[EM] More CWs who maximize SU with more than 1 issue-dimension

MIKE OSSIPOFF nkklrp at hotmail.com
Tue Jan 27 00:07:01 PST 2004

Yesterday I defined a kind of radial symmetry, in which the voters' 
population density distribution is the same along every ray leading from the 
center of the distribution.

But it seems to me that I've heard a weaker definition of radial symmetry: 
It requires only that for every ray leading from the center, there's another 
ray leading from the center in the opposite direction, along which the 
population density distribution is the same as on the original ray.

It seems to me that if voters have even that weaker radial symmetry, the CW 
is always the SU maximizer.

I'd said that maybe, if there's a point that's median with respect to all 
the space's dimensions, the CW is SU maximizer.

What I meant by "median" there is: Say the issue space has N dimensions. An 
(N-1)-dimensional section through the issue-space--I'll just call that a 
section. If a section divides the voters into 2 equal parts, I'll call that 
an equal section. If an equal section is made perpendicular to each of the 
rectangular co-ordinate axes, and if they all intersect at a point, I call 
that point the "median point".

So I'd said that maybe if there's a median point, then the CW will be SU 
maximizer. I don't think that there's a CW at all if there isn't a median 
point, when we're looking at it in terms of rectangjular co-ordinates, so 
I'll drop the wording "If there's a median point".

I found an example in which, when Euclidian distance is used, the CW seems 
to not be SU maximizer.

But if city-block distance is used, then the CW is SU maximizer.

You may prefer city block distance, or you may prefer Euclidian distance. 
But there's no compelling case for one over the other as the only meaningful 
distance measure.

What that means is that there's a meaningful sense in which the CW will 
always be the spatial SU maximizer, even with issue-space with more than 1 

And I claim that it's unfair to invoke non-spatial disutility, because, as I 
was saying, there's no reason to believe that, in a wv election, an 
unpopular, unknown, or non-committal candidate would have the voter-median 
position all to himself.

The claims in this posting are from a brief glance at the subject. 
Conclusions from a brief glance are sometimes incorrect, but these seem 
right. If you look it up, you'll probably find that these statements are 
correct. Maybe some here can comment on the accuracy of these statements. If 
one or more of them isn't correct, then I'd be glad for someone to mention 

Mike Ossipoff

High-speed users—be more efficient online with the new MSN Premium Internet 
Software. http://join.msn.com/?pgmarket=en-us&page=byoa/prem&ST=1

More information about the Election-Methods mailing list