[EM] The voter median candidate generalized to multidimensional issue spaces

[Forest Simmons]

On Thu, 9 Jan 2003, Eric Gorr <ericgorr at cox.net> wrote:

> An interesting method. If my understanding is correct, people would
> vote in the same manner in your method as they would with
> Condorcet...i.e. ranking the options.

Any kind of ballot that is more expressive than lone mark plurality could
be used.  Each row of the starting matrix represents one ballot with the
ballot marks, whether grades, yes, no, ranks, ratings, etc. all encoded
numerically.

Note that approval ballots don't fully order the candidates, but (given
more than two factions) the point X in the row space of A will almost
surely have as many distinct entries as there are candidates, because of
the the rotations, etc.


>
> Assuming my understanding is correct, can you provide an example which
> would provide a different winner (or set of winners) between your
> method and Condorcet (using either Beatpath Winner or Ranked-Pairs)?
>

Examples forthcoming ...

Forest

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