[EM] So "preference" means "representation of preference"?
Craig Carey
research at ijs.co.nz
Wed Mar 5 05:44:04 PST 2003
At 03\03\04 22:37 -0800 Tuesday, MIKE OSSIPOFF wrote:
>
>Markus--
>
>You wrote:
>
>I absolutely agree with Blake: "A method is just a function from
>a hypothetical set of preference orders to a set of winners.
>Where the preference orders come from is of no concern."
>
>I reply:
>
>A method is a function from a set of not-so-hypothetical representations
>of preference order to a set of winners.
>
>Every conceivable voting system has as its input some sort of
>representation of voters' preferences or ratings of the candidates.
>
>Preferences and representations of preferences obviously aren't the
>same thing. You & Blake seem to be invoking the academic authors,
>as if their misuse of "preference" is official & correct usage because you
>found it in a journal.
>
>I'm sorry, but I've never heard of a voting system that uses hypothetical
>preferences, whatever that means. As I said, they all have
>as their input a not-hypothetical set of representations of preference
>ordering.
That sounds evasive. Why not correct Mr Schulze's error of not attaching'
weight information before put the information into a set, and then make
a correct reply?.
To Mike Ossipoff ("MIKE OSSIPOFF").
I invite you to answer these 2 requests for information:
Here is a method.
A method that Mike has "never heard of" prior to this message: a
2 candidate
-------------------------------------------------------------------------
In the following, "a" and "b" are real numbers representing the 'weight'
(or count) of the ordered preference lists shown. The thing that is a
preference list and a weight is a to be named a ballot paper.
Preference List : Weight(or count)
(A) : a0
(AB): a1
(B) : b0
(BA): b1
The election always has except when a=b'. The winners are:
(a0+a1<b0+b1) implies B wins
(b0+b1<a0+a1) implies A wins.
-------------------------------------------------------------------------
This would be easy for MIKE OSSIPOFF.
If Mr OSSIPOFF posts up algebra then I would let the rule be swapped
with another.
At the moment I ask that the 4 paper 2 candidate 1 winner rule be
checked with definition of "Favorite Betrayal Criterion" that is
at the webpage with the URL shown and also below:
QUESTION 1 !
Does this FBC pass the 2 candidate 1 winner [FPTP] method defined
above?.
----------------------------------------------------------------------
http://www.electionmethods.org/evaluation.html
|Favorite Betrayal Criterion (FBC)
|Statement of Criterion
|
|By voting another candidate over his favorite, a voter should never
|get a result that he considers preferable to every result he could
|get without doing so.
|
|Complying Methods
|
----------------------------------------------------------------------
The favorite is not the 1st preference because of the word "over".
Can the FPTP method be failed by your Favorite Betrayal Criterion ?.
QUESTION 2 !
(1) Please consider the last US Presidential election where Mr Al Gore
was running.
(2) Please regard that the FPTP method used in that election is
re-stated so that it has 2 ballot paper preprocessing stages.
The 1st stage converts papers to STV papers.
The 2nd stage undoes that conversion and it returns FPTP papers.
This is not less secret than the sincerity information you often
write a bit on. So it is hyper-secret.
Here is a diagram showing the real world details, as asserted or
whatever:
FPTP papers --[STAGE 1]--> STV papers --[STAGE 2]--> FPTP papers -->
--> Electoral College
---------------------------------------------------------------------
Update on new sincerity definition
MIKE's definition of "sincerity" now has it be undefined when the
very last preference is not filled in. These methods depend on that
definition of the word sincerity:
(1) Condorcet Criterion (CC)
(2) Generalized Condorcet Criterion (GCC)
(3) Strategy-Free Criterion (SFC)
(4) Generalized Strategy-Free Criterion (GSFC)
Here is the exact wording of the latest "sincerity" definition:
>A sincere vote is one with
> no falsified preferences or
> preferences left unspecified
>when the election method allows them to be specified (in addition
>to the preferences already specified).
