More 0-info pairwise strategy

[MIKE OSSIPOFF]

Demorep said:

>Mr. Schulze wrote-
>
> > Suppose that p(B,A) is the calculated probability that
> > you change the winner from candidate B to candidate A
> > when you vote A > B > C > D sincerely. Suppose that
> > p(B,C) is the calculated probability that you change
> > the winner from candidate B to candidate C when you
> > vote A > B > C > D sincerely. Suppose that u(X) is your
> > von Neumann-Morgenstern utility of candidate X. Then:
> > If p(B,A)*(u(A)-u(B)) < p(B,C)*(u(B)-u(C)), it is
> > advantageous for you to vote A = B > C > D insincerely
> > instead of A > B > C > D sincerely.
>----
>D- Assuming there are polls before the final mark- the- ballots election,

Let me say that that we're here assuming the opposite. We're
assuming that there's _no_ information available about the other
voters, their rankings, their utilities, their intentions, their
party affiliations, etc.

>then with
>
>N1 AB N2
>N3 BC N4
>N5 AC N6
>
>the N numbers must all be rather close for "you" (i.e. John/Mary Q. voter) 
>by
>yourself (without any other devious voters) to change the results sincerely
>or insincerely.

That goes without saying. Whether the method is Plurality, Approval,
Condorcet or Margins :-(, the only way you affect the outcome
is if the other voters are in a tie or near-tie. A pair of same-
voting voters could reverse a pair-comparison or a magnitude
comparison between 2 defeats. So one could stay home on election
day, or take part in case he, or he & his friends, might change
the outcome favorably. If you make the latter choice, then
, when considering strategies, you're only interested in the
closer cases. You're interested in the possibilities where
a certain 2 candidates are the frontrunners in Approval or
Plurality. Or, in Condorcet or Margins you're counting the
ways that a count could be close enough for you to change it.
The difference is that with Condorcet & Margins there are many
more situations to check out.

>
>I assume no one is in favor of abolishing pre-election polls (???) which
>would provide all sorts of information.

True 0-info isn't likely, but it's interesting as a simplifying
case of the more general situation where probability info is
available. Also, there could be some kinds of situations, maybe
not usually in public elections, where 0-info conditions almost
obtain or do obtain.

Also, I caution against taking polls at face value. Pollsters
have been known to falsify the numbers in order to favor a
particular outcome. Place your trust only in polls taken by people
whose political beliefs are similar to yours, or whom you have
good reason to trust the honesty of. The advantage of the
Cycle Condorcet's methods is that they have the least need for
polling information.

Mike Ossipoff





______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com