[EM] Losing with 99% of the vote: Mike tell us why

[Craig Carey]

I see that Mike retained his FBC definition. He thinks it passes the
Approval Vote method. He and I are in a dispute over whether reading
and understanding one of his definitions is a valid use, if the result
contradictions his assertions on what the definition means.

Can anybody post in text that proves
that Mike Ossipoff has once tested any non trivial method with one of
his rules?. I.e. on any extant rule on any method in any year say in
the last 1yr. I presume not a single subscriber can find an instance of
Mike Ossipoff applying one of his rules to a method with a rigour to
prevent him making a single mistake or some vague link in the reasoning.

---------------------

I ask that Mike Ossipoff tell the mailing list what is visible on
[his] Approval Vote ballot papers. This is of great importance, e.g.
the to question of whether Approval Vote was always a really
misnamed hoax designed to irritate the Internet and the list.

Especially these matters are asked about:

  (a) checkboxes vs. indications of 1st, 2nd, 3rd, etc.

  (b) a statement on how many marks [subvotes] can be made on the
     [Mike Ossipoff style] Approval Voting papers.


I remind Demorep to answer my questions too. They were good questions
and the readers that are interested in the future of the United States
of America would want to very reliably know where Demorep stood on
a few basic issues (fundamentals, in [(high tech)] STV-style methods).


--------------------------------------------------------------------------


                         * QUESTION TO MIKE *

                 REQUEST FOR MIKE TO MAKE A STATEMENT


This is to Mr Mike Ossipoff.

I request under the guidelines of this mailing list, this question:

To Mr Mike Ossipoff: what is the probability of this election example
occurring?

 >   99 ABCDEFGHIJ
 >    1 J

What is the probability. You had a method in May 2000 of saving the
Approval Vote by saying that bad examples were improbable. We asked
for the probability and you evaded answering while not writing nothing.


--------------------------------------------------------------------------

At 00:50 31.05.00 -0400 Wednesday, DEMOREP1 at aol.com wrote:
 >Mr. Carey (being back on the EM list) wrote-
 >
 >I request, under the guidelines of the list, the probability of this
 >  example ever occurring:
 >
 >   99 ABCDEFGHIJ
 >    1 J
 >---
 >D- How about the probability of the below ???
 >
 >2 AJ
 >1 J

Under the Approval Vote, candidate J wins with 1/3 of the vote.
Note that Mike was careful to not say that A should win a 2
candidate election if it has 2/3 of the vote. We should assume
that Mike believes (believed) that A should lose and that he felt
that the example should not have been posted to the list in a
way that suggested that the Approval Vote was flawed.

 >
 >or ---
 >
 >3 AJ
 >1 A
 >2 J
 >1 C

Under the Approval Vote, candidate J wins with 2/7 of the vote and
candidate A loses with 4/7 of the vote. The principle of truncation
resistance which Mr Mike Ossipoff rejects, says that A wins the
above if and only if A wins this:

  4 A
  2 J
  1 C

Mikes believes that J should win if 3 papers are marked (AJ) and
that A should win if that is altered to (A).

 >
 >Add zeros and added choices (such as 500 CJ) for larger examples.
 >
 >No need----
 >
 >Approval is defective by treating all actual votes cast as being the same
 >(i.e. unranked) when in reality, of course, voters rank choices (if given
 >that option).
...

==========================================================================

Now here read Mike Ossipoff's reply in defence of what seems to me to
be the world's lead credible preferential voting method that gets
advocated in this mailing list.


At 05:37 01.06.00 +0000 Thursday, MIKE OSSIPOFF wrote:

>-
>>D- How about the probability of the below ???
>>
>>2 AJ
>>1 J
>>3
>
>In the above example, the A voters are voting for their last
>choice, something that no one would ever have any reason to do
>in Approval.

Mike's statement above appears to be false. Voters can have reasons
and Mike is not right to restrict the thinking done by unknown
voters. Mike made a false statement above and it looks as if I am
expected to see it as true.

Suppose there are 4 cockroaches under Mike Ossipoff's house. What if
they were put under an artificial vision system and allowed to vote,
with software providing the ballot paper counts to another subroutine
that calculated the Approval Vote winner. Mike Ossipoff told the list
in May that his 'roaches wouldn't do that. What if one was in pain, or
couldn't think with clarity?. Where is the reasoning for Mike's
conclusion that he knows what voters will think or want?. This is part
of Mike's defence to free himself from the hint of an allegation that
he knew the probability of an election.




>>or ---
>>
>>3 AJ
>>1 A
>>2 J
>>1 C
>>7
>
>In the above example, A has a 4/7 majority, but his voters
>apparently belive that they need to vote for J, because they
>think that C will get more votes than A. But C only gets 1

Why is the word apparently used?, when there is no data at all
allowing the word "apparently" to be appropriate. It is all a
fiction. I don't what is apparent about the motives.

