EM vote on SW methods: summary of votes to date

Mike Ossipoff dfb at bbs.cruzio.com
Fri Jul 26 00:36:05 PDT 1996

Yes, as Steve said, Condorcet (which I've previously been calling
"plain Condorcet") is the 2nd place winner by every meaningful
way of looking at it. Not only is it the winner in a second count,
after Smith//Condorcet is removed from the rankings, but it also
is less beaten than IRO (Instant Runoff), according to votes-against,
and even by margins (not that margins mean anything to me; I
only checked out the margins with respect to Smith//Condorcet).
So why did I call IRO "Smith//Condorcet's closest rival", or 
something to that effect? Good question! I said that because
I was looking at the _ratio_ by which Condorcet & IRO were
beaten by Smith//Condorcet. But that was really sloppy of me,
because ratio isn't relevant to any standard that means anything
to me, or which is relevant to things that are important to voters
or electoral reformers. How embarrassing!


As for how to count equal rankings in an IRO count, I claim
that both equally-ranked 1st choices should receive a _full_
vote. When ranking them in 1st place, surely the intent is
to _fully_ vote both against everything else. Then, of course,
if 1 of them were be eliminated (which didn't happen) then
the other would keep its 1 vote, since the intent is still to
give it a full vote over the other alterntives. If both 1st
choices were to be eliminated, then of course the single 2nd
choice would get 1 vote, indicating that it is being voted
against everything less-liked.

If that sounds arbitrary, it isn't really: The principle behind
it is that everything in the ranking is intended to receive 1 full
vote against everything lower-ranked.

I word the brief IRO count rules slightly differently to 
summarize that possibililty, the possibililty of equal rankings:

Repeatedly, eliminate from the rankings the alternative that occupies
or shares highest position in fewest rankings.

Either continue this till only 1 alternative remains un-eliminated,
or stop as soon as 1 or more alterntatives occupy or share highest
position on more than half of the rankings, declaring as winner
the alternative occupying or sharing highest position in the
most rankings. I probably would prefer the 2nd of those 2
procedures, though the 1st one is simpler & briefer to state,
and, since brevity of definition is the only good thing that
can be said about IRO, that briefer 1st procedure would seem
the one to propose, if IRO were being proposed, with the
possibility of equal rankings.

For the purposes of our IRO count, the 2nd procedure in the
preceding paragraph seems more in the spirit of majority-rule,
stopping when 1 or more alternatives top a majority of the
rankings. Continuing to do eliminations, as in the 1st procedure
in the preceding paragraph, isn't as good, since most of us
don't consider eliminations a good thing--so it's better no
to do it any longer than necessary, and to quit when something
occupies or shares highest position in a majority of rankings.

Though I said that's more in the spirit of majority-rule, I
want to emphasize that really no version of IRO is genuinely
in the spirit of majority rule. It's just a matter of not
doing the un-majoritarian elimination process any more than
necessary. Much insult to majority rule may have already
been done by the time that something tops a majority of the rankings.
All I'm saying is that at least it's better not to continue the

I'm not saying that those 2 procedures would give different
results in our election. One thing for sure, another advantage
of stopping the count when 1 or more alternatives occupy or share
highest position in a majority of the rankings is that it
could greatly reduce the job of the count.



More information about the Election-Methods mailing list