Votes Against Tiebreaker

Mike Ositoff ntk at
Sun Oct 18 19:24:27 PDT 1998

> The problem is that if there is not a YES/NO vote, then the *absolute* support
> for any winner using VA is questionable.  Both B and C may not have YES
> majorities.

A choice between several candidates is a relative choice, not
a matter of absolute acceptance.
Sometimes, unavoidably, it's necessary to have one of several
alternatives, even if none of them can get an absolute majority.

> I must again bring up the law making analogy.  If a bill comes before a
> legislative body, then it either gets or does not get a majority in favor of
> it.  Having 2 or more bills on more or less the same subject does NOT do away
> with the majority requirement.   

As I've said, let's not use current practice as a standard for
how things should be done. Any, the successive Y/N vote, where
each alternative is voted on in turn, replacing the previous one
if Yes wins, is just another single-winner method, like VA.
Not as good as VA, because it has the offensive strategy possibility
of my voting against your alternative when it's paired with a
pushover, so that my alternative, later, will go against the
pushover instead of your alternative. But you can easily 
retaliate by voting against my alternative when it comes up
for vote. Therefore, "Sequential Pairwise", while not the best,
is still very good in meetings, if we want something whose
time requirement is less than pairwise count, and which takes
about as long as successive eliminations (the meetings form of

Essentially, in those votes, you're not voting on the _absolute_
desirability of the alternative; you're voting on whether it's
better than the previous winner. You've got to realize that
multi-alternative elections are about _which_ alternative wins.
It's about which is better in comparison to the others, not
about their desirability in some absolute sense.

Sure, I refuse to vote for candidates who are absolutely unacceptable
to me, so absolute considerations are present. I'd even enjoy
giving them a "No". But Y/N voting greatly complicates strategy,
in votes that, however I may feel about absolute sleazes, 
must procedurally be treated as relative choices, since something
or someone has to win, regardless of whether any of them are
absolutely acceptable.

> All of a sudden things are supposed to be different when electing public
> officers (such as Presidents, Prime Ministers or Supreme Court Judges) using
> VA-- i.e. it becomes OK for candidates to be elected with majorities against
> them.  IMHO, the voters/politicians will never accept such a reform.

Remember that VA won't avoidably elect someone with a majority
against him.

> I mention again that lesser of two evils (LO2E) false (i.e. insincere)
> majorities in general elections are generally the result of having 2 plurality
> winners in Democrat and Republican primaries and/or in partisan top 2 runoff
> primaries and/or nonpartisan top 2 primaries.

It's the result of FPP's failure of the 1st Choice Criterion.
Runoff also fails it, though not as badly as FPP & IRO.

Mike Ossipoff


More information about the Election-Methods mailing list