Consensus, Condorcet(0), and Condorcet(1/2)

[Mike Ossipoff]

Hugh R. Tobin writes:

[Mike says: I'm going to make some comments before I reply
to specific parts of the copied message]

Tobin talks about "agregate happiness". That's a consideration
that leads to point systems. I've already talked about the
disadvantages of point systems. It isn't useful to compare
these methods by a standard which, to be met, would require
a completely different kind of method, one which one of us
are proposing.

Pretty much all of what Tobin says in this letter, in advocacy
of compulsory or default falsification of people's ballots,
has been answered in my previous replies--and my replies haven't
been answered (unless they're answered in Tobin letters that have
been written more recently than the one that I'm currently
replying to).

So, for the replies to Tobins' statements in this letter that
I'm replying to, I refer you to letters I've posted here 
already. 

Though I shouldn't repeat all those replies, and though it isn't
necessary to repeat replies that I've already posted, there
are a few things that I might add, and maybe even a few brief
statements that could be (even if they needn't be) re-emphasized.
But the bulk of my reply to this letter can be found in my previous
replies to these same statements when Tobin previously made them.

Tobin speaks of how Condorcet, without the compulsory or
default falsification (I'll begin calling it "non-falsified
Condorcet") tempts some voters to vote preferences that they
don't feel, including order-reversed preferences. As I said:
Go for it. Voters who order-reverse in non-falsified Condorcet
will most likely regret it, and are unlikely to repeat it.

Tobin again says that the voter who wants to really cast a vote
against equally disliked lower choices is forced to ask his
friends to agree that some will vote one of them over the other,
and some will vote the opposite, or at least that the voter
will be tempted to order-reverse or falsify his own ballot
if it isn't done for him by the method's rules.

Well what's wrong with that? If the method doesn't do it
for him, of course he can do it for himself, and that's
as it should be. Order-reversal casts twice as strong a
vote-against someone as those half-votes anyway.

Not that that's the only solution to the problem that Tobin
claims exists:

Hey, here's something you've overlooked: Ask your brother to run,
and ask your voters to rank him directly under you. That will
cast a _full_ extra vote against everyone else, instead of
the wimpy half-votes that you seem to like.

Or, optionally, save candidates the trouble of doing that by
allowing the use of NOTB, so a candidate could ask his voters
to rank NOTB directly below him.

Or there could be a fictious "dummy candidate" that voters
could rank where they wanted. Or each candidate or party, etc.,
could designate its own dummy candidate for its voters to make
use of. This doesn't violate the principles of Condorcet's
method, and doesn't violate voters' wishes as does
falsified Condorcet. Better a dummy candidate than a
compulsory or default falsified ballot.

There are plenty of solutions other than compulsorily, or by
default, falsifying people's ballots. Assuming, that is, that
there's even a problem. 

Whether there's a problem depends on what you want. I've talked
about what I want from a voting system, and I've told why
I claim that it's what millions of voters would like, and that
it's what electoral reformers want. You'd give that up
(the goal of getting rid of the lesser-of-2-evils problem)
in order to solve a problem for which there's no evidence of
anyone considering it a problem.

But in any case, even if we grant that it's a problem, I've
told you several ways that it can be dealt with, without
making Condorcet into falsified Condorcet.

Another thing: I've asked you more than once to express what
you want in the form of a criterion, to clarify exactly what
it is, and to facilitate comparison of the crierion or criteria
met by falsified Condorcet,  vs the criteria met by non-falsified
Condorcet.

Re: your new example: One can always show a natural circular
tie, and call it evidence in support of giving the election
to whomever you want to give it to. It isn't at all clear
why the fact that the A voters didn't vote A over B should
mean that B should win. Are you implying that ratio should
be the measure of beaten-ness, instead of votes-against? I've
very thoroughly expressed the reasons why I claim that votes-against
is better than ratio or margin.

Now, I'm going to reply directly to some statements in your
letter, but probably only a few. But after my last such
reply in this copied letter, I'll indicate that there are
no more replies after that point in the letter.

