[EM] Dan W-J; IRV is 2nd best (=worst); CVD tells Maryland 'no research ever'

Craig Carey research at ijs.co.nz
Wed Dec 24 12:17:02 PST 2003

[1b] The Centre of Voting and Democracy told the State of Maryland that
   it does not do research into the best options. Instead it interests
   itself in equal voting minus the fairness-for-other-individuals parts
   or something

[1a] The Alternative Vote (1 winner STV) is 2nd best.


This can be named the Alternative Vote Twin Towers example: a support
rise that is nearly 50% in size and that is entirely FPTP papers, is
occuring while the candidate ('A') being supported changes into a


|Winner: · ·A· · ·|Winner: · ·B· · ·|  Alternative Vote winners
| Tot: · 513000· ·| Tot: ·1024001· ·|
|· ·A· · · 1001· ·|· ·A· · 512000· ·| Change = +510999
| * BA · · ·999· ·|· ·BA · · 1001· ·| Change = +2
|· ·CBA· · 1000· ·| * CBA· · 1000· ·|
|· ·DBA· · 2000· ·|· ·DBA· · 2000· ·|
|· ·EBA· · 4000· ·|· ·EBA· · 4000· ·|
|· ·FBA· · 8000· ·|· ·FBA· · 8000· ·|
|· ·GBA· ·16000· ·|· ·GBA· ·16000· ·|
|· ·HBA· ·32000· ·|· ·HBA· ·32000· ·|
|· ·IBA· ·64000· ·|· ·IBA· ·64000· ·|
|· ·JBA· 128000· ·|· ·JBA· 128000· ·|
|· ·KBA· 256000· ·|· ·KBA· 256000· ·|


I doubt I can get this edited up properly. I am not sure but I may have
left too much in or out. I will CC this off to the CVD lawyer.
Presumably he has a stake in the CVD ideal of knowing far less than
very many people, about the stably named Alternative Vote [I won't
name it STV]


Since about 24 July 2003 the Instant Runoff Message of the
Maryland Center of Voting and Democracy, has been it has been
hooked onto a 1 winner preferential voting methods that is
NOT the best.

On that date I concluded about 5 days research and showed that
another 1 winner method was far fairer, i.e. more monotonic.

Readers may read the old message of July here:

    From:   Craig Carey <research at ijs.co.nz> 
    Date:  Thu Jul 24, 2003  10:44 pm
    Subject:  IRV measured: it is over 2x unfairer (than 1/3 quota AV)

It did not announced with a bigger show since I was writing on
San Francisco in those months a better method may be something that
is not good enough.

A search at the "Charities & Fundraisers" part of the Maryland state
website shows what it told the government that its purpose was:

|  "Purpose: Contributions will be used to further public
|   understanding of electoral rule changes and to support
|   efforts to improve elections and representation through
|   propretional representation and instant runoff voting."


I will comment on the CVD written purpose given to the Maryland
government and estimate if they said they free themselves to promote
better voting systems than IRV.

(1) US "public understanding" is not mathematical.

(2) "rules changes" permits changes that lock out better possibilities.
   It also allows changes for the worse.

(3) "improving" US elections could be a sloshy test that lets in
   better methods that are 6x worse.

(4) A purpose to act around the Instant Runoff permits them to reject
   all better preferential voting methods. Doubtless they will do that
   without complaining a lot.

(5) The words representation and proportional are matching up with
  only the first of these two. IRV would passes (5a)'s rules and if
  the CVD is ignore (5b) (is the case) then its text ignores the
  case implying that IRV ought be improved:

 (5a) Summing right. Two cases:
   * Outside of preferential voting: the same percentages (maybe with
     added biases but not multiplied-in biases)
   * Inside of preferential voting: Block Vote style naming of a
     candidate causes its subtotal to be incremented by the weight
     of the paper.

 (5b) Strict rules protecting voters from dumb or unfair behaviour
   of the preferential voting method while they are doing "what if"s.


That seems to be extremely clear: the Centre of Voting and Democracy
gave awful principles to the state of Maryland that just make it
plain that it has no official intention of ever offering to any
city, another awful but clearly better option than the
Alternative Vote. If they say that they will do research and nothing
happens for decades (should be enough time to get San Francisco to
flip over to the England's Alternative Vote) then maybe the tax
exempt status would need defending.

"Public understanding" may mean training up Green youths on 1-2-3-ism
or something.


The pictures in triangles show that a 1/3 quota would be just the thing
lead to a pass under a monotonicity test (if 3 candidates):

The method I tessted in just inserts the 1/3 quota in a partially
incorrect way, and compliance with monotonicity is only better and the
rule is still failed.


This is now defined to be the
    Rob Richie 1,2,3 preferential voting method,
    (a variant of the Alternative Vote):

(1) Eliminate all but the best 3 (while ignoring preferences except the

(2) [a>(a+b+c)/3, =(a>(b+c)/2)]
   Of the 3 remaining candidates, if only 1 candidate has more votes
   than the average of the other 2, then that candidate wins (i.e. the
   other 2 are under the 1/3 quota).

(3) Otherwise the Alternative Vote is used.

There are the 3 steps, hence the "1-2-3" or "123".
Remembering why IRV is the worst preferential voting methad that a city
is probably going to be sucked into using. can be as simple as 1-2-3.
(In New Zealand the Green party co-leader wrote it was as easy as
counting from 1 to 10, somewhere on greens.org.nz).