The Condorcet variant methods are likely all indifferent to whether
the last preference is present or not. Yet CC and GCC contain the
word Condorcet.
Mike's own website is destroying 4 of the rules
Mike wrote to this list and said that the "favorite" was who I wanted
it to be. That subsequently got contradicted by Mike.
---------------------------------------------------------------------
At 2000\March\10 06:28 +0000 Friday, MIKE OSSIPOFF wrote:
From: "MIKE OSSIPOFF" <nkklrp at hotmail.com>
To: election-methods-list at eskimo.com
Subject: Re: [EM] FBC and undefined definition (was Demorep: Technical Words
Date: Fri, 10 Mar 2000 06:28:11 GMT
...
>>At 17:01 07.03.00, MIKE OSSIPOFF wrote: ...
>> >>At 12:23 05.03.00 , MIKE OSSIPOFF wrote:
>> >>...
Here is the 6-March-2003 version and below is the 29-Feb-2000 version.
|By voting another candidate over his favorite, a voter should never
|get a result that he considers preferable to every result he could
|get without doing so.
>> >> >> >> >> >Favorite-Betrayal Criterion (FBC):
>> >> >> >> >> >
>> >> >> >> >> >By voting a less-liked candidate over his favorite, a voter
>> >> >> >> >> > should never gain an outcome that he likes better than any
>> >> >> >> >> > outcome that he could get without voting a less-liked
>> >> >> >> >> > candidate over his favorite.
>> >> >> >>...
>>...
>> >>Can you write the idea "over" in a mathematical form?. Despite the
>> >> use of words, I can't understand the definition of FBC (because of
>> >> lots of problems with the wording: "over" [undefined, a preference
>> >> nearer the 1st?], less-liked candidate [refers to a preference
>>:> list?], "likes better" [no voter is around to comment so a preference
>> >> list (perhaps) describes 'likes'], any [can mean either all, or
>> >> some], favorite [must be a 1st preference or else described in
>> >> attached notes or etc.]).
>
>The only one of those objections that makes any sense at all
>is your objection about "any". Maybe "every" would have been
>less objectionable. "Every" expresses the meaning that I intended.
>However I'm not so sure that I agree that "any" is ambiguous there.
>
>Conceivably you could think that it meant "Any one particular outcome,
>that outcome being an outcome that he could get by..."
>
>But the combination
>"any....that..." is standard enough that I don't think anyone
>but you would be in doubt about the meaning.
>
>Anyway (Oops, I used that word again), if it will clarify the
>meaning for you, replace "any" with "every".
>
>> >
>>...
>>So maybe "over" means approximately:
>> x<>y =>: "x over y [in p]" = x in tr(p,y),
>>
>> (Using the "tr()" notation I used before, p is a paper, y a
>>candidate)
tr(p,y) meant "truncate the paper p just after the preference naming
candidate y, if y is named, otherwise leave it unchanged." [Could
truncate immediately before and there it returns a set.]
>
>No, that isn't what "over" means. I doubt that you know what
>your pseudomathematical diarrhea means. It's obvious to anyone
>that you vote Gore over Bush if you give Gore a higher position
>in a ranked ballot, or a higher point score in a nonranked ballot.
>Then, if your ballot is a ranked ballot, Gore's position is
>over that of Bush in the same sense that we use the term
>"ranked higher". Are you going to ask what "ranked higher" means?
>And if your ballot is a nonranked ballot then your point-assignment
>to Gore is numerically over your point assignment to Bush.
>I could have spelled out how you vote X over Y on every type
>of ballot, but the above seemed so obvious that there was no
>need to waste the space or lengthen the criterion in order to
>do that.
>
You will need to look at all of the ballot papers before you can
reject FPTP ?. Recently you said it was embarrasing to pass
FPTP by mentioning MMC. Why was there no advice to mention
FBC and MMC at the same time (?).
G. A. Craig Carey
http://groups.yahoo.com/group/politicians-and-polytopes
More information about the Election-Methods
mailing list