> 
> At 08:53 AM 7/28/96 -0700, Steve Eppley wrote:
> >Hugh T wrote:
> >>Steve E wrote:
> >>>Which of the following is the greater violation of democratic 
> >>>principle in a 2-candidate race:
> >>>  1. Electing candidate A even though A lost to B (46 to 54).
> >>>  2. Electing candidate A even though A lost to B (34 to 46, 
> >>>       with 20 who have no preference between them).
> >-snip-
> >
> >>But in the tiebreak we are comparing the results of at least three
> >>pairwise races.
> >-snip-
> >
> >Sure, I know that.  But there's an important connection.  What is
> >the "largest votes against" tie-break measuring?  What's the meaning 
> >of each pairwise race?
> >
> >The C>A=B voter is already counted as a vote against A in the AC
> >pairing and as a vote against B in the BC pairing.  You seem to
> >be saying that this voter should also be counted in the AB pairing
> >because s/he'd benefit if we do that.  I don't think that's a
> >sufficient reason, and I think it would muck up the meaning of 
> >what's being measured.  
> >
> 
>  I had thought it was agreed that a voter ranking candidates equally should
> at least
> have the option to cast half-votes each way, and the question was whether

Oh no. That wasn't agreed. I agreed that, if the public demands it,
or requires it before they'll accept Condorcet, then obviously
it would be undemocratic, and electorally unwise, not to include
the ballot option to cast false contrary half votes. 

But I've claimed, and still do claim, that it makes no sense
to cast such votes. Any reason that you can give for casting
false contrary half-votes could be served better by other
voting strategies &/or options that I've named.

It, I claim, would be extremely undesirable to officiall endorce
& encourage, preference-falsification by actually listing
it as an option on the ballot.

But I'm for being democratic, and so of course if that were
what voters wanted to do, though it would make no sense, then
I say, if that's the case, then it should be added as an
option. Adding it as an option without that kind of a reason
would be unnecessary & harmful.

> there should
> be multiple choices with explanations of their potential consequences on the
> ballot (and,
> if so, what should be the default), or a fixed rule designed to capture the
> intent of sincere 
> voters' rankings.  If half-votes are not even allowed, then voters who are
> sincerely
> indifferent as between lower-ranked candidates (and who do not have a
> strategy based
> on expectations of which will beat the other) must engage in vote-pairing or
> random 
> selection in order to give full effect to their votes for their first
> choices.  As a candidate, I would
> encourage my supporters to rank all candidates below me randomly if they
> have no preference
> among them or are completely ignorant about the others, so that I will have
> the best chance to 
> win a pairwise tie.  Is it desirable to encourage voting such false
> preferences?   (One could

But earlier in this reply, in the introductory comments, I listed
several better solutions to that alleged problem.

> argue it provides an incentive to become informed and make choices --
> obviously those who 
> favor systems that throw out all truncated ballots think so -- but at least
> in those systems the 
> importance of complete rankings is easily understood by the voters, which is
> not the case in 
> Condorcet-0 -- it subtly robs the truncating voter of a portion of his total
> voting power, in a small
> number of cases,  to elect his first place candidate).

As I said, there are _legitimate_ things that can be done by
candidates & voters wanting to use Condorcet as a points system.
I listed some of those things before I began replying directly
to the copied statements.

> 
> I submit that if we want the pairwise races to reflect sincere preferences,
> so that an apparent
> Condorcet winner really is the candidate who would win against each other
> one, then we should try
> to minimize incentives for voting insincere preferences.  My half-vote
> proposal does
> this far better than the alternative of always counting equal rankings as zero.
What? Your falsified Condorcet forces voters to abandon their
favorite, to protect a compromise from what your method would
otherwise do to it. Your method takes away all of the lesser-of-2-evils
properties of Condorcet's method.

> 
> The other point that I intended to raise, by providing an example showing
> all votes rather than
> responding to the question as you posed it, was that your comparison omits
> information that 
> is relevant to the aggregate unhappiness of the voters with various
> outcomes.  That only 46% 
> voted against a candidate in her only loss is more impressive if the reason
> is that a large number 
> of voters ranked her equally with another at the top of their ballots (or at
> least above all other members of the Smith set), than it is  if a large
> number ranked her equally with another at the bottom of their ballots.
> In the former case it could be that electing either A or B would maximize
> the political utility of the voters

If your standard is the maximization of utility, then you should
be proposing the points assignment methods. True, Condorcet
could be used as one, and voters could do so in the ways that
I've listed, ways that are harmless to Condorcet's valuable
properties.

