Lowest Majority loser tie breaker

[Mike Ossipoff]

DEMOREP1 at aol.com writes:
> 
> Yet another Condorcet circular tie breaker for single (or multiple)
> executive/judicial elections ----
> 1. The voters vote yes or no on each candidate and rank number each
> candidate.
> 2. The simple Condorcet method is applied to those candidates who get a
> majority yes vote.
> 3. If there is a Condorcet circular tie, then the candidate who has the
> lowest yes majority loses and the Condorcet method is redone. Such loser has
> had his/her chance to win head to head. Repeat if necessary until the winner
> remains (or winners remain). 

As I've suggested previously to Demorep, he'd be doing a better job
of advocating all of his new methods if he'd tell us in just what
way they're supposed to be better. What criteria do they meet? What
standards do they meet? How do they do what voters want?

Demorep's latest proposal is close to IRO. Repeatedly eliminate the
candidate approved by fewest people. We've been over the ways
such methods can fail.

Someone asked who Demorep is, and I'd like to suggest that
maybe Demorep is Bill Clinton :-) because that's someone who'd
be well-described by that bipartisan name. But even if that isn't
who it is, the name implies acceptance of the 2-party system,
and the notion that politics consists only of Democrat &
Republican. 

Demorep once answered my request for claims about standards
& criteria by re-stating his rule & saying "That's my
standard & my criterion." Well yes, would that be the
"hippopotamus logic" that Don spoke of? If the only standard
or criterion met by Demorep's methods is one written so that
it repeats their rule, then we shouldn't be at all surprised
if Demorep's methods meet those criteria better than does
any other method :-)

> 
> Item 3 is new. I repeat again the absolute necessity that the winner have
> majority "positive" support. The main single winner competitors to Condorcet-
> approval voting, top 2 in runoff, instant run-off  - each generally produce a
> majority winner (which may be the wrong majority winner in a head to head
> Condorcet test).

Wrong majority winner? How about false majority winner. IRO & Runoff
can make someone look look like a majority winner, but that isn't
quite the same thing.

> 
> A Condorcet tie breaker winner using the least votes against in his/her worst
> defeat will be absolutely blasted as being chosen by a minority (since such
> winner would be opposed by one or more majorities). 

What are you talking about? In a circular tie everyone is beaten.
Any choice from a circular tie could be "blasted". But certain
majority principles & criteria can still be protected, and that's
what Condorcet does.

> 
> The crisis in U.S. politics (and in many other countries) is due to the lack
> of guaranteed majority rule in all elections (legislative, executive and

Successive elimimination of candidates with fewest "yes" votes will
violate majority rule in a way similar to the way that IRO does.
Yes, majority rule is important. No, Demorep's methods don't
protect majority rule.

As the proponent of all those methods, it's Demorep's responsibility
to actually demonstrate to us what they'll do. Don't just
define them.

> judicial) due in great part to historical accidents in each country (along
> with some deliberately adopted monarchial and/or oligarchial election
> features in each country-- such as the infamous U.S. Electoral College for
> electing the U.S. President).
> 
> The lesser of two (or more) evils situation is a manifestation of such lack
> of majority rule. The next U.S. CRISIS is Nov. 5, 1996.

Oh? I thought the next yawn was Nov 5, 1996. But it's true that
the national boredom with political choices is largely causesd by
an extreme lack of selection, and that lack of selection is the
result of the LO2E problem & the spoiler problem, and those
things are closely related to a lack of majority rule. As
I've said, majority rule, & getting rid of the LO2E problem
are closely related.

> 
> Could any of the 1996 U.S. President candidates on the Nov. 5, 1996 ballot
> get a majority of yes votes ? I doubt it.

That brings into question the desirability of a choice rule that
would disqualify everyone. I'm not saying that I wouldn't like
that. I'm leaving all the partisan races blank on my ballot
on the 5th, including the Presidential election. 

In fact, I was just going to post a message to the effect that
this upcoming election has shown me how desirable NOTB is.
First of all, though, I emphasize that, as Steve pointed out,
if we had Condorcet's method, we'd have a far better selection,
and it would be doubtful whether anyone would need to vote
NOTB over everyone (which is what I'd do in the upcoming 
election).

We've already discussed Demorep's propsal to take y/n votes on
each candidate & disqualify anyone who gets "n" from a majority
or (was that it?) who doesn't get 'y" from a majority.

If it was the latter, then I oppose it. Disqualificatin shouldn't
be based on indifference, but only on active opposition.

As I said, we talked about this. We seeem to debate each
issue over & over again. A proponent makes his proposal &
arguments, and then waits a while, and recycles the whole
thing again.

As I said before, there's nothing wrong with taking y/n
votes & disqualifying anyone with "n" from a majority. But,
as I showed last time we talked about it, such a rule isn't
really necessary with Condorcet's method, which, without
help, does a good job of denying victory to majority-rejected
candidates.

Still, I didn't oppose y/n being added, and the disqualification
of anyone with "n" from a majority, and I don't oppose it now.
I'd use it myself on Tuesday.

In fact I've proposed (did you notice, Demorep?) a method
that I called "Yes/No". It was intended to be a simple, nonranked
method that would improve on Approval, in the event that we
couldn't get rank-balloting in some jurisdiction:

Voters may give any candidate a "y" or a "n" or neither. 
Any candiate getting "n" from a majority of all the voters
is disqualified. The winner is the un-disqualified candidate
with the most "y" votes.

Though I don't oppose adding the y/n feature to Condorcet
elections, as a way of disqualifying candidates, I claim
that it isn't needed, and that therefore it could be an
unnecessary complication that would lengthen the specification
of the proposal.

On the other hand, adding y/n & disqualification by a majority
voting "n" could have tremendous popular appeal, to a public
that is so justificably cynical, and wants more than anything
else to vote against people. Condorcet counts their votes
against candidates, but adding the y/n feature really makes
it obvious & out-in-front, and that may be a very desirable
thing to help acceptance.

But I disagree with Demorep's propsals to, in various ways,
use y/n to modify Condorcet's method itself, as opposed to
just disqualifying people. By so doing, Demorep would make
a completely different method that would have none of Condorcet's
unique & important properties.

***

By tahe way, there are several ways of using NOTB, one of which
corresponds to disqualification by majority:

I'll list them from mildest to strongest:

(In all these procedures, if NOTB wins, then everyone is
disqualified)

1. If NOTB is in the circular tie then everyone is disqualified
2. Any candidate beaten by NOTB is disqualified
3. Any candidate over whom a majority of all the voters have
   ranked NOTB is disqualified
4. No disqualification except the disqualification of everyone
   if NOTB wins.

These proposals should be considered by taking into account that
the actual rule might be that if a everyone is disqualified
then the winner would govern for a very short interim duration,
till there's a 2nd election.

#1 may be too extreme. It would be too easy for a likely-to-lose
faction to engineer a circular tie in order to shorten the
winner's term in office.

#2 could cause strategy vulnerability, when truncating voters
could engineer the elimination of the Condorcet winner and the
election of their candidate, by letting NOTB beat the
Condorcet winner by some small margin.

I like #3 best, so far, though I haven't checked out how it
does under order-reversal conditions.

In any case, of course if NOTB wins, then everyone should be
disqualified.

***

I re-emphasize that NOTB, or any other majority disqualification
system, won't really be needeed, because Condorcet does a
good job of that, and because there will be a good enough
selection that we'll all have candidates we like better than
NOTB.

***

Mike



> 
> .-
> 


--