outweigh its costs in certain scenarios. 
	In only-ordinal WV methods (including river), I think that CWO's merits
probably outweigh its costs. In margins methods, the benefits are probably
greater, because the burying strategy is more of a problem in margins
methods. In IRV, I think that CWO's merits vastly outweigh its costs,
since it essentially brings Smith-efficiency to the method, while keeping
some of IRV's anti-burying properties.
	When debating the merits of CWO Condorcet, we have to ask ourselves how
often we will see sincere majority rule cycles, and how meaningfully we
can choose any member of the sincere Smith set over another based only on
ordinal information.

my best,
James
http://fc.antioch.edu/~james_green-armytage/voting.htm

_________________
The message above is a reply to the message below:
>Dear James!
>
>I'm very confused about CWO. If I understand it correctly, it will not
>at all help avoiding strategic voting but will rather introduce a new
>and very drastic way to strategically alter the outcome! Assume there's
>a cycle, like
>  x A>B>C
>  y B>C>A
>  z C>A>B
>with x,y,z < n/2, and assume that A wins. Then B would most certainly
>withdraw in order to alter the outcome to C, since, like her supporters,
>B will probably prefer C to A!
>
>So, that makes CWO absolutely unacceptable for me! Only the winner may
>withdraw in my opinion.
>