[EM] Issues, Condorcet, and IRV (was: IRV vs. plurality)

[Eric Gorr]

At 11:41 AM -0400 8/13/03, James Green-Armytage wrote:
>"Eric Gorr" <eric at ericgorr.net> writes:
>>Let's say, we have something like:
>>
>>03 AB
>>01 C
>>
>>Now, either A or B is going to win.
>>
>>However, there is one last ballot to count. The person casting this
>>ballot would greatly prefer that A wins, but sets the cutoff to
>>include B.
>>
>>A & B still tie.
>>
>>Tie is resolved in B's favor.
>>
>>That voter ended up with a worse result from their point of view by
>>participating and setting the cutoff in the wrong place.
>
>Eric, I don't understand. If the person doesn't vote, then there is a tie
>between A and B. If the person does vote, there is still a tie between A
>and B. I don't understand how the person's vote made the result any worse
>from their point of view.

They wanted to give a clear win to A.
They instead allowed B to win by setting the cutoff in the wrong spot.

>"The participation criterion says that the participation in the election
>by a same-voting group of voters should never worsen (due to the opinion
>of this group) the result of the elections"

Right.

The voter participated and got a worse result because the cutoff was 
in the wrong spot.

>So, if you *delete* the ballots cast by those voters altogether, it
>changes the outcome to something they prefer. This is the test of the
>participation criterion; if this can happen given rational votes, then the
>method fails the criterion.

If you delete the ballot and then determine the winner, A could 
win...which is what they prefer.