[EM] possible improved IRV method

[Juho Laatu]

Hi,

I have some generic comments on demonstrating
vulnerabilities of different voting methods.

There is no problem free method and on this list I
have seen numerous demonstrations of theoretical
vulnerabilities of (probably) all the proposed
methods. In most cases those demonstrations do not
convince me since one can easily generate at least
some kind of (theoretical) demonstrations against any
method. As a result, when talking about the practical
benefits of different voting methods, I'm mainly
interested in vulnerability demonstrations with
following type of characteristics.

They should be described as credible scenarios from
real life. I'd like to see e.g.
- credible description of the candidate setting (e.g.
left wing and right wing candidates + one centrist)
- credible description of what the voters think (e.g.
left wing oriented voter with sympathies to the
centrist)
- number of voters
- how much do the voters have information on the
expected result of the election (e.g. polls)
- are the voters experts on voting methods or just
normal people
- is the election highly controversial or not (i.e.
are the voters willing to do lots of work to get the
outcome that they desire or are they just satisfied
with dropping their vote in the box)
- are people supposed to vote strategically (not
according to their sincere opinion) on their own
initiative and reasoning, or will they get strategic
guidance from the press, from their party openly, or
secretly (i.e. are the planned strategies known by the
others)

In some cases we might have also some additional info
available:
- will the votes be public or secret
- is it possible to see after the election if the
strategy was used or not
- how does the community feel about insincere voting
(is there a general trust on everyone voting honestly
or are people afraid that others will use tactics (and
is this the case in real life); are strategic voters
considered clever or immoral)
- will there be future elections and everyday work
that would be influenced by the fact that someone used
some strategy heavily (against some others) (=life
continues after the election and one must pay for what
one did)

Sometimes it may be also useful to mention what the
targets of the election are. I believe at least some
supporters of IRV feel that it is not right if some
small centrist party ("weak candidate") wins the
election and the major ones lose. For this reason IRV
supporters may not like Condorcet methods. Condorcet
methods can be characterized as seeking for a
compromise candidate while IRV favours more the strong
candidates. I consider the Condorcet approach to be
more like a neutral approach. If there are additional
requirements (like favouring the strong parties /
steady first place support / strong supporting party
machinery behind the candidate) they should thus be
mentioned in the assumptions.

In summary, pointing out one theoretical vulnerability
of one method doesn't prove much on how usable that
method is in some real life situation. Description of
a realistic real life situation where the method fails
is a stronger statement on the usability of that
method in the described type of real life settings.

BR, Juho Laatu


> On Jun 29, 2006, at 3:48 , Eric Gorr wrote:
> > Jonathan Lundell wrote:



		
___________________________________________________________ 
All new Yahoo! Mail "The new Interface is stunning in its simplicity and ease of use." - PC Magazine
http://uk.docs.yahoo.com/nowyoucan.html