> who ranked C last, so 46% really reflects the total "unhappiness" from the
> election of A.   In the latter case it seems  that 46% understates aggregate
> unhappiness.  In that case I do not see that we should measure
> aggregate unhappiness  as being less than it would be if the "C" voters who
> ranked A and B
> equally had instead split equally between them. 

It's the points assignment systems whose goal is to measure
aggregate unhappiness.

> 
> Consider this example:
> 
> A=B>C  20
> A>C>B  12
> B>A>C  20
> C>B>A  26
> C>A>B  22
> 
> A beats C 52-48
> B beats A 46-34
> C beats B 60-40
> 
> Isn't there a better case for electing A over C here, than in my earlier
> example where  the
> lower "turnout" between A and B reflected abstention by voters who all
> preferred C to both of them?

Based on what criterion or standard? As I said, you can use
a natural circular tie to justify claiming that anyone should
win.

> 
> (snip)
> 
> >>From an older message:
> >
> >>Consider a slight modification of this scenario: Dole's strategists
> >>get a minority of his supporters, 18 per cent of the total, to
> >>order-reverse (alternatively, they really prefer Nader), the rest
> >>sincerely vote Dole, Clinton.
> >>
> >>28 Dole, Clinton
> >>18 Dole, Nader
> >>20 Clinton
> >>34 Nader, Clinton 
> >
> >With Condorcet(0) the Dole reversers would be burned at the stake if
> >they elected Nader instead of Clinton, so reversal isn't a plausible
> >scenario in my opinion.
> >
> Whether it is plausible with these specific names is beside the point.  As I
> argued elsewhere, an attempt at reversal that barely fails is not implausible
> in some circumstances.  Likely it will be those Dole voters who see the least
> difference between Clinton and Nader who will be willing to take the chance of
> reversal, so strategy shades into sincere preference -- in practice we would not
> really know how many were reversing, and how many sincerely were so disgusted
> with Clinton that they would even prefer Nader.  
> 
> >The alternative--that 18 D>N>C voters really do prefer Nader--is more
> >plausible.  A question we shouldn't ignore about the alternative
> >scenario is: who *ought* to win with those sincere preferences?  
> >Why not Nader?  If the win is instead accorded to Clinton, 48 voters 
> >will be happier and 52 will be unhappier.  If the win is accorded to
> >Dole, 46 voters will be happier and 54 will be unhappier. 
> >
> 
> If Nader wins because the Clinton voters could have elected half-votes but
> ignorantly voted only for Clinton in the belief that a second ranking would be
> used only to decide between Dole and Nader (whom they find equally
> unacceptable),
> then I count this as a failure of an unduly complex  system.  If Nader wins

It's you who'd add complexity to the system by adding an option
(or worse, a compulsory or default ballot interpretation).

I've already told you that those half-votes by the MIddle voters
would be a mistake, and that if the Middle voters vote no 2nd
choice, then order-reversal can only backfire, and therefore
wouldn't be attempted. 

> because 
> the Clinton voters did not even have the option to cast half-votes, then I
> have to ask, 
> should Nader really prevail in the case where he is ranked equally with Dole
> at the
> bottom of all Clinton voters' ballots, when had the Clinton voters all voted
> a second
> choice, and two-thirds of them had actually preferred Nader to Dole, Nader
> would 
> have lost?  And should the Clinton voters be coerced into voting false

Won't happen if the Dole voters know what they're doing at all. If
they're sophisticated enough to order-reverse, then they're 
sophisticated enough to realize that it will backfire everyk
time.


> preferences in 
> order to maximize the chance to elect their candidate?

But your half-preference votes involve false votes anyway.
No one's forced to vote false preferences in non-falsified
Condorcet.

> 
> -- Hugh Tobin
> 
> >---Steve     (Steve Eppley    seppley at alumni.caltech.edu)
> >
> --
> Registered ICC User
> check out http://www.usefulware.com/~jfoltz
> 
> 
> .-
> 


--