(For clarification, Im going to say I reply now before my replies made in
this posting, to distinguish them from replies someone has quoted from
earlier postings).
Warren says:
It [Participation failure] shows voting honestly can hurt you (versus not
voting at all)
in every Condorcet system.
I reply now:
So pity the poor Condorcet voter who is strategically forced to stay home on
election day. <smiley>
When a method offers as much as Condorcet does, there is sure to be some
kind of tradeoff. Something that someone lacking in honesty can seize upon
and try to portray as a major problem. Sorry, but I dont consider
Participation failure a serious problem. It isnt a strategy dilemma. It
doesnt prevent voters to dilute or abandon their genuine preferences. It
doesnt force them to let their favorite lose to a compromise, as RV will
do.
--WDS:
There seems to be the idea in either Ritchie's or other minds
that, if you allow equalities in rankings in a Condorcet voting systems
(and/or, handle them via "winning votes")
then "order reversal" will not be required of a strategic voter.
I reply now:
Falsification of preference is the only offensive strategy that can cause a
problem for wv Condorcet. I refer you to SFC and GSFC. With SFC- complying
methods, if youre part of a majority who prefer the CW to someone else
(without necessarily knowing that s/he is the CW), and if you and your
majority vote sincerely, then that less-liked candidate cant win.
(Ive plausibly defined sincere voting for the purpose of my criteria, and
posted that definition here and at the barnsdle website.
My criteria can be found at:
http://www.barnsdle.demon.co.uk/vote/sing.html
In other words, you and your majority are strategically free to rank
sincerely, under that criterions plausible premise conditions.
Thats why its called the Strategy-Free Criterion. Thats why I say call it
the pinnacle of the promise of rank-balloting.
Warren continues:
I believe that idea is false. I suspect that in every Condorcet system,
whether
rank-equalities are allowed or not, and whether "winning-votes" are used or
not,
there are election situations where you (a voter or co-feeling small bloc
of voters) must cast a vote which
is fully-dishonest about one or more orderings, i.e in which you say A>B
when you honestly feel B>A. If you do not do this in your vote,
then you get a worse election winner.
I reply now:
I have a big surprise for Warren: Ive been pointing out for a long time
that Condorcet wv doesnt strictly meet FBC. No one claims that it does.
But wvs FBC failure is unlikely, and is unlikely to cause a strategy fear
that will intimidate most voters.
Some time ago I was recommending Kevins method that meets FBC, SFC, & SDSC,
because, as I said then, that method has something for everyone--even for
the timid voter who insists on FBC compliance in order to not vote
Compromise over Favorite. I spoke to someone who wanted to vote all the
Democrats over Nader in an Internet BeatpathWinner poll, without believing
that the Democrats honesty or policies were as good as Naders. And I
couldnt assure her that there could never be a benefit in doing that,
because Condorcet doesnt meet FBC. That was what attracted me to the method
that meets FBC, SFC & SDSC.
Since then I feel that the guarantees of SFC, GSFC and SDSC will be
sufficient to reassure most people that they neednt defensively
order-reverse. If someone is interested in strategy enough to consider
defensive order-reversal, then s/he can be expected to look at the
guarantees of SFC, GSFC and SDSC. If s/he is in a majority such as the one
specified in those criteria, then s/he has nothing to worry about and
neednt consider defensive order-reversal.
And if you arent in such a majority, then what can you really expect
anyway? So I prefer SSDs added advantage of GSFC compliance to that other
methods added advantage of FBC compliance.
But I emphasize that if Warren wants FBC, then he neednt look farther than
the method that meets FBC, SFC and SDSC. He neednt settle for the little
that RV offers.
Warren continues:
One way to set up such a situation (which should work against most of
the Condorcet systems discussed on EM) is this.
You honestly feel A>C>the other candidates.
If you do nothing or vote honestly, then C will be the Condorcet winner.
If you vote A>the others>C then C will no longer be
the Condorcet winner allowing A to win.
If you rank the others EQUAL to C then C will still be the Condorcet winner.
I reply now:
Warren has discovered offensive order-reversal. Weve talked at length on EM
about why it isnt a problem for wv Condorcet. I hope it isnt necessary to
repeat all of that here for Warren. Ive said it in recent posts, during the
past few days, or at least during the past week. I shouldnt have to repeat
it again now.
Ossipoff: ...elections in which equal ranking is disallowed. Of course
no one is proposing such a version of Condorcet.
--WDS: Tideman in his 2006 book recommends exactly that.
Just because the most recent and important book in an area recommends
something, does
that mean that Ossipoff should retreat one iota from his stance that "no
one" does?
I reply now:
I meant that no serious voting system reform advocate advocates such a
method. Ive often said here that all or nearly all voting system academics
have their heads up their ass.
Warren quotes past discussion:
WDS: In IEVS, presently, equal rankings are forbidden in rank-order
methods.
M.Ossipoff: which (like Warren's other assumptions) makes the results
meaningless.
WDS: I do not agree I ever made any "assumption" here.
M.O.: If you didn't assume or believe the premises you based your
simulation
on, then even you must not have believed that your simulation's results
would have any relevance to real-world elections. Shall we call them
simulation-premises instead of assumptions then, to avoid any
speculation about what you were thinking? ...if your simulation
is based on counterfactual premises, then its results won't mean anything.
--WDS: I did not make this "assumption."
I reply now:
Perhaps Warren didnt notice that Ive quit saying that he made that
assumption. I now say merely that its a premise of his simulation.
Warren continues:
I did not "base my simulation
on" it. I did not make these "premises." I did not make these
"simulation-premises."
I reply now:
Im sorry, Warren, but you did, according to what you told us on EM, about
your simulation and what went into it. As I said, Garbage in, garbage out.
Warren continues:
It would indeed be good if you avoided
speculation about what I think.
I reply now:
Yes, and Ive quit expressing speculation about what you think about the
real-world-relevance of your simulation premises.
But, according to what you yourself said, your simulation used those
premises.
Mike Ossipoff
----
election-methods mailing list - see http://electorama.com/em for list info
