Explaining Condorcet for Single Winner Elections

Mike Ossipoff dfb at bbs.cruzio.com
Sun Apr 7 01:14:08 PST 1996

Demorep's paragraph 6a suggests a good way to avoid criticism based
on the Condorcet Loser Criterion: Right after stating that someone who
beats everyone wins, it's perfectly natural & fitting to state that
anyone who loses to everyone is disqualified. At that point, before
even talkinga about Condorcet's circular tie solution.

Though _any_ added rule could be considered to add some complication
& be some loss of brevity, stating the Condorcet loser disqualification
at that appropriate place makes a lot of sense, if we really have
to add a rule to protect against Condorcet Loser criticism.


But as for the vote transfers, that part is more like MPV. Condorcet's
method doesn't need any vote transfers, since all of a voter's preferences
are counted, and the voter doesn't just have 1 vote, on 1 candidate.
So any method with those vote transfers would have to have a different


Returning to the previous topic, the Condorcet Loser disqualification
rule could be replaced with a Smith Criterion rule, saying that if
there's a group of alternatives that all beat everything outside the
group, then the winner must be chosen from among that group. That
rule would automatically disqualify Condorcet losers, and would also
meet all the other academic criteria, including all of the candidate-
counting criteria that aren't met by plain Condorcet (not that they're
really important, except for show, & for avoiding academic criticism).

If the Smith Criterion rule were used, then the rules would then say
that the alternatives qualifying under that rule shall be chosen
from by Condorcet's "votes-against" rule, which would then follow.
The Condorcet Loser disqualification rule might be simpler to write
into the count rules, since it wouldn't need the clause described in
this paragraph.

Alternatively, instead of that statement of the Smith Criterion,
there could be a statement that the winner shall be chosen from
the smallest group of alternatives that all beat everything outside
that group. (It's important to say "group" instead of "set", for
public acceptance). This would avoid the clause described at the
beginning of the previous paragraph, but could sound more complicated
because it talks about "...the smallest group of alternatives that...".


I'm sure that any of these approaches would be ok. The approach
of just starting with a proposal for plain Condorcet, and then
adding a Smith Criterion or Smith set rule, or a Condorcet Loser
disqualification rule, at such time as the method is challenged
on candidate-countng grounds, would also be a good possibility.




Mike Ossipoff


More information about the Election-Methods mailing list