[EM] IRV's criteria successes?

MIKE OSSIPOFF nkklrp at hotmail.com
Mon Feb 4 18:17:54 PST 2002



Markus said:

in so far as IRV meets majority for solid coalitions and independence
from clones, IRV can hardly be called "erratic" compared to primary
with runoff.


I reply:

If I tell you that someone drives erratically, that fortunately
doesn't mean that he does everything that can be called erratic.

IRV is erratic because it requires strict ordering, collects a ranking,
and then makes irreversable decisions based only on a fraction of
the ballots' information, looking only at 1st choice votes.

But it was inevitable that eventually someone would post about IRV's
criterion successes.

When Markus says "majority for solid coalitions", I assume that he
means what I call Mutual Majority, and Bruce Anderson used to call
Generalized Majority Criterion.

I renamed it Mutual Majority because I didn't find it to be very
general. It's about a fortuitous special case. Here's Mutual Majority,
written in a way that's contrived so that Approval won't pass:

If there's a group of voters who are a majority of all the voters
and who all prefer all the candidates in a certain set to all the
candidates outside that set, and if those voters vote sincerely,
then the winner should come from that set of candidates.

[end of definition]

Sounds good, but a mutual majority is also a situation where IRV
will demonstrate its failures of FBC & WDSC, showing its lesser-of-2-evils 
problem and its gross and avoidable violation of majority rule.

Say you believe, rightly or wrongly, that there's a majority who
prefer both Worst & Middle to Favorite. Why should you vote Favorite
in 1st place? No reason to. He can't win. But you can gain by voting
Middle in 1st place, in case he needs your vote to avoid immediate
elimination and the subsequent transfer of his votes to Worst.

The Clone Criterion, too, is about a special case. Sure, with IRV,
adding clones won't hurt your faction, but other ordinary & typical
situations will make IRV violate FBC & WDSC.

WDSC says:

If a majority of all the voters prefer A to B, then they should
have a way of voting that ensures that B won't win, without any
of them reversing a sincere preference.

[end of definition]

Approval passes FBC & WDSC. Plurality & IRV fail both.

Mike Ossipoff







_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com

FFrom election-methods-list-request at eskimo.com  Mon Feb  4 18:32:21 2002
Received: (from smartlst at localhost)
	by mx1.eskimo.com (8.9.1a/8.8.8) id SAA12511;
	Mon, 4 Feb 2002 18:32:05 -0800
Resent-Date: Mon, 4 Feb 2002 18:32:05 -0800
Message-ID: <3C5F3878.9020408 at postmark.net>
Date: Mon, 04 Feb 2002 17:42:16 -0800
From: Blake Cretney <bcretney at postmark.net>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.7) Gecko/20011226
X-Accept-Language: en-us
MIME-Version: 1.0
To: election-methods-list at eskimo.com
Subject: Re: [EM] Sincerity for criteria, Blake
References: <F235XQQez1enAqpfkNs0000af3a at hotmail.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Resent-Message-ID: <"GwsbI.0.P33.bGqNy"@mx1>
Resent-From: election-methods-list at eskimo.com
Reply-To: election-methods-list at eskimo.com
X-Mailing-List: <election-methods-list at eskimo.com> archive/latest/7263
X-Loop: election-methods-list at eskimo.com
Precedence: list
Resent-Sender: election-methods-list-request at eskimo.com

  MIKE OSSIPOFF wrote:

>
> Blake continues:
>
> Maybe there are people who are so in awe of mathematics that when you
> prove that a particular definition (of your own design) applies, they
> feel that you have proven your point. If I define the "Great Method
> Criterion" and prove that Ranked Pairs passes it, they will take it as
> proven that Ranked Pairs is a great method.
>
> I reply:
>
> I use a number of criteria that aren't of my own design. But
> I use some that are of my own design too. I never ask people
> to value a criterion just because I wrote it. My criteria are
> mostly lesser-of-2-evils criteria, and it doesn't take any
> convincing for most people to accept the importance of the
> lesser-of-2-evils problem. The relevance of my defensive
> strategy criteria to the lesser-of-2-evils problem is so obvious
> as to not require any argument.

>
> The difference between your standard and mine, Blake, is that
> mine is one that's shared by lots of people who express concern
> about its violation. Lots of voters and reform advocates
> express concern about the lesser-of-2-evils problem. It's recognized
> to be the one big thing wrong with Plurality.

