Proving B loses with heaps of votes (Re: [EM] Query for Approval advocates

[Craig Carey]

Correction: in my last message (partly quoted here), change the words

. . .!> "also wins V4B: V4 ="
to
. . . . "also wins this next election"

---------------
>(5) By (4) and the axiom of Truncation Resistance, B wins V4:
>
>V4 =
>   (AB.)  34.33%
>   (BC.)  28.57%
>   (C..)  37.01%
>
>(6.1) By (5) and the axiom/corollary of Monotonicity, candidate B also
!> wins V4B:
!>V4 =
>   (AB.)  34.33%
>   (BC.)  32.57%  [up 4%]
>   (C..)  33.01%  [down 4%]
>
>(6.2) By (6.1) and the axiom of Truncation Resistance, candidate B also
> wins V4B:
>
>V4B =
>   (AB.)  34.33%
>   (B..)  32.57%
>   (C..)  33.01%
>
>(7) By the axiom of the Right number of winners, candidate A loses
> election V4B (and candidate C also loses election V4B).
>
---------------

Here 4% of the vote moves from (C) papers to (BC) papers.

(Monotonicity leads to more wordy arguments, more complex algebra,
and use of before-hand ruled-out papers).

I am quitting the EM List (again; unless someone discovers something
soon).

Also Rob Richie sent in a message to the "instantrunoff" mailing
list (of groups.yahoo.com) that identifies a new category of
enemy (a category that contains "election officials", rather than
say, all candidates and all of the public).

--------------------------------
>Bottom-line: election officials at this point are unlikely to be your ally
>-- and even worse, will be your enemy. If they won't get out of the way, we
>may need to find ways to go around them even as we push to understand why
>they are so resistant to a reform that on so many levels is simply "good
>government."
--------------------------------

I would be reluctant to draw conclusions since the plan was to harm just
about every single person and then lose in the whole attempt is an
insecure basis for a precepts on how to handle enemies. He ought not
be categorizing anyway for that is not a use of reasoning, I suppose.

A peek at IRV's 3 paper equation can have it failed for about 3 seconds.
With REDLOG, some subroutine can fail it in less than a second.
Still no good news on the numerical front. That For Loops program I
wrote takes maybe 2-6 hours to get to a statisfactory conclusion that
IRV is perfectly failed with numbers estimating the scandalousness of
the defect.



C.C.     --   Notes on Preferential Voting:
 http://groups.yahoo.com/group/politicians-and-polytopes