Mixed Condorcet-Plurality

[Buddha Buck]

At 03:00 PM 04-10-2001 -0500, you wrote:
>Demorep,
>
> > 34 ABC
> > 33 BCA
> > 32 CAB
> > 99
>
>By the reasonable assumption that voters put the most energy into their
>first rank choice, and lower ranks demonstrate cyclic preferences, I'm
>completely willing to abandon majority rule and give victory to A with 34%
>of the vote in your example. This is a good demonstration for me where
>plurality is a reasonable choice.

It doesn't seem to matter to me...

Plurality: A got the most first-place votes, so A is the winner.

IRV: C got the least first-place votes, so is eliminated.  All of C's votes 
transfer to A, so A wins 66:33

Condorcet:  There is no Condorcet Winner, but the list of pairwise defeats 
is (ordered by strength): B>C 67:32, A>B 66:33, C>A 65:34.  Tideman (Ranked 
Pairs), SSD, and most other methods I know of off the top of my head would 
all drop the C>A defeat first, leaving A undefeated, and thus the winner.

Approval: Not Applicable (Not enough information on ranked ballots to use 
Approval)

If all the applicable methods give the same result, I'm uncertain what the 
example is trying to show.  That Plurality sometimes gets it right?

Here's an example where Plurality, Condorcet, and IRV all yield different 
results!

45 ABC
35 CBA
20 BCA

Plurality: A wins
IRV: B is eliminated, and 20 votes transfer to C, C wins
Condorcet: B defeats A 55:45, B defeats C 65:35, B wins

>Tom