The CVD (a strictly unjust personnel selecting algorithm promoter) actually
has a lawyer (who is a "field officer"):

    [Mozilla 1.5 says the HTTP dates the page at 18-Nov-2003, 04:37]

| General Counsel
| Dan Johnson-Weinberger
| Dan Johnson-Weinberger is the Center's General Counsel. He is licensed
| to practice law in California and Illinois.
| Dan is a 1997 graduate of the University of Illinois (B.A. in
| economics and political science) and a 2000 graduate of the University
| of Chicago Law School.
| Dan has been published in the Chicago Tribune, Chicago Sun-Times, Roll
| Call, The Hill, the Daily Herald, the Chicago Reader, the Daily
| Southtown and the Miami Herald. He has appeared on National Public
| Radio and is an occasional guest on the nationally syndicated radio
| program Beyond The Beltway.
| In the last year and a half, Dan has spoken to classes and groups in
| Florida, Wisconsin, California, Illinois, Indiana and New York. He
| encourages invitations to speak to college classes or political
| groups.
| Dan spends the majority of his time in Chicago where he focuses on the
| drive to revive cumulative voting for the Illinois House of
| Representatives. He also spends time in Springfield, Illinois, where
| he lobbies the Illinois General Assembly, and was the main driver
| behind passage of a law that allows county boards to give cumulative
| voting rights.
| Dan helped to found the Midwest Democracy Center in 1995 with a group
| of Illinois electoral reformers (www.midwestdemocracy.org ). He worked
| on several political campaigns and enjoys walking precincts.
| Dan can be reached at djw at fairvote.org or by phone at 312.933.4890.
| His website is www.djw.info and his blog is www.djwinfo.blogspot.com
| (opinions expressed on his sites are his own and not necessarily
| shared by the Center). Inquiries are encouraged.

Mr W-J liked the Cumulative Vote.
(I.e. if voters should fill in 7 votes but fill in 2, then the paper's
weight is multiplied by 7/2, etc., etc.)

Seven-winner STV (say), is similar except it uses the weights from 1 to 1/7
instead of from 7 to 1.

The alternative is the Block Vote ('k seats, tick k checksboxes') which
does not multiply by a rational number.

A way to improve STV is remove the transfer values that make STV follow
the idea of the Cumulative Vote.

It would be replaced with the style of the Block Vote: no dividing by
a denominator.

Dan W-J got his CVD webpage to fail to say he liked IRV. However it says
he promoted the Cumulative Vote in Illinois. A way to improve the
proportionality of STV is hack out the Cumulative Vote theme in it.

I say that progress here has no analogy with Dan's interests (for
Illinois: decisions on Cumulative Vote are probably best made decisions
for local decisions).


Some story using model trains can supply an analogy to the plan to
remove transfer values from STV. Suppose the election elects 4 winners.

(1) Old STV case:
  Counting a weighted ballot paper, is like getting a model train up to
   speed as it rises over 1 track with 4 different slopes.

(2) STV replacement case:
  The paper can start 4 trains that each have a track with only 1 slope.
  Now the power<=1 rule can be failed more quickly.
  However some computer algorithm controls when and where the track
  bends and become horizontal and power is cut.  By some means, the
  rule is not violated, and the tracks have only a slope corresponding
  to a transfer value of 0 or 1.

(The idea of the description could be improved.)


Topic: IRV vs the United Nations Covenant on Civil and Political Rights [of 1976]


| ...
| Article 25 General comment on its implementation
| Every citizen shall have the right and the opportunity, without any of
| the distinctions mentioned in article 2 and without unreasonable
| restrictions:
|   (b) To vote and to be elected at genuine periodic elections which
|      shall be by universal and equal suffrage and shall be held by
|      secret ballot, guaranteeing the free expression of the will of
|      the electors;
|   (c) ...

That UN text really seems to be not designed to eliminate IRV.
It does not split up maximal power voting rights up into 2 parts
(proportionality and some 0<=power<=1 rule(s)). Alternative maybe it
could be said that I have not got the 2 unified (it seems hard to do).

However when the IRV of the CVD is extrapolated into the UN rules,
it fails badly.

(Recently I posted up an example showing the cascading from some
probably nobody at the bottom, and at each cascade the quantity of
votes shifting doubled. It was losing FPTP votes. [Refer to: the 2003
messages with 49.9% in the subject fields].


Here is the CVD list of donors: http://www.fairvote.org/about_us/index.html :

|    Supporters of the Center
|    The following foundations have provided grants to the Center for its
|    programs. Those foundations with asterisks have provided support for
|    activities in 2003 and beyond.
|    * Arca Foundation
|    * San Francisco Foundation  http://www.sff.org/
|    * Deer Creek Foundation
|    * Solidago Foundation
|    * Ford Foundation
|    * Stern Family Fund  http://www.sternfund.org/
|    * HKH Foundation
|    * Stewart Mott Charitable Trust  http://www.srmtrust.org/main.asp
|    * Joyce Foundation  http://www.joycefdn.org/
|    * Vanguard Foundation  http://www.vanguardsf.org/
|    * McKay Foundation  http://www.mckayfund.org/
|    * Working Assets  http://www.workingassets.com/
|    * Open Society Institute  http://www.soros.org/
|    * Z. Smith Reynolds Foundation  http://www.zsr.org/
|    * Rockefellers Brothers Fund  http://www.rbf.org/
|    * and individuals from across the country

Not very interesting so far.

Since there was no link to the Ford Foundation website, I checked their
Grants database. It seems that the Ford Foundation stopped funding the CVD
(it is not perfectly clear that it did actually stop, but it seems
extremely likely).

The CVD got US$80,000 in 2002 AD.

Reference: http://www.fordfound.org/  : Grants section:
Under: "Peace and Social Justice / Governance and Civil Society / 2002"


A fairer IRV election has got more glass (or plastic) eyes that read
some extra checkboxes. If the Alternative Vote is used then the city
ought have a less proportional method if it gets something fair.

More information about the Election-Methods mailing list