[Election-Methods] Fwd: FYI - FairVote MN Responds to Lawsuit Against IRV

[Juho Laatu]

Dave already explained the use of the equal sign. I was just lazy to  
invent any specific order in my example ballots so I used simple  
ties. Voters are free to list all of them in their order of preference.


Condorcet methods are good methods in such sense that they are not  
very vulnerable to strategic voting and on the other hand they allow  
voters to give quite a lot of information on their preferences (and  
all this info will be taken into account when determining the  
winner). In some other methods ballots carry more information (like  
numerical candidate ratings) but they have more difficulties with  
strategies.

Condorcet methods are not very widely used today. Partly this may be  
due to the higher complexity of Condorcet methods. They are  
relatively tedious in large scale with manual counting. But computers  
can do the calculations easily. One just needs to collect locally all  
pairwise comparison results (how many voted X>Y etc.) and then send  
this data upwards.

Another reason why Condorcet methods are not used in wide scale today  
is that there are quite number of them and it is not quite obvious  
which one to take into use. Actually all of them may be good enough  
when compared to many methods in use today but selecting one can be  
difficult. Experts also have different opinions on which of the  
"Condorcet completion methods" is the ultimate best. All Condorcet  
methods elect the same winner when one of the candidates beats all  
the others (=the Condorcet criterion). But Condorcet methods have the  
interesting property that in some situations A may be preferred to B,  
B to C and C to A. This is where different Condorcet completion  
methods differ from each others.

Third problem of Condorcet methods is that there are some nasty  
strategies that may be used. It is however quite difficult to apply  
them, especially in typical large public elections where voters make  
independent decisions. And if someone tries to apply strategies they  
easily fail and may rather hurt the strategists. In most practical  
situations the voters can just forget the strategies this and rank  
the candidates sincerely. Condorcet methods are thus relatively  
strategy resistant although some strategic voting scenarios exist (in  
the most commonly iscussed one voters can try to introduce an  
artificial loop in the opinions and thereby fool the system).

I listed some of the problems of Condorcet methods above, but in  
general Condorcet methods are really good general purpose single  
winner methods for typical political elections.

There are also Condorcet methods that have been enriched with  
additional information like approval cutoffs, numeric ratings and  
preference strengths. These enhancements are an interesting area of  
study but in most cases the basic (pure ranking based) Condorcet  
methods are good enough and already provide most of the benefits and  
avoid strategic voting related risks (and are simpler).

Juho


On Dec 27, 2007, at 5:02 , Kathy Dopp wrote:

> On Dec 26, 2007 6:53 PM, Juho Laatu <juho.laatu at gmail.com> wrote:
>>
>> Condocet methods do not put any additional weight on first position
>> on the ballot. Vote Gore>Bush>Paul>Nader is considered to be equally
>> strong in saying "Gore is better than Bush" than e.g. vote
>> Nader>Paul>Gore>Bush.
>
> Juho,
>
> Thanks for clarifying.  I understand better the method now.  So in
> Condocet, if you really dislike a particular candidate, it is best as
> a voter, to list all the other candidates in order of preference -
> except for the one you might dislike the most?
>
>>
>> With votes
>> 25: A>E>B=C=D
>> 25: B>E>A=C=D
>> 25: C>E>A=B=D
>> 25: D>E>A=B=C
>
> I do not get the "=" signs.  Do you mean that voters are limited to
> listing two candidates in ranked order and that it does not really
> matter what they list as their third choice since all third choice
> candidates are equal?
>
>> Condorcet methods elect E (since E would win any of the others 75-25
>> in a pairwise comparison). E didn't have a single first place
>> supporter but many obviously considered E to be a good compromise. Is
>> this ok to you?
>
> Yes. I think this Condocet method actually gives a reason for using
> ranking with multiple candidates.  I think IRV is awful, but this
> seems to be OK.
>
>>
>> Condorcet methods simply collect the pairwise preferences from the
>> ballots and base the decision on that data (without any potentially
>> unfair elimination rounds).
>
> Yes. This is far fairer and makes more sense to me than IRV.
>
>> Putting more weight on the first
>> preferences is not used, mainly since it would then be more
>> problematic to keep the method sufficiently strategy free (=voters
>> can now quite safely mark their sincere preferences on the ballot).
>
> Yes that does seem true - although I have not sat down to really
> ponder and study it because I'm working on other things like achieving
> verifiably accurate vote counts which I believe are more crucial first
> steps.
>
> It is very important IMO that voters can actually mark their sincere
> preferences without having to strategize and hypothesize on what other
> voters may do to overcome the flaws of the system like is necessary
> with IRV.
>
> Kathy
>
>>
>> Juho
>>
>>
>>
>
>
>
> -- 
>
> Kathy Dopp
>
> The material expressed herein is the informed  product of the author
> Kathy Dopp's fact-finding and investigative efforts. Dopp is a
> Mathematician, Expert in election audit mathematics and procedures; in
> exit poll discrepancy analysis; and can be reached at
>
> P.O. Box 680192
> Park City, UT 84068
> phone 435-658-4657
>
> http://utahcountvotes.org
> http://electionmathematics.org
> http://electionarchive.org
>
> History of Confidence Election Auditing Development & Overview of
> Election Auditing Fundamentals
> http://electionarchive.org/ucvAnalysis/US/paper-audits/History-of- 
> Election-Auditing-Development.pdf
>
> Vote Yes on HR811 and S2295
> http://electionmathematics.org/VoteYesHR811.pdf
>
> Voters Have Reason to Worry
> http://utahcountvotes.org/UT/UtahCountVotes-ThadHall-Response.pdf
>
> "Enlighten the people generally, and tyranny and oppressions of body
> and mind will vanish like evil spirits at the dawn of day," wrote
> Thomas Jefferson in 1816
> ----
> Election-Methods mailing list - see http://electorama.com/em for  
> list info