[EM] Updated article

[Paul Kislanko]

Again, I'm just a po' li'l ol' voter here. But there was a resonation with
the "need to have accurate polling data".
 
Neither I nor any other voter in my precinct gives a hoot or assigns any
credibility to the "have to have something to publish so we'll publish the
results of a badly-conducted poll" garbage that we're currently getting.
 
Any analysis that depends upon "accurate polling data" is probably doomed. I
was the happy recipient of a random poll call not long ago, and even the
nice-sounding part-time intern who was asking the questions agreed with me
that the questions were pretty stupid. There was no way they could get my
opinion from my answers to those questions. 
 
Let's pretend for the purposes of Election Methods that the only "polls" in
the debate are those that said "Dewey beats Truman!"
 
 


  _____  

From: election-methods-bounces at electorama.com
[mailto:election-methods-bounces at electorama.com] On Behalf Of rob brown
Sent: Saturday, December 24, 2005 6:02 PM
To: Jan Kok
Cc: election-methods at electorama.com
Subject: Re: [EM] Updated article


Thanks for your feedback Jan, comments below....

On 12/23/05, Jan Kok <jan.kok.5y at gmail.com> wrote: 


Rob,

Overall, I like the previous, short version and the current version
about equally.  The short version spent a larger fraction of the text 
talking about the problems with Plurality voting.  The problems with
Plurality are the common ground that pretty much all voting reformers
(IRVists, Approvalists, Condorcetists...) agree on.  Those problems
are also the most likely "hook" to get ordinary citizens and political 
activists interested in voting reform.  So, I like to put a lot of
emphasis on the problems with Plurality when I promote voting reform.


I tend to agree.  I think people very quickly lose interest in voting reform
when they see that those who advocate reform are so factionalized
themselves. 

At least I pointed out that (in my opinion) DSV-approval and the various
Condorcet methods are pretty much equal.  I see no reason to dwell on their
various imperfections in documents intended for the masses.  I continue to
have the opinion that the "cycle resolution" imperfections of Condorcet and
DSV methods are academic and would have few real world consequences if we
just agreed on one method or another.  In other words, here is my vote: 

[X] DSV- Approval Strategy A
[X] Condorcet- Shultz
[X] Condorcet - Ranked Pairs
[X] Condorcet - Minmax
[X] Condorcet - R Brown scoring algorithm
[ ] IRV
[ ] Approval
[ ] Range
[ ] Plurality
[ ] Dictatorship


(and yes I am aware of the irony of unapproving Approval....) :)



It's a pity that Approval isn't your favorite method.  If it was, you
could keep the article pretty short. :-)

You say "If the voter guesses wrong, there is a good chance that
voting may actually do more harm than good to his own cause."  I am 
afraid the casual reader won't understand what you are talking about,
unless you reword this to be more descriptive, or show an example.  


Good suggestion.  See my update.  http://karmatics.com/voting/movienite.html
The permanent url of this revision is
http://karmatics.com/voting/movienite-r3.html (with r1 and r2 being the
older ones).  New stuff is in green, ommitted stuff in red.  

I think this explains the problem pretty clearly. 



I don't see any way to avoid explaining the problem clearly, because 
that problem is the justification for allowing voters to cahnge their
votes, and for introducing ranked ballot methods.  I assume you are
talking about the problem that if a voter underestimates the strength
of his Favorite candidate (relative to a Disliked candidate who seems
to have a chance of winning) and votes for a Compromise candidate in
addition to his Favorite (in order to keep Disliked from winning),
Compromise may win, whereas Favorite might have won if there were not 
so many votes for Compromise.

By the way, allowing people to change their votes in Approval voting
serves the same function as pre-election polls.  If we had Approval
Voting as our public election method, I expect that candidates, 
parties, special interest groups and the media would all want to have
polls that would accurately estimate the results of the election. 


Here is the problem with that as I see it:  it works ok for, say,
presidential elections.  Last time I voted, though, there were a ton
different seperate elections on the ballot, mayor, district supervisor,
sherrif, etc.  I had no idea of who was likely to win....its hard enough to
just learn about all the candidates, much more so if I have to keep track of
who is ahead in the polls. 

Furthermore, this doesn't work in other types of elections like the movie
example I used.  From a practical point of view, this means that we can't
get people used to our "better" election method through things like web
polls, or in voting for the officers of small organizations, etc.  By having
the expectation that people will have accurate polling info served to them,
you have limited the usefulness of the method.  I find that extremely
unfortunate, and unnecessary as well.  The DSV method I proposed seems to
have all the advantages of Approval, without the "need to know polling info
and strategy" downside. (of course, I think Condorcet methods do as well,
but with DSV Approval, it is easier to demonstrate that it is so because of
the way it is derived from Approval) 


-rob 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20051224/76de27e7/attachment-0003.htm>