[EM] Re: questions about sprucing up

Kevin Venzke stepjak at yahoo.fr
Fri Dec 24 21:02:24 PST 2004


Forest,

 --- Forest Simmons <simmonfo at up.edu> a écrit : 
> First let R, G, and B be position vectors for the vertices of a nearly 
> equilateral triangle, say R=[0,0], G=[100%,0], and B=[50%,87%].  Then 
> (with a random number generator) you pick x and y less than 100% at 
> random, calculate z as 100%-x-y, and use your method to find the winner 
> for the election
> 
>     x RGB
>     y GBR
>     z BRG
> 
> Finally color the point P=x*R+y*G+z*B with red, green, or blue, depending 
> on whether the winner was R, G, or B.

Great, I finally understand this! I wouldn't have thought such a trick
was possible.

I played around a lot with this, especially
A, B, C>B
A>B, B>C, C>A
and
A>C, B>C, C.

It's interesting to be able to quantify how often different methods elect
A given the first set of ballot types when A lacks a majority. FPP picks A
most often (largest region), followed by Craig's IFPP, IRV, and then Margins.

For the second types (the standard cycle example), WV agrees with FPP.
IRV rotates this in one direction, and Approval (i.e. approve all non-last)
rotates it in the other direction.

Neat tool. Thanks a lot.

Kevin Venzke



	

	
		
Découvrez le nouveau Yahoo! Mail : 250 Mo d'espace de stockage pour vos mails ! 
Créez votre Yahoo! Mail sur http://fr.mail.yahoo.com/



More information about the Election-Methods mailing list