[EM] DH3 pathology, margins, and winning votes

[RLSuter at aol.com]

My objection was to Warren's rhetoric. To say the problem
is a "serious" one is one thing. To say, as Warren did, that
it is a particularly worrisome "pathology" with the potential to
cause "massive destruction" is a very different thing. Also,
lumping Condorcet in with Borda is pretty outrageous. Nearly
everyone agrees that Borda has major vulnerabilities regarding
strategic voting that far outweigh any effects the DH3 problem
may have.

As far as empirical data about actual public elections
in which different voting methods are used, it's true that
there is a lot of data regarding one method, IRV, and
small amounts regarding a few others, but there is none
that I know of regarding any versions of Condorcet and
few if any regarding range voting. Even most of the data
regarding approval voting is with organizational elections,
not public elections.

-Ralph Suter

In a message dated 8/27/06 11:59:17 AM Eastern Daylight Time, 
easmith at beatrice.rutgers.edu writes:

<< RLSuter at aol.com (RLSuter at aol.com) wrote:
 >What is most lacking in this and other discussions
 >on this list about strategic voting is empirical data about
 >how people vote in actual public elections in which different
 >voting methods are used.
 
 See http://www.RangeVoting.org/rangeVborda.html#kiribati for a discussion of
 DH3 in a real-life, public Borda election (in Kiribati, a Pacific Island
 state). (Also noted is a more local example, at
 http://www.RangeVoting.org/rangeVborda.html#bossbulb.) I am not personally
 convinced that Condorcet _with an appropriate completion mechanism, one not
 vulnerable to DH3 pathology (approval being the easiest to implement IMO,
 but I'd be interested in seeing a range voting version of that) in case
 of cycles_ would be vulnerable in any realistic case, but Warren is correct
 to point out the problem as a serious one. >>