[EM] Re: CIBR examples, and its CC failure

[Juho Laatu]

Hello Ken,

On Jun 10, 2005, at 19:07, Ken Kuhlman wrote:
>  So, CIBR appears to be less than ideal, which stems from the fact 
> that the weakest candidate isn't necessarily eliminated first.

I'm not sure what the negative effect of not eliminating the weakest 
first are. But I just want to point out that the structure that the 
correlation derives from the given votes looks very static to me. One 
could just run the correlation analysis through and not eliminate 
anyone while doing it. The end result would be a tree like structure of 
candidates where different branches represent different 
groupings/ideologies.

(This kind of static information may be more useful to some comparison 
based methods than to the runoff type thinking that CIBR represents. 
But this just reflects the fact that I think that correlation is a 
useful tool also for other methods.)

BR, Juho