[EM] This installment wasn't sent

MIKE OSSIPOFF nkklrp at hotmail.com
Mon Oct 9 21:06:47 PDT 2000


Hi--

It seems to me that this installment of my reply to Janet might not
have gotten sent, and so I'm sending it now. My apologies if I'm
sending it twice. But first let me add a brief p.s:


When I spoke of the CVD claim that it's ok to vote for Nader in
1st place in IRV, because he's a sure loser, sure to get immediately
eliminated, I pointed out that that isn't true. But, aside from that,
what do you think of a method that only lets you fully support your
favorite if he's a sure loser?

Another thing: IRV, because of its capricious choice of whose
voted preferences to count and whose not to count, likely is
unconstitutional, because it violates the U.S. Constitution's
"equal protection under the law" clause, a clause often applied to
voting rights. IRV violates the basic principle of one person, one
(counted) vote. Not counting some people's preferences violates
1-person-1-vote, and the equal protection clause.

Now, here's the letter that I might not have succeeded in posting
yesterday:

The lesser-of-2-evils problem is the reason why we want a better voting
system. We therefore have no use for a method that doesn't improve
on Plurality in regards to that problem.

I've talked about IRV, but how do other methods do? If we can't
get a method that belies Arrow's impossibility theorem, then how
good a method can we get? That's the subject of this brief letter.

Let me mention again two websites where you'll find information about that:

http://www.electionmethods.org

http://www.barnsdle.demon.co.uk/vote/sing.html


But let me say a little here:

Approval is a briefly-defined method: It's the same as our current
Plurality method, except that voters may vote for more than 1
candidate if the so wish, and, when they do, they give one whole vote
to each candidate for whom they vote. As now, the candidate with the
most votes wins.

Approval makes some important lesser-of-2-evils guarantees that IRV
can't make:

With Approval, no one ever has any need to vote anyone else over his
favorite. I've shown how common that need will be with IRV. With
Approval, everyone can always fully vote for their favorite, every
time.

With Condorcet's method (defined at the websites) there are common
, probably typical, conditions under which no one need consider any
kind of strategy, but can sincerely rank as many candidates as they
want to. That of course stands in stark contrast to IRV, with which it
will be common for voters to need to rank someone else over their
favorite.

The above 2 paragraphs are about the criteria called FBC, SFC, &
GSFC, at the websites. There are other criteria too, relating to the
lesser-of-2-evils problem. Those criteria are defined at the websites.
They're SARC, UUCC, WDSC, & SDSC. IRV fails all of them. Approval and
Condorcet each pass some important ones, of those criteria.

I shouldn't lengthen this letter by defining those criteria, so I'll
just refer you to the website. Let me just add that I consider WDSC
my basic minimum requirement for an adequate method. Approval & Condorcet
comply with WDSC. IRV fails WDSC. WDSC is about need for order-reversal
in the typical lesser-of-2-evils situation.

In conclusion, when we ask what we want from a voting system, and
put it in precise terms, as precise yes/no tests, we find that IRV
fails every one of those tests--if we care about the lesser-of-2-evils
problem or majority rule.

Mike Ossipoff




_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Share information about yourself, create your own public profile at 
http://profiles.msn.com.



More information about the Election-Methods mailing list