[EM] Classifying 3-cand scenarios. LNHarm methods again.

[robert bristow-johnson]

On Apr 17, 2010, at 9:25 PM, Markus Schulze wrote:

> In my opinion, "Condorcet" refers to a criterion
> rather than to an election method.

actually Markus, i mostly disagree.  "Condorcet", with no other  
qualification (like Schulze or RP) does not *fully* describe a method  
because it doesn't specify how it deals with cycles.  but cycles don't  
always happen, and i would bet that they rarely happen in the real  
world.  the ballot evidence in Burlington in 2006 and 2009 show a  
clear Condorcet ordering of all candidates.

but setting aside for the moment the means of dealing with a cycle (or  
ties), Condorcet *is* a well-defined method that has a ballot  
definition (Ranked, as opposed to Score or Approval or the Traditional  
vote-for-one) and a method of tabulation that is consequently  
different than others of the same ballot such as STV or Borda or  
Bucklin.  it's not a fully defined method, but enough of it *is*  
defined to make a meaningful comparison with existing methods such as  
IRV,  Plurality, or delayed runoff.

i realize that with Schulze or Tideman, the method of tabulation and  
resolution can take place right from the beginning without doing the  
"generic Condorcet" and then applying Schulze or RP in case of a  
cycle.  i realize that.  but without worrying about the cycle, there  
is a method and it is well defined.

no disrespect intended, i think the Ranked Ballot is the correct  
ballot (Score requires too much information from the voter causing  
voter uncertainty in how to mark the ballot, Approval or Traditional  
overly limits contingency information from the voter, again causing  
uncertainty in how to mark the ballot to best support a voter's  
political interest) and a Condorcet-compliant method is the correct  
way to tabulate the ballots.  and among the Condorcet-compliant  
methods, Schulze is likely the best, but it is *not* the most  
transparent for the proletariat and any of these non-traditional  
methods seems to have a problem getting past some persistent ignorance  
(which is something we continue to struggle with in politics) among  
voters.  but *which* Condorcet-compliant method (among the ones that  
are reasonably meaningful) continues to appear to be a bit of ivory- 
tower academic navel gazing.  in my opinion.  some are better than  
others, but it's unlikely to make any difference with any frequency in  
real elections.

just one jaded person's opinion.

--

r b-j                  rbj at audioimagination.com

"Imagination is more important than knowledge."