Merit of Iterative Plurality & Bucklin

[Mike Ossipoff]

When the iteration that Steve described for Iterative-Condorcet
is applied to Plurality or Bucklin, it seems to me that the resultling
methods are better than ordinary Bucklin, but not as good as
ordinary Condorcet (By "ordinary", I mean without the iterative
feature added).

And it seems to me that Iterated Plurality & Iterated Bucklin
don't meet the meet-able the criteria that I consider important & have
been rating methods by (GMC, LO2E-1 & LO2E-2).

Maybe this is the place to add that it seems to me that
a likely good rule to add to Iterative Plurality would be to 
stop the count when 1 or more alternatives has a vote
total at least equal to half the number of voters, and
award the election at that time to the alternative with
the most votes (as in IR-1).

I've previously claimed that Bucklin is the best of the
methods that are easily hand-counted in large elections, but
now I believe that Iterative Condorcet & Iterative Bucklin
are the most deserving of that title, of the methods that
I've heard of so far.

Since ordinary Bucklin, itself, is iterative, being a stepwise
Approval, the 2 levels of iteration could make it confusing
to the public, or to organizations to which it is offered.
So Iterative Plurality seems the method to recommend for
an easily hand-counted method for large elections. Not that
there's a need for hand-counting of elections nowadays, even
in organizations. Any organization big enough to have enough
voters & candidates to make hand-counting difficult would
surely have many members with computers, who could do a
computerized count.

So not only is easy hand-counting not relevant to public elecions
anymore, but it probably isn't even a requirement for organizations,
and so the title of "best easily hand-counted method for large
elections" is something with more curiosity interest, or academic
interest only, unless one is recommending to a large organization
that isn't willing to use a computer in its count.

I should add that anything said this early about merits of those
newly defined methods is tentative, but I believe the statements
I've made about that in this letter are correct and could be
demonstrated if necessary.

My reason for posting about the merit of those methods is
that I previously said I didn't know about it, and I didn't
want the matter to remain completely un-commented-on.


Mike

 





--