[EM] Losing with 25% FPTP; CVD, Richie: monotonicity lost if missing 2nd prefs
Craig Carey
research at ijs.co.nz
Sun Mar 24 11:53:03 PST 2002
Greetings Election Methods List members
I have a mailing list named the instant-runoff-vote mailing list.
If any wants to take it over then they may write to me. A likely outcome
is that the list closes.
http://groups.yahoo.com/group/instant-runoff-vote/messages
It has 7 members in it which is not bad for a mailing list on the reform
of the lives of councillors and mayors in the USA.
Much of the message is an attack the US spokesperson advocating the
negating the will of the public, and corruption inside of preferential
voting [ref example below], Mr Rob Richie. A brief contact allowed
Mr Richie to conclude he evidently knew what I did about IRV. I myself
found out that the CVD has really not go an interest in the Alternative
Vote, sort of whatsoever, but it is a large agency hinted the 'about'
page (it ha donors: a <<"good">> voting method costs extra).
This is the final sentence of the letter from the leader, Rob Richie.
------------------------------------------------------------------
>
>No, you and I know from past experience that dialogue
>between us isn't worth the time.
>
>- Rob
------------------------------------------------------------------
PS. there has not been communication or dialogue between me and
Mr Richie. It am sure that that is the message I got where the name
appeared at the bottom. What a terrible idea for a dialog. If I guess at
the meaning right, it means a communication between two parties that
has to be protected from the public or else scandal is a risk. I can't
say that everything I know was learn't from Mr Rob Richie.
I want to fully agree with Mike Ossipoff's posted comments saying he ought
work more and what would of interest is that practice with at quantifier
eliminator be used.
Here is a copy of a private message I received from Rob Richie.
Roughly the procedure for the ethical conduct of testing that the search
has to restart when the worst case bug is found.
-----------------
Since this is disclosure of private e-mail that is full of clear
errors and not built to withstand public scrutiny, then please have a
sensitive view to Mr Richie's comments [a R.R. negating of the intent
transform should not be applied to my comment]
------------------------------------------------------------------------
At 2002\03\23 07:52 -0500 Saturday, Rob Richie wrote:
>Craig,
>
>I would think you would know that you can generate
>whacky results with any system. Your example below
>of course is political nonsense (why would every B
>voter rank C second, but no C voters rank B second),
>just as some of the examples that can be used to
>mock every system you would propose.
>
>No, you and I know from past experience that dialogue
>between us isn't worth the time.
>
>- Rob
>
>
OK, that is just in a plain Ossipoff-ian style (seriously flawed at a
deep level and it could have been improved by search for problems and
finding out why they are there without the slightest use of declarations
of views).
>Message text written by Craig Carey
>>
>
>
>Dear Mr Rob Richie
>
>Subject: 25% of Republican votes go missing in IRV
>
>At 2002\03\09 23:19 +1300 Saturday, Craig Carey wrote:
>...
>>
>>The Alternative Vote
>>
>>· · · · · · · 1st · · 2nd
>>· · · --------------------· · · ·change
>>· · · ·AB · ·3333 · ·3333
>>· · · ·B· · ·3332 · ·3334 · · · + · ·2
>>· · · ·C· · ·3335 · ·6666 · · · + 3331
>>· · · --------------------
>> AV Winner: · C · · · B
>>· Total:· · 10000 · 13333
>>· · ·IFPP: · · C· · · ·C
>>
>>
>>
>>The number of candidates is increased by 33.33%. The voters (papers)
>>can show a 99.939994% purity of intent to have Mayor "C" elected. The
>>Alternative Vote denies about 1/4 of the electorate a right to have a
>>non-negative effect to their vote.
>>
>>The Alternative Vote gets the 2nd election wrong (i.e. different from
>>IFPP), rather than the more closely balanced first 1 winner election.
>>
>>The last 25% have their overall vote negated: 3333/13333 = 24.998125%
>>
>>[In the 2nd election, IFPP has B lose since 3334 < (3333+6666)/2 =
>>4999.5.]
...
I missed the bug. It is remarkable for having only 3 papers and having the
changes involve on FPTP papers.
Had Mr Richies said that the outcome is obviously improbable then that
could be argued to be having the viewpoint restricted to a lattice
placed throughout the polytopes. Any "its obvious" argument would not
hold up for small numbers of candidates and the person arguing would
be found wrong if not all of the normal vectors had to be constrained
by whatever the rule was. However voters were said to have put the wrong
marings. Whatever.
Mr Rob Richie is in error. Obviously he should never criticise voters
before I get around to naming the country, which was the fact.
Imagine that B voters hate candidate A, and A's voters hate C.