To the readers: fill out this Approval Voting paper (or a similar
one):

A: [ ], B: [ ], C [ ]

Be careful: You might be a person Mike was thinking of when
he wrote that you will "think that C will get more votes than A".

How did Mike know that. Again no reason, just an assertion.

---

>vote, making C the lowest votegetter. Again, the example is one
>in which the A voters have made an implausibly big mis-estimate.
>Especially since previous election results are available and
>give an accurate estimate of the strength of the various factions
>or parties or candidates.

How can voters that made their 1st preference be for A, be making a
"big mis-estimate" when they carry a weight of 57.14% and your
Approval Voting method makes A lose?.

Mike can't simply say that the voters are wrong: there may not be
any voters: in fact there might only be papers. It was Mike who
introduced the idea of voters, and I, and perhaps Demorep, can't see
any argument at all that makes them part of a defence that can
protect the Approval Vote.

Anyway, has Mike got any other criticisms of the supporters of
candidate A ?. Are these complaints or criticisms?. Suppose all the
A voters denied that there was a mistake. Mike is contradicting
the experts, say in highland Chile. How does that defend the
Approval Vote Mike?.

At this part of Mike Ossipoff's document, Demorep's question on the
probability of this election, was getting fully ignored:

   2 AJ
   1 J

That could be a 2 or 3 candidate election. Mike would seem to want to
say it is OK in a 3 candidate election to have J win over A (3:2), but
it is just unreasonable to have J win over A (3:2) in a 2 candidate
election because the last preference is being specified. ((Does Mike want
to amend his definition of the Approval Vote?, now that it appears to
be a method that embeds well in one example, but it believes it
shouldn't.))


The probability defence of Mike was never retracted.
Me and Demorep mainly wanted an answer to our questions, and I thought
Mike was seeking to make an explanation that was extremely dull and the
right thing was instead to reject forever, the Approval Vote.

The Approval Vote could have been rejected but that attempt failed
seemingly because the questions were very clearly evaded. I suggest to
Mike that he holds an "ulterior motive" that prevents him from conversing
with this mailing list on the Approval Vote. He refused to state a
probability on the likelihood of the example.
Mike: was it bigger than 0.5 or smaller than 0.5 ?? [I of course refer to
Demorep's question about the 2:AJ,1:J election paper ratio.]

Mike told the list he could run into a lack of enthusiasm by councils.

He wrote suggesting he wanted to have a bit of success in recommending
the Approval Vote. Why the aim of success when the methods recommended
will make a candidate that should lose, win rather than lose? (in a
1 winner election).


>>No need----
>>
>>Approval is defective by treating all actual votes cast as being the same
>>(i.e. unranked) when in reality, of course, voters rank choices (if given
>>that option).
>
>And wouldn't it be nice if those rankings would always be sincere.

Sincerity is unknown and irrelevant. That has been resolved to my
satisfaction. Mike knows it too: all voters can vote for A and all
papers can have no 2nd preferences and all preferences can be for B.
It is not "nice" at all to always want voters to be since: instead their
sincerity is absolutely irrelevant. The sentence of Mike should certainly
not have contained the word "always". Whatever methods Mike uses to
create such a high concentration of totally false sentences per
paragraph ought be a topic of interest to the mailing list. Writing
down definitions and putting them online and then ignoring them is one
way for Mike to make mistakes. E.g. FBC passes the Approval Vote. When
I read FBC I find that it fails the Approval Vote. When Mr Ossipoff
said FBC passes the Approval Vote, he might not have been reading his
own definition.

The definitions are dumb and especially undefined. If FBC fails the
Approval Vote and Mike claims assertively without referring to the
definition that FBC passes the Approval Vote, then what is the actual
purpose of ignoring the rule?. Does he regard his own rules as being
threats like questions from Demorep and hence will refuse to follow
fairly and honestly their exact wording (which can't be done at all
well since they are undefined excluding maybe monontonicity).

Why didn't Mike just answer the question of Demorep on the probability
of the example?.

In October, Mike claims to recommend the Approval Vote.
That is the method that makes A win this 1 winner election:

           999,999,999,999,999 ABCDEFGHIJ
                             1 J

(The last preference could be empty or filled).

Subscribers will recall that Mike Ossipoff defended the Approval Vote
using something approximately named "Utility theory".

That example plainly seems to get the wrong winners when the required
number of winners is 2 and 3, also. For 2 winners, the Approval Vote
would elect candidates J and A. Candidate J could be Mr Ossipoff
himself and he could be voting for himself. The other
999,999,999,999,999 voters might have been a single voter with a
relative weight matching their real ability to answer simple questions.

I have asked a question of Mike Ossipoff.