Re: [EM] Majority Winners and 3-Level Approval

[Alex Small]

Stephane wrote:

>Why just stop there?
>3-level approval is composed of
>preferred, acceptable, disapproved
>This can easily be matched to a preferential truncated ballot:
>P1 A2 (and D truncated to manifest my desapproval)
>
>What if there is more than 3 candidates and I would like
>to express my full opinion:
>So it generalizes to:
>P1 Q2 A3 B4 (and D and E truncated to...)
>I rank all of my approved candidates.

I don't follow you?  What are Q and B?  How do you use lower rankings to
decide a winner?

I can see 2 ways to use 4+ levels.  The first is to do just like Bucklin,
going to lower and lower levels until you get a majority, but allow people
to put multiple candidates in each level.  I don't know much about Bucklin,
so I don't know what defects it has, but the greater freedom of putting
multiple candidates in each level should remedy some problems.

The second is to use CR for lower levels.  For example, if we rated
candidates on a 0-10 scale, first you look to see if anybody has a majority
of 10's.  Failing that you look and see who has the highest average
rating.  Or maybe you look for a majority of 9's and 10's, or where ever
you want the "preferred" and "acceptable" cutoffs to be.  (zero would of
course be unacceptable).

Alex

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em