The lesser-of-2-evils refers to a sense on the part of many voters that 
they aren't voting for someone they particularly like; they are only 
trying to keep someone they particularly dislike out of power.  So the 
best they can hope for out of an election is a lesser evil.  From this 
simple start, you derive various standards and criteria.  To do this, 
you add your own assumptions and standards.  So you are no longer simply 
responding to the voter's concern.

There are other problems with your argument, though.  Until plurality is 
replaced, most arguments are going to be directed against it, but that 
doesn't mean that its problems are the most important kind, as you 
imply.  Just that it's problems currently effect more people. 
 Gasoline-driven cars produce air pollution.  Perhaps nuclear cars will 
produce fission leaks.  People seem more concerned about cars producing 
air pollution than fission leaks, so I guess air pollution must be worse.

But actually there's good reason to believe that reformers aren't 
primarily concerned with the lesser-of-2-evils problem.  The biggest 
single-winner campaign is for IRV, and this is because reform advocates 
often become obsessed with quite different strategy problems than the 
lesser of 2 evils (as least as you understand it).   They may be wrong, 
but it was you who invoked the ad populum argument.

---
Blake Cretney


rom election-methods-list-request at eskimo.com  Mon Feb  4 18:32:21 2002
Received: (from smartlst at localhost)
	by mx1.eskimo.com (8.9.1a/8.8.8) id SAA11554;
	Mon, 4 Feb 2002 18:30:12 -0800
Resent-Date: Mon, 4 Feb 2002 18:30:12 -0800
Message-ID: <3C5F3805.5030506 at postmark.net>
Date: Mon, 04 Feb 2002 17:40:21 -0800
From: Blake Cretney <bcretney at postmark.net>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.7) Gecko/20011226
X-Accept-Language: en-us
MIME-Version: 1.0
To: election-methods-list at eskimo.com
Subject: Re: [EM] 02/03/02 - STV for Candidate Lists:
References: <v01520d00630bd21c2204@[64.79.81.3]>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Resent-Message-ID: <"TpvY02.0.Sq2.pEqNy"@mx1>
Resent-From: election-methods-list at eskimo.com
Reply-To: election-methods-list at eskimo.com
X-Mailing-List: <election-methods-list at eskimo.com> archive/latest/7261
X-Loop: election-methods-list at eskimo.com
Precedence: list
Resent-Sender: election-methods-list-request at eskimo.com

  Donald Davison wrote:

>02/03/02 - STV for Candidate Lists:
>
>Dear Adam,
>You wrote: " While it [Party List] is not quite as efficient in making
>every vote elect a representative as STV is, it is highly proportional,
>highly democratic, and extremely simple to implement and understand. "
>
>Donald:  While I agree to the good things you say about Party List, it does
>have a number of faults.
>
>One: Its high proportionality is only for party.
>
>Two: The voter is not allowed to cross party lines.
>
>Three: The order of the candidate list of each party is suspect.
>
>The mathematical correct way to determine the order of candidates for a
>party would be to use STV, but then why not use STV for the entire election
>and drop Party List?
>Yes indeed, why not?
>

I think that closed party list has a lot to say for itself.

1.  It allows districts/constituencies to be eliminated.  They tend to 
corrupt the process.
2.  It provides a high degree of proportionality.
3.  It is simple both to use and understand.  Voter's need not even be 
literate.
4.  More importantly, the elections themselves are easier to understand, 
because they are so partisan.  Purely local races are confusing to 
voters, and require a lot more money in order to educate/propagandize 
them.  This is why American style personality politics has the problems 
it does.  Voter's are confused, and politicians are constantly desperate 
for large sums of money from people like the folks at Enron.
5.  Although there are strategy issues, party list doesn't have many of 
the peculiarities of other methods.  For example, in STV and cumulative 
voting there are strategies involving getting an even split of 
supporters between the candidates of one party.  In mixed systems, there 
are strategies involving getting a district candidate who shares one's 
ideology, but not party affiliation.
6.  Voters are voting for a known entity, a pre-determined party list. 
 Voter's aren't forced to make every decision, but they are able to 
respond effectively when list makers make bad choices.  That's a more 
realistic model of democracy than the micro-management model.

Let me point out that even in STV it will still be necessary for parties 
to choose which candidates represent them, although they do not put them 
in order.  Presumably Donald would find this choice "suspect" as well. 
 And what on earth does he mean by "mathematically correct"? 
 Mathematics has no preference as to how we govern ourselves, or even 
whether we live or die.

---
Blake Cretney




More information about the Election-Methods mailing list