[EM] Spoilers

Martin Harper mcnh2 at cam.ac.uk
Sun Mar 25 14:52:28 PST 2001


Tony Simmons wrote:

>>> From: Martin Harper <mcnh2 at cam.ac.uk>
>>> Subject: Re: [EM] Spoilers
>> 
>>> Just what exactly is "sincere approval" in
>>> this case? How do I measure it? {genuine
>>> question: I don't want to guess from what
>>> you've written}
>> 
> 
> I just mean that the election method is Approval,
> and the voters are writing on the ballot what
> they actually think of the candidates.

So, you might say that some voter has "sincere approval" of some 
candidate when, in a poll, they truthfully reply to the question "do you 
approve of Fred?" with the answer "yes"?

Problem is, it's entirely possible to find yourself in an election where 
you  "sincerely approve" of all (or, more likely, none) of the 
candidates - it's not clear to me that we should disenfranchise such 
voters by requiring that the candidate with the most "sincere approval" 
should win. In general, while I agree that "sincere approval" is related 
to being the best person to elect, I don't think it is exactly the same.

>>>> Doesn't this sound a lot like the way Plurality
>>>> and IRV suppress support for a third party by
>>>> forcing voters to be insincere?
>>> 
>>> Approval does suppress support for an
>>> *apparent* third party. Plurality and IRV do
>>> this too - but they both also suppress support
>>> for *genuine* third parties as well. Approval
>>> won't do this. In addition Approval only
>>> forces voters to be strategic, not insincere,
>>> which is "nicer" in some way.
>> 
> How is a third party apparent but not genuine?

I meant that it is apparently third, not apparently a party... ;-)

For example, if (flight of fantasy time) Nader had had unanimous first 
choice support, but the international capitalist conspiracy suppressed 
this information and made him appear to have only 10% first place 
support in the polls, then he would be an apparent third party. In an 
approval election being apparently third would suppress the quantity of 
approval he got.

In your example, Road Runner was an apparent third party, (and Mickey 
voters voted accordingly), but (in the measure of "sincere approval") 
was actually a frontrunner. In practice, in large elections, polls are 
good enough to make apparent third parties very rare.

> I take it "insincere" means ranking or otherwise
> evaluating candidates in a way that reverses what
> the voter actually believes?  (just checking to
> be sure I know the difference between insincere
> and strategic).

Basically, yes - there have been debates about this on the list, but 
that seems to be roughly what people mean by the words. Exact 
definitions vary, myself, I go for something like the following:

A "Bad Preference"(BP) occurs when a voter either prefers A to B, but 
votes them equal, or thinks A and B are equal, but votes one over the 
other. A "Very Bad Preference"(VBP) occurs when a voter prefers A to B, 
but votes B as superior to A.

A sincere vote is one where it is not possible to remove a VBP without 
introducing (at least) one more VBP, and it is not possible to remove a 
BP without introducing (at least) one more BP or VBP. A vote with no 
VBPs or BPs is trivially sincere.




More information about the Election-Methods mailing list