[EM] 09/20/02 - The Manipulation Test:

[Donald E Davison]

09/20/02 - The Manipulation Test:

  -------------- Forwarded Letter and Comments ------------
From: Saari at aol.com
Date: Wed, 18 Sep 2002 04:06:43 EDT
To: donald at mich.com

Mike Saari wrote: "I still have seen no resolution to a very simple question."

Dear Mike,

I recall that you did present this question before, but I don't remember on
which list.  I must have deleted the post (Mike `gets no respect').
It's not my problem if no one on a list will not reply to some of your
posts, welcome to the club.  This is not a perfect world, people don't
always do what we would like them to do, get use to it.  I'm going to
answer your letter now because I can use your letter as a carrier to make a
point or two of my own (he also wants to bad mouth Approval Voting).

A question like yours should be presented to a list, then no one can be
blamed for not responding.  I have a question for you: `If a letter falls
onto a list and no one responds, was there any noise?"

Anyway, I am forwarding your letter to a few lists, maybe there will be
some noise this time.  One of the lists is the Election Methods list, a
`hot-bed' of Approval Voting cult members.  Surely, one or two of the
cult's members will give you the resolution you are seeking, (the `hot-bed'
will give Mike some respect).  For sure, you're not going to get what you
want from me.

Mike Saari: "Suppose there are only two candidates, or suppose that all
candidates have been eliminated down to only two.

Here is the sample scenario:
All of the voters rate candidate "B" as "very good".
60% of the voters rate candidate "A" as "excellent"
The other 40% of the voters rate candidate "A" as "awful".

Donald: Candidate A is the winner.  I'm telling you that right up front.  I
don't need to hear a bunch of twisted logic to the contrary.

Mike Saari: "(assume these "ratings" are honest expressions of the true
feelings of the voters,..."

Donald:  No No Mike, No can do.  If we are going to have a debate we both
need to agree on the `ground rules'. I do not agree to assume these
`ratings' are honest expressions of the true feeling of the voters.'  This
is a flaw in your example.  The true expression of each `B' voter would not
bullet vote candidate `B' in the face of the fact that the `A' voters all
rated candidate `B' in second place.  Honest expression of true feelings
would be mutual, in other words, most of the `B' voters would rank
candidate `A' as second choice, unless the faction decided to manipulate
the election and try to gain an advantage by bullet voting, which is most
likely what is happening in your election, but you'll only get away with it
one time.  The `A' faction will also be bullet voting in the next election,
and then goodbye Approval Voting, for Approval Voting is what you are
selling with your simple minded question.

Mike Saari: "...although how such data can be gathered in real life is a
separate question)."

Donald:  Don't be sly, we all know it's Approval Voting you are selling.

Mike Saari: "So which candidate is the "better" choice - the one liked by
everybody, or the one hated by 40% of the group?"

Donald:  Answer: Candidate `A'.  See I told you you're not going to get
from me the answer you are looking for.

Mike Saari: "The question is not answerable with logic."

Donald:  Sorry, but I must insist that we must agree on any `ground rules'.
the question is answerable by logic or illogic.

Mike Saari: "I simply note that virtually every known "voting system",
including IRV, will inexorably choose candidate "A" as the winner (since
more than half prefer A over B).  The fact that A is intensely disliked by
many (whereas B is liked by everybody) is disregarded under most so-called
democratic voting systems."

Donald:  You have just implied that most all election methods are not
democratic.  You should not fight two battles at once.  You run the risk of
someone only replying to this one point, if so, he would sidestep your main
question, leaving it unanswered, like it was unimportant, like as if not
replying at all to your letter.  You already know what that feels like ( Ha
Ha ).

Mike Saari: "There are some who believe, philosophically, that the
candidate preferred by the majority MUST be the correct choice."

Donald:  CORRECTION:  Almost `ALL' people believe `the majority MUST be the
correct choice.'

Mike Saari: "Such a belief is unfounded, though, if you realize that this
method creates a "40% are unhappy" group, whereas a system with some
possibility to choose the less "exciting" (but liked by all) candidate B
creates a group which is more contented overall."

Donald:  If this is the best argument you can present for Approval Voting,
then you lose, because almost everyone wants their own number one to win.
The Approval supporters are fools to think Nader would have won if the last
Presidential election was an Approval Voting election.

Mike Saari: "I don't think that a true democracy should choose intensely
disliked (by 40%) candidates over a "liked by all" alternative, but that is
exactly what will happen for the scenario just described."

Donald:  Well, think again.  You are wrong, but your example is not a total
waste, it does point to another flaw in Approval Voting.  Your example is a
case in which one candidate (A) is leading in the pre-polls by 60 over 40,
but some other candidate wins the election, the Approval election.  Almost
all the public believes in the policy of the majority winner being the
`true majority winner', if I may use that term (oh oh, you're in trouble
now, the Approval cult doesn't like for you to use that term, that term
belongs to them).
The public is not going to stand for the `true majority winner' losing the
election because of some weird method - out with the method.

Mike Saari: "How does IRV respond to such criticism?"  Thanks, Mike

Donald:  Irving allows the voters to make lower choices without those lower
choices harming an earlier choice, this is not true of Approval Voting.
But of more importance is the final test of a method, that is, The
Manipulation Test.  The policy of manipulation has been raised by Tom Ruen
and next is my description:

  T H E   M A N I P U L A T I O N   T E S T   by Donald Davison

"If a method can be manipulated, then, by God, let's teach the public how
to manipulate the method.  The ensuing chaos will force the public to seek
and accept a better method."


