[EM] Condorcet basics

[Dave Ketchum]

Agreed that there can be variations among what is basically Condorcet, BUT:
      Voter's ranking of candidates only shows for A>B that A is 
preferred, over B, but not by how much.
      Ranking A=B shows equal liking, and two voters ranking A=B and A=B 
have same effect as one voting A>B and the other A<B.
      Truncation is permitted and truncating both A and B does not get the 
counting that explicit A=B would provide (truncating B while keeping A is 
an implicit ranking of A>B).
      All of the above only produces a matrix showing counts for A>B, A<B, 
etc.  Matters not whether all the ballots are counted into a single 
matrix, or into separate matrices with the matrices added together such as 
from separate polling stations or absentee ballots added in later, etc.
      Matters not how the voters indicate the ranking specified above, 
provided the ballot design tries for minimal error and provides for this 
for machine voting as well as for manual voting with ink and paper.

A couple thoughts inspired the above:
      Ballot design is important, but should be a separate project from 
counting methods.
      I claim A=B=C=D should be permitted, and should have the indicated 
value for each pair of such candidates.

On Wed, 6 Oct 2004 15:10:02 -0500 Paul Kislanko wrote per
Subject: RE: [EM] Re: Does MAM use the Copeland method?

 > That example was the first that convinced me that there was a lot of 
work to
 > be done to get one that can be explained to the voting public.

AND - being explainable and understandable is important.
 >
 > For the same set of ballots, pairwise comparisons result in a different
 > winner based upon which voting method is employed. So which one you use 
 > is an article of faith, not reason.
 >
 >
 >>-----Original Message-----
 >>From: election-methods-electorama.com-bounces at electorama.com
 >>[mailto:election-methods-electorama.com-bounces at electorama.com
 >>] On Behalf Of Ted Stern
 >>Sent: Wednesday, October 06, 2004 2:54 PM
 >>To: election-methods at electorama.com
 >>Subject: [EM] Re: Does MAM use the Copeland method?
 >>
 >>On 6 Oct 2004 at 12:13 PDT, Paul Kislanko wrote:
 >>
 >>>Ok, let us just get this straight. I don't give a flying flip about
 >>>any criterion, and I don't have a favorite method.
 >>>
 >>>I merely observe from the original ballots that 5 of 9
 >>>voters prefer C
 >>>over A. So those are the ones who will be unhappy if A is elected.
 >>>
 >>But how unhappy will they be?  Only 3 of those voters
 >>strongly disapproved A.
 >>The 2 voters who wanted C in first place are actually happy
 >>that B didn't win (B wins approval), since B>C.
 >>
That response makes me unhappy.

First, we have no knowledge as to how unhappy which voters may be.

More important, we have a cycle in which any possible choice will result 
in a majority wishing for different choices.

Possible responses to cycles:
       A Runoff?  These do not necessarily please all.
       Make resolving cycles part of the method.
 >>
 >>>That majority will initiate a referndum that changes the
 >>>voting method
 >>>because it selected the "wrong" candidate from the VOTERS
 >>>perspective.

This does not work - it is changing the method because of unhappiness - 
WITHOUT - checking to see if the method moved to can be expected to 
provide less unhappiness next year.
 >>
 >>What makes you think it would pass?  Bucklin and Approval
 >>give the election to B, but the A voters are happy, and the C
 >>voters would rather not end up with B winning, so 2/3 of the
 >>voters would vote to keep Condorcet.
 >>
Again, explaining cycles is the heart of resolution, rather Than changing
method.
 >>
 >>>You academics can say A is right, but if that is so, this just
 >>>demonstrates that Plurality does as well as anything.
 >>>
 >>Just in this case.  Who should win this example? (from Rob LeGrand)
 >>
 >> 98:A>C>E>D>B
 >> 64:B>A>E>C>D
 >> 12:B>A>E>D>C
 >> 98:B>E>A>C>D
 >> 13:B>E>A>D>C
 >>125:B>E>D>A>C
 >>124:C>A>E>D>B
 >> 76:C>E>A>D>B
 >> 21:D>A>B>E>C
 >> 30:D>B>A>E>C
 >> 98:D>B>E>C>A
 >>139:D>C>A>B>E
 >> 23:D>C>B>A>E
 >>
 >>    A   B   C   D   E
 >>A --- 458 461 485 511
 >>B 463 --- 461 312 623
 >>C 460 460 --- 460 460
 >>D 436 609 461 --- 311
 >>E 410 298 461 610 ---
 >>
 >>
 >>B>E : 623
 >>E>D : 610
 >>D>B : 609
 >>A>E : 511
 >>A>D : 485
 >>B>A : 463
 >>A>C, B>C, D>C, E>C : 461
 >>
 >>A wins Beatpath, B wins MAM (and plurality), C wins MinMax, D
 >>wins IRV, E wins Bucklin.
 >>
 >>Ted
 >>--
 >>Send real replies to
 >>	ted stern at u dot washington dot edu
-- 
   davek at clarityconnect.com    people.clarityconnect.com/webpages3/davek
   Dave Ketchum   108 Halstead Ave, Owego, NY  13827-1708   607-687-5026
             Do to no one what you would not want done to you.
                   If you want peace, work for justice.