[EM] DH3 pathology, margins, and winning votes

Warren Smith wds at math.temple.edu
Sat Aug 26 18:09:26 PDT 2006


yes, this is (as Chris Benham said) the same as truncating,
at least as "truncating" is usually interpreted.

By the way, note that IRV is immune to the DH3 pathology while
Condorcet methods get hurt by it.   So DH3 is a good reason to
prefer IRV over Condorcet methods.   However, there are also
reasons to prefer Condorcet over IRV, such as, monotonicity,
Condorcet winners get elected, and IRV can suffer from
the winner=loser (reversal failure) pathology whereas some
Condrocet methods such as Schulze are immune to it.
So then you are saying that you are unhappy for good reason with
Condorcet and also unhappy for good reason with IRV...  at which
point you realize that Range Voting is the cat's meow.  :)



More information about the Election-Methods mailing list