Approval Voting can be manipulated out of existance.  Irving cannot be
manipulated.
You, or any other person, is invited to try to manipulate an Irving
election so that the election elects the wrong candidate, that is , a
minority candidate of a minority faction.

Copies of Tom Ruen's letter on his manipulation suggestion and my reply to
him are below:


 ------------ Tom Ruen's Letter -------------
To: <stv-voting at yahoogroups.com>
From: "Tom Ruen" <tomruen at itascacg.com>
Date: Fri, 6 Sep 2002 23:44:44 -0500
Subject: [stv-voting] Re: Droop-STV results from CV elections

Dear STV-supporters,

I was a little disappointed to hear no opinions about my thoughts of using
STV power within subgroups of voters of a CV [cumulative voting] election.
I know I get too wordy, but I'm still hoping for a reply. I'll try again
briefly.

It is an interesting argument I think - to notice that STV falls out of
cooperative interest of CV voters. (There may be other cooperative
strategies, but I can't guess now.)

Cumulative voting suffers from the identical spoiler effect of plurality
election. Both methods reward coalitions of voters [that] are willing to
work together before the election and agree where to put their united
support.  In plurality, if 50% of the voters agree to agree, they have all
the power.  In CV, if 1/(N+1) of the voters agree to agree, they can
control one seat among N. IRV and STV both serve as methods to figure out
who agrees and how much power that offers.

This perspective may be a fair way to explain STV within a CV context, at
least if allowing fractional votes. Take a group of similar minded voters,
and let them agree to work together to get what they want. Let them all put
a vote on their favorite candidate, and see which candidates get the most
votes. If some candidates get too much, then rules can be applied to allow
those supporters to move some votes for a next favorite, so long as the
first is guaranteed victory. Next candidates with far too little support
may be abandoned and votes voluntarily moved to help a stronger compromise
near the threshold for victory.  In the end, after all the possible
controlled victories are guaranteed, the group can offer their "extra"
votes for the strongest compromise candidate they can help but can't
guarantee victory. In the end the group has CV ballot votes for all members
to maximize their collective power to elect winners.

There is no law against voters working together like this, and once some
start doing this, other may see it is in their best interest to follow.
And there's no competition of groups - any two groups using this strategy
separately are still better off working together so less votes are wasted
on winners. At some point, it will be clear we should just STV for all
voters since it offers a simple and fair structure for this cooperation to
occur.

Is this story too obvious?  It was somewhat of a surprise to me.  It is
potentially practical in my corporate board CV elections.   Big share
holders recognize their gamble and *if* the big ones are willing to work
together before the election, my point will be made and we may be one step
closer to the clarity of STV.

Tom Ruen

 ------- Reply to Tom's Letter ------- 9/20/02 Draft:
To: stv-voting at yahoogroups.com
Date: Tue, 10 Sep 2002
Subject: [stv-voting] 09/10/02 - Tom Ruen's Droop-STV:

Tom Ruen wrote:
"I was a little disappointed to hear no opinions about my thoughts of using
STV power within subgroups of voters of a CV election. I know I get too
wordy, but I'm still hoping for a reply. I'll try again briefly."

Dear Tom,  I would like to encourage you to continue posting your
thoughtful letters, some of them are good.  You have been disappointed
before, for this is not the first time no one has replied to one of your
posts.  You will just have to bear with us, most of us are not as quick as
you are, we need some time to realize what it is you are talking about - HA
HA.
  Besides, I for one usually do not get around to dealing with my email
within one day.

Anyway, correct me if I'm wrong, but I think your policy is as follows:

     `If an election method can be manipulated, then, by God, let's teach
the public how to manipulate the method.  The ensuing chaos will force the
public to seek and accept a better method.'

If this is your policy, it sounds good to me, I think you are on to
something, let's do it.

This policy would educate a faction on how it can use the Droop quota
and/or averaging votes at the ballot box in order to gain an advantage.
Once one faction gained an advantage, other factions would be forced to do
the same in order to compete.  This will result in most voters facing the
ordeal of averaging their votes at the ballot box.  Hare Preference Voting
experienced this ordeal in its history.  Its solution was Droop Preference
Voting.

If this is your policy, then I'm with you.  Let's devise ways to manipulate
the different election methods, single and multiseat.

In the case of your board election it might be possible to operate this
scheme with a quota less than the Droop quota.  Because there are no votes
transferred in Cumulative Voting,  many votes are excluded in the surplus
votes and the votes of the lower candidates, which means that less votes
will be required to elect a member.  I remember that you stated that one of
you board elections only had six candidates for five seats.  That is a low
number of candidates per seat, but my guess would still be that two or
three candidates were elected to your board with less than a Droop quota of
votes.  [two were elected with less than the Droop quota]

With more candidates, like ten for the five seats, we could expect two or
three to be elected with maybe even less than fifty percent of a Hare
quota.  This scheme will work better if it has a lower target.
The scheme will work with your CV board elections, but CV is only one
little used election method, the big prize will be Plurality-at-Large.

      `Your CV board election today, tomorrow the world.'


Regards,
   Donald Davison, host of New Democracy at http://www.mich.com/~donald
                        Candidate Election Methods
   +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
   |                        Q U O T A T I O N                          |
   |  "Democracy is a beautiful thing,                                 |
   |        except that part about letting just any old yokel vote."   |
   |                           - Age 10 -                              |
   +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+

         Please be advised that sending email to me allows me to
         quote from it and/or forward the entire email to others.






















----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em