Confirmed!: Condorcet efficiency of IRV > 2-stage runoff

[Forest Simmons]

On Fri, 20 Sep 2002, Rob LeGrand wrote:

> Steve asked:
> > Can you provide me with some examples where the IRV method elects
> > the Condorcet Candidate, but the 2-Stage Runoff does not?
>
> Here's one:
>
> 40:A>C>D>B
> 25:B>C>D>A
> 20:C>D>B>A
> 15:D>C>B>A
>
> Plurality picks A, top-two runoff picks B and IRV picks C, the
> Condorcet candidate.
>
Good example.

I assume that the rankings are supposed to be sincere, since there is
nothing particularly nice about an insincere CW.

How would Plurality and TopTwo do under near perfect information with
fully rational voters?

It seems that the faction of 40 would have nothing to lose and something
to gain by voting

40 C>A>D>B in the TopTwo Runoff and
40 C in Plurality,

and that the other factions would know this and consider it either
undesirable (in the case of of the faction of 20 C>D>B>A) or futile (in
the other two factions) to demote their first choices,

so that IRV's only superiority would be under zero or
near zero information in this example.

Forest

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