[EM] Mr Simmons: provide the FBC definition please

[Craig Carey]

Lying again, are we?.


From:  Forest Simmons <fsimmons at p...>
Date:  Sun Aug 26, 2001  11:14 am
Subject:  Re: [EM] Three Tier, Dyadic via CR, etc.

 >
 >We were launched on this particular search of elimination methods
 >satisfying the Condorcet Criterion by a desire to fix IRV in a way that
 >would be palatable to the same folks that now support IRV.
 >

Perhaps the search will find nothing. Mr Simmons is giving an aim of
finding something that does not exist. It would help here if Mr Simmons
would comment on (1) monotonicity and (2) that property of STV that
has the win-lose state of a given candidate, unaffected by changes
after (and not including) its preference. Will (2) be lost and (1)
worsened?. The word "fix" seems plain enough to me, but I presume that
their it is, to the degree it is not undefined, an ideal of making IRV
somehow worse.

One problem IRV people may have is that they sense goodness with less
perceptiveness than government officials can. It is like a black
universe that this list's typing monkeys could start to fill: what is
there in Condorcet that can make any method appear to be better to an
individual voter that wants a method to be free of obviously dogy/wrong
behaviour. What happens here is that chat lasting for years results
in no complete consideration of the 3 candidate problem. We could say
that all of the 9 or 15 paper simplex that holds all 3 candidate
methods has to be searched through. Mr Simmons used the word "search",
but perhaps the simplex won't be searched at all.

The search was restricted to elimination methods and I presume that
that is a extreme restriction that imposes bad properties except that
increase information flow between stages can make a method with more
stages than candidates be a method that has the 2 numbers comparable.

An elimination method is one that removes all vertices of a candidate
in a single step. That can be an extremely large number of vertices
that vanish in a single step. It is remarkable constraint and not one
that is natural (but it is a known idea), and factorising is something
that is naturally done. Suppose the IRV people say that the improvement
has to be so great that the perhaps quite wrong restriction prevents
the attaining fo the improvement of a size that is large enough?.

How do you choose these constraints (i.e. 'a method shall have stages
where elimination of k candidates occurs'), Mr Simmons?.

I presume that Mr Simmons can't produce a plot for the papers
(AB), (B), (C), or for (AB), (A), (B), (C), that shows that there is
any reason to imagine the pairwise comparing style ideas will be of
use in improving IRV. It has been the history of this list to
consider ideas that can't make it past the simplest 3 candidates tests.

Mr C. Layton's "Fluffy the Dog" example seems to pose a problem for
people that could see Condorcet as plausible,

What about pairwise comparing's ability to properly implement vote
bartering and perhaps bribing, in hierarchies of factions in elections
with hundreds of candidates. Does someone want to tell us how restricting
consideration to only 2 candidates at a time, could possibly allow the
method to properly simulate the real power disputes between votes
representing factions of candidates, with each faction having at least
100 candidates. Who said the comparing had to use summing to subtotals
when there was truncating after just 2 candidates and that is done for
all nc * (nc - 1) / 2 combinations. The readers of the list might be
able to regards the lists prolific pro-Condorcet idealists (Schulze,
etc.) as part of a community of monkeys in a zoo. All that steel that
locks them in (in the picture) could not have a correspondence with any
outer world because Condorcet gets rejected.



 >We're going to have to simplify eventually if we want this journey to
 >end up there.
...

Simplify which equations, Mr Forest Simmons?. Given the problem with
disagreeable statements and no equations about anywhere, what is
needed by Forest Simmons would be a lot more complexity in the exact
treatment and transformations of the quantifier logic rules.

Perhaps subscribers have their own beliefs about rules (like aquarium
fish) and they could put on show. These fish would be shaping the
course of civilisation that is exhibited inside cities and it would
be nice to see what the rules exhibit. IS there a consensus that the
fish that would say "right number of winners" is missing from so many
people's collection of axioms, that the list can agree that methods
that pick the wrong numbers of winners are OK. To not make "right
number of winners" an axioms makes the mathematics very much more
complicated.



 >Comment:
 >
...
 >That's one reason why we have to pit the minmax loser against the
 >Copeland loser: i.e. to make sure we don't eliminate the CW if there is
 >one.

 >
 >Unfortunately Copeland elimination by itself is not good enough to
 >guarantee the Ranked Pairs winner in a three way contest.
 >



At 01.09.04 13:32 +1200 Tuesday, Craig Carey wrote:
 >
 >
 >Mr Forest Simmons wrote this. Since [he] was not replying to my request
 >for information privately I have rejoined to ask the question in this
 >list. What exactly is the equation of this concept called "FBC" ?.
...

FBC was never defined by Mr Ossipoff. Usage of it could involve a tariff
before the far side of a better method could arrived at.