[EM] Spoiler Effect on Wikipedia

[Daniel Bishop]

Eric Gorr wrote:

> At 6:04 PM +0000 11/11/04, Paul Crowley wrote:
>
>> On Thu, 11 Nov 2004 12:08:31 -0500, Eric Gorr <eric at ericgorr.net> wrote:
>>
>>> hummm....there appears to be two opposing points of view here.
>>>
>>> Chris B. claims that IIA satisfaction does imply ICC satisfaction.
>>> Markus S. claims that it does not.
>>
>> I'm responsible for the edits to that page that make that claim, but
>> if it's wrong please do fix it. Markus S - I'm very surprised that
>> IIA does not imply ICC, could you give an example? I mean the strong
>> version of IIA, the one that no reasonable system satisfies.
>
> I think you mean the strong version of IIA which no ranked ballot 
> method can satisfy.
>
> Unless I am mistaken, Approval Voting does satisfy IIA

It only meets IIA under the assumption that adding or removing a 
candidate does not affect people's votes on the other candidates. This 
is an unrealistic assumption.

For example, suppose that the votes are:

5: A
4: AB
3: B
1: BC
9: C

Then A has 9 votes, B has 8 votes, and C has 10 votes, so C wins. But 
suppose that A withdraws from the election, and so the votes become

5: (spoiled ballots)
7: B
1: BC (spoiled ballot)
9: C

But I think it's more realistic to expect that the 6 "spoiled ballot" 
voters would instead vote for their choice between B and C, which would 
make the outcome

12: B (the former A, AB, and B voters)
10: C (the former BC and C voters)

A's withdrawal from the election changes the winner from C to B, and 
thus Approval voting fails IIA.