What about this: by P2 all the (B) papers can be changed into a sum of
(BA) and (BC) papers. It is wrong to consider the reciprocating of
2nd preferences between A and B if C is to be ignored. Only through
being inexact can the CVD make such a concentrated collection of errors
using very few words. Anyway it looks like the door on the matter and
USA's destiny is sealed since it is allegedly as waste time to
communicate with. The CVD got a tip off from me last year sometime that
IRV is a going to be show corrupt once I get around to it. But
-----------
There is another problem. If the suggestion I got is followed then the (B)
papers are replaced with (BA) papers.
I am not free to change the (B) papers into (BA) papers. There is an
improper suggestion to me to alter the example. That is a request to have
the outcome be biased. I am not changing the example.
Any change to the example makes the new example irrelevant.
Furthermore, any changing of the (B) papers into (BA) papers makes IRV
get both winners wrong. That seems to be actual argument. The winner of
both ought be C but Richie would free IRV from the allegation implying
it was picking the wrong winners. having A win the 1st and B win the 2nd.
Another problem is that once a 4th paper that is a (BA) paper is allowed,
then still in the tetrahedron the method is showing nonmonotonicity.
What else could the fact given that Mr Richie has opted for having this
equation say when A wins:
(A wins) = (b<a)(b<c)(c+Zbc<a+Zba) or ...
Now the triangle was changed into the tetrahedron and Mr R proliferated the
nonmonotonicity by adding another vertex. However incorrectly assumes that
took the hint on needed to alter the preferential voting method (achieve a
biased outcome like the book or ID Hill and R. Richie says that women and
others can get if they are elected under [AV perhaps]
----------------------------------------------------------
Lets investigate and see if Rob Richie was attempting to use hints to
free IRV from the claim it was vote negating and that it got the
winners wrong and so forth.
>>· · · · · · · 1st · · 2nd · · 3rd
>>· · · -----------------------------
>>· · · ·AB · ·3333 · ·3333 · ·3333
>>· · · ·B· · ·3332 · ·3334 · ·3334
>>· · · ·C· · ·3335 · ·3335 · ·6668
>>· · · -----------------------------
>> AV Winner: · C · · · B · · · C
>>· Total:· · 10000 · 10002 · 13335
>>· · ·IFPP: · · C· · · B|C· · · ·C
>>
If we suppose the full power gold standard FPTP papers are just the type of the
virtuous trustable paper that Candidate C has got then we can see that
the 2 (B) papers of unknown power, have an influence equal to x (C) papers,
where 3331 <= x <= 3333
The power of the FPTP papers that B has is about 1666 times that of a plain
(C) ballot.
IRV is corrupt and there is presumably some aspect of the C papers having
too little power.
Nevertheless that can be quite real to a candidate.
Again that is another example that can be plotted in a triangle.
Though I have not got the way to calculate the size of the excessive power,
it is certainly the case that value to pin the wrongness of IRV's corruptness
at, is the number 1,666.
I got "0<=power<=1" equal suffrage (i.e. my P4) defined without defining a
power number.
It is a pity that I can't go back to the CVD and find out if they regard
the power excess, as exceeding 400 or something.
It will negate a support rise of 33% when some of the support rises on the
2nd preference. If the support would rise on the 3rd preference then an even
bigger fraction of the public is harmed. The IRV method does not have any
such bugs, so far as is known. I presume it ought be possible to have a change
from maybe 42% of the that only increasing support, with the IRV method
negating that support rise.
------------------------------------------------------------------------------
If there are 10 papers and the region where candidate A wins is of the
shape of a hypercube, then the CVD would slip across to the idea that an
attempt to spot-probe IRV in the search for more atrocious behaviour like
what I sent out to the fashioner of the lives of mayors and other political
candidates, may search trough the regions near vertices. A richer place for
human rights depriving bugs would be near the centres, I guess. But the
calculation gets the number large. For 10 candidates the table above shows
that the dimensionality is 6,235,290. It could be reduced by a factor of 10
by truncating after the papers of the candidate under consideration.
Number of vertices = 2 ** 6235290, = x
Here is that number in a more readable form.
log x = 6235290 * log 2
1877009.3216636653 = about 2*(10**1,877,009) vertices.
This is not worth more than a few moments (I am only after opinions,
including an opinion on whether a GIF of this could have pastel yellow
at the topic)
The assertiveness
of the claim, truth
/\
/ \
/ \
/ \
/ \
/ \
/ \
/ \
/________________\
IRV is IRV is good enough to be
good enough desired and its good
and passes are found by considering
just as cleanly as normal vectors alone.
a metal engineer's nut How the IRV method is
would pass through a the material expression
larger precisely of the desire of the CVD
measured hole to advance their knowledge
-------------------------------------------------------------------------------
To kill a bit of time, I make out some arguments on how IRV might
be defended against claims that the method is flawed. It is not just
the case that the CVD can't think of any good alternative to the
Alternative Vote, but actually when about 0.3 of journey towards
FPTP is traversed and away from IRV, a continent/realm of monotonic
methods can be arrived at. That is a statement about 3 candidate
elections. What the CVD needs to defend itself against is perfectly
monotonic and truncation resistant and P2 compliant preferential
voting methods that are maybe inefficient in providing
proportionality.
