[EM] FBC and undefined definition (was Demorep: Technical Words

[Craig Carey]

I believe that FBC is a dud of an idea so long as it applies to a
  single preference named the "the favorite". However, Mr Ossipoff
  wrote about "precise language" so a few extra doubts added to the
  ones I made earlier would not be out of place.
  


At 16:01 07.03.00, MIKE OSSIPOFF wrote:
 >
 >Demorep:
 >
 >I don't know that any of our proposals to the public contain
 >the technical words that you were objecting to. We all know
 >that proposals to the public should be expressed briefly
 >in plain language.
 >
 >But that doesn't mean that we shouldn't use whatever words
 >are needed when discussing methods' relative merits here.
 >
 >Also, criteria can be written in precise language, available
 >to be checked on by anyone who wants more than the plain-language
 >statements that we'd use to tell people what a method can do,
 >or what it won't do, or what it won't make voters do.
 >
 >For instance, I might tell someone that Approval will never
 >give anyone incentive to vote a less-liked candidate over their
 >favorite. If someone wants that in precise language, then
 >they could be referred to a more carefully-worded statement of
 >FBC.
 >

I have not seen any carefully worded statement of FBC.

I wasn't going to write. FBC is not one of the more interesting
  rules, except for its resistance to being defined (rather than
  being improved).

FBC is a good example of why this statement has to be presumed
  false until shown true:

   "criteria can be written in precise language, ... If someone
    wants that in precise language [then they only need request
    that at the Election Methods mailing list]"


---------------------------------------------------------------
At 17:01 07.03.00, MIKE OSSIPOFF wrote: ...
 >>At 12:23 05.03.00 , MIKE OSSIPOFF wrote:
 >>...
 >> >>FBC is contrived valueless overly weak rule.
...
 >>At 12:23 05.03.00 , MIKE OSSIPOFF wrote:
 >>...
 >> >> >> >> >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.]).
 >
...
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)

What is the maximum number of papers that a voter can cast?.
Your not defining "Better" suggests you might want that to equal the
  number 1.

...
 >... "Better" isn't undefined. It's in any dictionary of English.
 >I don't intend to define for you every English word that I use.

The dictionary meaning of the word "better" is a description of a
  quite different item. When two different ideas/objects have the same
  name then there is no proof or suggestion because of that, that two
  distinct ideas/things are identical. Why refer to a dictionary?.

  On how many preference lists is the "liked better" term dependent ?:
  one or two?. This is not a "participation axiom" type rule (i.e. one
  that Mr Ossipoff calls a similar idea, SARC), which is dependent on one
  preference list. (I.e. rules about the removal of votes and all their
  preferences lists).

Why did you suggest I should understand the word better: to suggest
  that voters ought too?. Some voters will refuse to acknowledge any
  opinion on the candidates.

A question for readers, in my opinion, is: do they concur that it
  is obvious, that their opinion matches up with Mr Ossipoff's on
  the fact of, whether or not, the "likes better" function is:
  (a) dependent, or (b) independent; of the 2nd preference list, i.e.
  the list in the state after the change ("than any outcome that he
  could get without voting a less-liked candidate...").

 >the
 >>  preference list before or after the "voting" ? (i.e. the inserting of
 >>  the preference, if any. If a text on any attached statement writen by
 >>  the voter is followed, then it will 'before inserting' [or "over"-ing]).
 >
 >I have no idea what all that is supposed to mean. If you don't
 >know what it means to like one candidate better than another,
 >then I can't help you.

That is a statement that suggests to me that you think the matter
  is simple, when in fact it is not. If you want the 'likes' of this
  function to be simple, then some constraints will be needed (e.g.
  1 winner, or undefined for any method except the Approval Vote).

That is no definition of the Boolean function "better". I asked for
  a definition and no definition was given.

The idea of "better-ness" is not one that readers can guess at.
In earlier messages to the mailing list, I introduced my binary
  satisfaction formula. The response being made suggests to me that
  Mr Ossipoff does not understand the how complex the considerations
  of the readers must be when trying to reconstruct what it can only
  be that Mr Ossipoff must have been thinking of. Now, it is not clear
  that Mr Ossipoff has been thinking clearly. It is likely that I
  need not try to fix this FBC rule, because a decent fix would have
  to alter FBC by a lot. A dispute presumably would result, and in any
  case, it should be possible to get Mr Ossipoff to stop falsely claiming
  the Approval Vote gains something from FBC. Couldn't that be deferred
  until FBC is defined?: apparently not.

  If Mr Ossipoff doesn't define "better", then correspondence could
  perhaps take up another 27K. I presume few others will write as I will,
  so Mr Ossipoff might be able to terminate the discussion on FBC without
  having ever defined using the criteria he set out. Of course, readers
  will be aware that I can easily have a new FBC version defined in just
  5 lines of text.

What is the idea of the function "Better":

Is it independent of the preference list that is in the 2nd (the
  after) election state?. Can you admit that FBC is a mathematical
  item at all?. Is it a function that fully ignores opinions?.

The word "any" is still there. In English, sometimes it means "for
  all", and at other times, "there exists". My opinion, is that there
  is a real and unnecessary lack of clarity due to the use of the word
  "any". Couldn't the definition be improved?. 

Mr Catchpole, for example, tends to edit his definitions in those
  arguments where he was indeed referring to his written definition.

...
 >If you have a serious objection or claim that you'd like answered,
 >please post it concisely and clearly in a short posting. I have

The difficulty I am having in getting you to take FBC seriously,
  and why not, given that this list has been told that it supports
  somehow the Approval Voting method, has made my long message
  appropriate. All I am after is a definition.


Consider this example which is a little irrelevant:
  Suppose a voter was considering changing this:
      (1:AB,2:B,     <others>; B wins)
  into this:
      (2:AC,0:B,1:C, <others>; C wins)

Suppose P1 said B continued to lose if AB changed into AC and if 2:B
  was changed into (1:AC,1:C). A small detail there is that a number 
  became negative but it was concealed, i.e. (1:AB) changed into 
  (-1:AB,2:AC). Is the FBC definition of "likes" similarly marked with
  difficult to initially see obscurities?. A definition of FBC, sent to
  the list, could remove uncertainty from the minds of subscribers
  about what it is that FBC is, exactly. If they knew what FBC was,
  then possibly they would know what it was that they liked, given the
  arguments Mr Ossipoff has already made.

Mr Ossipoff: do you want a contest where subscribers try to improve
  it?: have it apply to multiple preferences simulaneously, etc..

...