The Allure of IRV

[Forest Simmons]

I agree with Mike's main point (that we should go after the public, and
not worry too much about converting IRV proponents) as well as his
secondary important point that rank balloting is not the right thing to
take to the public at this point.

By keeping things simple we can use simplicity as an argument in favor of
our proposal and against IRV.  That's the kind of argument that will
appeal to the general public.

At this point I think that one of the strongest arguments against ranked
ballots is that in Australia where ranked ballots are available, most of
the voters vote above the line instead of filling out the ranked ballots.

Before I start arguing that in public, I want to make sure that it is
correct.  I don't remember who first reported that or in exactly what
connection, because I was interested in other things at the time. But now
I want to make sure I get it right, because it could be the most
convincing evidence that the general public doesn't want ranked ballots,
except perhaps vicariously, letting their favorite rank the candidates for
them by proxy.

Craig Layton, are you there?  Correct me.  Is the "above line option" only
there for PR elections?  Is it there for all STV elections, multiple
winner or single winner?

--------------

More on ranked ballots in public elections:

Several folks have suggested that ranked ballots are easy to fill out. For
us yes, but not for the general public. Suppose for example, that there
are ten candidates.  Does a number vote of 3 next to a candidate's name
mean that we consider the candidate to be third from the top or third from
the bottom?  After watching the winter Olympics and seeing that 10 is
perfect and 7 is lousy in figure skating, it would be easy to get
confused.

Suppose that you have 30 candidates to sort.  The most efficient
algorithms require about 150 comparisons.  Simple minded algorithms
require about 900 comparisons.  A disorganized person without a systematic
approach will stop far short of the complete ranking that will give all
465 comparisons.

Approval will (typically) allow the person to give the most important half
of these comparisons, without a strict ranking.

Some have said that truncation takes care of the ranking problem.  But
what if the two front runners are your last two choices (numbers 29 and
30)?

When we have thirty candidates, I believe that the most practical method
is the hybrid Approval/ProxyApproval method:

Use a one bit (two level) ballot.  A bullet vote represents a proxy
designation.  Voting experts that are not happy with any of the thirty
candidates' preference schedules can vote their own Approval ballot, and
it will automatically be counted as such, as long as they do not "bullet"
[which would be an unlikely choice of a voting expert who doesn't like any
candidate's preference schedule].

This can be adapted to PR in various ways, including Demorep's Candidate
Vote Transfer method, as well as ProxyPAV.

Someone might object by saying that 30 rankings are a tiny fraction of the
30 factorial possibilities.  But since the issue space is not likely to
have 30 factorial significant dimensions, most of those 30 factorial
rankings will not be needed.  As long as the number of non-clone
candidates is greater than the effective dimension of the issue space,
there should be a suitable ranking choice (among those offered by the
candidates) for most voters ... surely enough to ameliorate the spoiler
effect to the satisfaction of most voters, if not to the discerning
experts.

If IRV solves the spoiler problem to the satisfaction of IRV proponents,
then this hybrid method should solve it to the satisfaction of most
everybody else.


Forest

On Thu, 25 Apr 2002, MIKE OSSIPOFF wrote:

>
> Adam does have a point, that people to whom IRV has been promoted
> are more likely to insist on rank-balloting. That's especially or
> mostly true of organizational IRV proponents, or people who've
> been completely sold on IRV.
>
> But those people are often or usually impossible to reach. There
> are 2 kinds of IRVists who can't be reached: Organization leaders who
> are completely committed to IRV; and others who are inclined to go
> with what they've been told and won't listen to anything else.
>
> Sure, there are a few people who've heard the IRVists' arguments who
> listen & understand IRV's problems, and how those problems are avoided
> by better methods. But what percentage are they?
>
> The point of all that is that I don't expect that many committed
> IRVists are going to change anyway.
>
> So it seems to me that it isn't a matter of changing the IRVists.
> It's merely a matter of defeating them. And that should be easy
> when we tell people of IRV's problems, and point out that the IRVists
> have been a little less than honest in concealing these facts from
> those whom they promote to.
>
> Monotonicity & Participation, when described in frank terms,
> in terms of the absurd things that IRV will do, will discredit IRV
> even in places where the public has heard its promotional arguments.
>
> If the recent French election has eliminated the natural right
> compromise in the runoff, and jumped to an extreme, that will help
> illustrate IRV's problems too.
>
> So, if we can't change the committed IRVists, but can only try
> to defeat them, then which is the better proposal to offer as
> a counterproposal, Approval or Condorcet? I can't say for sure that
> I'm right, but it seems to me likely that we have a better chance,
> when IRV is discredited, by saying, "Let's forget about their approach,
> and go with something that's really simple, and easy to implement,
> and which is nothing other than Plurality done right." My concern is
> if we offer another rank-count that will just start a longer debate,
> and the IRVists can say "Oh come on, if it's going to be a rank
> method, let's stick with the one we were considering, the one that
> you're familiar with."
>
> Anyway, I'm not saying that I'm sure that I'm right about Approval
> being the stronger proposal against IRV, when reaching the public,
> not the committed IRVists, is the goal. I'm not sure, and of course
> it's something that needs to be discussed, both here in general, and
> in any particular placed where a public proposal is going to be made.
>
> Mike Ossipoff
>
>
> _________________________________________________________________
> MSN Photos is the easiest way to share and print your photos:
> http://photos.msn.com/support/worldwide.aspx
>
> ----
> For more information about this list (subscribe, unsubscribe, FAQ, etc),
> please see http://www.eskimo.com/~robla/em
>
>


----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em