Here are my Three Pillars of reformers
PILLARS OF PREFERENTIAL VOTING REFORMERS
I. "Better is needed". and
II. "It is good enough".
III. "No voting system is perfect". . Or "no method is perfect".
Does anybody want the "instant-runoff-vote" mailing list. E.g. to
squat over the name. Possibly this might be a year when the YAhoo
mailing lists cease to exist..
Those interested in more complaints in the same style (which might not
be many) could browse here: http://www.ijs.co.nz/irv-wrong-winners.htm
The above may errors in it. As ever my proof reading is inadequate.
My new catchy phrase: A method as bad as that is a method that is actually bad.
I want to say that Americans are an inspiration for all us. IRV is
validated as good by comments vaguely saying that a bug should be overlooked
as soon as found.
===========================================================
More comments.
---------------
Rob Richie tried to get rid of a single example showing IRV wrong. Lets
jump to the conclusions and this could be wrong, and suppose IRV was
exonerated. Thus obviously Rob Richie has the agency to check out IRV at
millions of dimensions. This is not quite the same as millions of points
like happens with the human genome project and if the CVD is not
substantially ignorant of preferential voting them it must have an
awareness of the significance of the high dimensions and how difficult
they make it to pass IRV. AS soon as we say that the CVD is reckless and
in fact only advocates an unfair methods that does not provide the United
Nations principle of equal then there is no suggestion of untruthfulness
which suggested by the following but subject to that condition.
A method passes equal suffrage only when the slopes of a region where
candidates win, are tested and pass. The CVD is not showing any progress in
accepting that view, nor have I ever seen any counter. Every single
example proving IRV is unjust, etc., is one making if clear that right in
the defect area is a face with a wrong slope. The CVD won't ever get IRV
to be fair while it does not consider constraints on slopes. Surely we
can ignore Condorcet. IRV is not far from a method based on the casting
"if wins here then wins there" shadows. Either the CVD knows nothing about
IRV or there is a problem with why the CVD is not declaring that it has
declared a basic aspect of a principle. If a scientist is given a
microscope and a piece of paper and a pen, then success at getting the
view drawn onto the paper is seemingly able to be easy, How old is the
CVD. It has been around since 1985, and they do not even have the word
polytope at their website. Maybe they renamed the word polytope. This is
a Green political all sorts of omissions by accident might occur.
I don't care who it is in the CVD but if they will measure slopes in
6 million dimensions and check ever face in an attempt to permit a summarising
view of IRV's worth, then any hand holding a slope measuring device would
have to be used very fast. So the CVD worker may have to emit Cherenkov
radiation if actually checking out a single IRV method in a good time. To
be balanced here, it absolutely seemed to be that Mr Rob Richie was saying
that once the single point was countered then IRV was back to as it was
before. It seems incredible that Mr Richie did not realize that had to be
just as bad before the test. There was a missing constraint so it could
have been over 10,000 tourist aliens in a spaceship. If that is unreal
the Instant Runoff Method that can't be good in the Arctic and bad in
the Antarctic, is a method the CVD would tell us would be used in some
part of America. He forgot. Would any member here forget to add a
constraint.
So was IRV ever tested. The Cherenkov radiation emission argument was missing
numbers behind it. [I can't recall the spelling: it is the blue light that
comes out in a cone from particles moving inside a medium, faster than
light would].
Suppose the following....
If there are 10 papers and the region where candidate A wins is of the
shape of a hypercube, then the CVD would slip across to the idea that an
attempt to spot-probe IRV in the search for more atrocious behaviour like
what I sent out to the fashioner of the lives of mayors and other political
candidates, may search trough the regions near vertices. A richer place for
human rights depriving bugs would be near the centres, I guess. But the
calculation gets the number large. For 10 candidates the table above shows
that the dimensionality is 6,235,290. It could be reduced by a factor of 10
by truncating after the papers of the candidate under consideration.
Number of vertices = 2 ** 6235290, = x
Here is that number in a more readable form.
log x = 6235290 * log 2
1877009.3216636653 = about 2*(10**1,877,009) vertices.
It may require Superman to test. The CVD says IRV is used but that would
be different from testing IRV. It is not tested and it fails tests that
it ought pass. In response to the curiosity that IRV fails so badly: what
is their statement?.
So IRV has to be sampled at at least a number of points approximately equal
to the digit 1, followed by 1.877 million zeros. Such a figure would be
very inaccurate.
I guess the CVD folk can't say when the IRV method was tested and passed.
Instead of opposing the CVD something intellectual physically inactive
that would replace it, seems to be a small conspicuous option.
G A Craig Carey, Auckland New Zealand
More information about the Election-Methods
mailing list