[EM] Option to vote simultaneous strips

[Forest Simmons]

Good point! Consider the rules of Gradual Approval amended to say that
by the n_th round the number of strips submitted should be at least n.

By the way, in addition to committee and proxy applications, another
potential application of Gradual Approval and other non-simple methods is
as an underlying method for Lorrie Cranor's DSV (Declared Strategy
Voting). 

I read Lorrie's dissertation recently.  It has a lot of good stuff in it.

The basic idea of DSV applies to all voting methods for which strategy is
an important consideration.  After a discussion of the general
possibilities Lorrie specializes to the case of (lone mark) Plurality as
the underlying method in her case studies, and assumes that all voters
would want to use a decision theoretic strategy that maximized their
expected utilities. 

She comes to a kind of mixed conclusion.  When DSV is based on Plurality
it shares some of the problems of IRV: it is non-monotone, tends to
eliminate a CW that has poor first place showing, etc..  It seems that the
main advantage over IRV is that it is harder to manipulate than IRV. 

The college campus survey that she conducted indicated that most folk (in
that population) wouldn't trust that application of DSV: they would rather
cast their own Plurality ballots in a regular Plurality election.

This somewhat disappointing result seems to have led Lorrie to put DSV on
the back burner; now she devotes her time to her job and family. 

On the positive side, as Lorrie points out, DSV has been used by private
firms for certain kinds of decisions.

She also mentions the need for further study of DSV with other underlying
methods and with other types of strategies, especially the "sophisticated"
strategies of game theory that go beyond the simple strategies of decision
theory used in her thesis.

It seems to me that Gradual Approval is an ideal candidate for an
underlying method. At each round each voter has to make a simple decision
that should be easily amenable to decision theory. The method cannot cycle
because the game is over in n moves if there are n candidates.

Forest 


On Mon, 3 Dec 2001, MIKE OSSIPOFF wrote:

> 
> 
> In Gradual Approval, it seems to me that the voter should have the
> option to simultaneously vote more than one strip. For instance,
> the voter knows immediately that s/he wants to vote for hir favorite,
> and not for hir last choice.
> 
> The Nader voter, in the Nader,Gore,Bush election, has it in hir best
> interest to turn in a "no vote for Bush" strip immediately, with hir
> "vote for Nader" strip. In that way, s/he can possibly convince other
> Nader preferrers that Bush won't win, encouraging them to not vote for
> Gore in a subsequent round.
> 
> Mike Ossipoff
> 
> 
> 
> _________________________________________________________________
> Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp
> 
>