[Icehouse] Design Competition?

[Dale Sheldon]

On Thu, 12 Jul 2007, Christopher Hickman wrote:

>> (I realize now that perhaps I wasn't too clear when I said I thought 
>> any Condorcet method was better than IRV, but that you needed a way to 
>> break out of Condorcet "ties" (cycles, really); Ranked Pairs is a 
>> Condorcet sub-method which does that.  Apoligies for any confusion I 
>> caused.)
>
> My understanding was that there is a possibility of a three-way tie, 
> where a given three entrants would each be ranked higher than one of the 
> other two and lower than the other, resulting in an M.C. Escheresque 
> circular paradox, thus requiring some kind of run-off vote.  So I am 
> mistaken and the ranked pairs method can handle this?

Yes.  I'd go into detail, but the Wikipedia article covers it nicely.

http://en.wikipedia.org/wiki/Ranked_Pairs

-- 
Dale Sheldon
dales at digitalfreaks.org