More standards

[Blake Cretney]

On Tue, 20 Oct 1998 20:32:37   DEMOREP1 wrote:
>Mr. Cretney wrote in part-
>
>Definition of Sincere Voting
>
>There should be some standard that makes sure a method
>matches what we consider a definition of a sincere vote.  For example,
>some people advocate Approval and define a sincere vote to mean absolute
>approval.  Others only think it means relative approval.  There should
>be some standard to decide who is right.
>
>4.  Sincere vote matches tabulation
>The best way of judging this I can think of is to ask, "is a sincere
>vote a rational vote assuming no knowledge of how others are voting."
>Of course, the "assuming no knowledge" part is necessary because
>every method can have strategic situations with such knowledge.  But
>since a rational voter votes in a way that best helps his or her
>candidates based on the method of tabulation, this standard implies
>a strong connection between the definition of the sincere vote and
>the method of tabulation.
>---
>D- Things are getting bizarre.  YES/NO is absolute.  Number voting ONLY shows
>relative support.
But how do you justify that position.  Is it just an arbitrary decision
of the method's designer, or does the definition of a sincere vote
have to conform to some standard.

Here's an example of what I mean.  I consider 3 alternatives, the first
two are candidates, the third (C) is whatever happens if no candidate 
gets an absolute approval majority.

Sincere preference
A > B > C
None absolutely approved of.

Imagine that you are this voter.  You don't absolutely approve of any of
the candidates, but you prefer one to the others, and want to avoid
the election-invalidation alternative.  Which is the rational vote, 
based on no knowledge of how others are voting, to vote sincerely, 
or to vote for the one candidate you prefer?  In this example, voting
sincerely will only increase the chance of election-invalidation,
your last choice.  It certainly won't increase the chance of your
first choice winning.

So if rational people will not vote sincerely, even if they have no 
information on which to base a strategy, is it meaningful to use this
definition of a sincere vote?  I don't think so.  I think that if a 
method violates SEC, the voters will come up with their own definition
of a sincere vote themselves, a definition that matches the way votes 
are actually tabulated.

It could be argued, however, that SEC is just another strategy problem.
All methods have some strategy problems, so why should we care about
this one?  Well, unlike other strategy problems, it isn't necessary for
the voter to have any knowledge about how others are voting.  It is
therefore less like a strategy, that the voter must design for each
election, and more like a general rule that the voter can always
apply in every election.



-----== Sent via Deja News, The Discussion Network ==-----
http://www.dejanews.com/  Easy access to 50,000+ discussion forums