[EM] STV's rejection: it's "not a defect, it's a feature!"

[Craig Carey]

I just wrote down some passing ideas on this Approval Voting method.


At 14:07 16.03.00 +1000, David Catchpole wrote:
 >I like to liken the Approval/Everything else argument to that between
 >Laissez Faire economists and everything else. If the market for votes was
 >free and everybody had equal information and was supremely rational, then
 >approval would be the laissez faire method of choice.
 >
 >On Thu, 16 Mar 2000, Craig Carey wrote:
...

A dictionary I have, says of the term: 'unrestricted freedom in
  commerce, especially for private interests'.

A problem with the analogy is that "equal information" might be better
  if replaced with 'equal good quality information'. The Approval Voting
  method can be a definitely difficult system for experts knowledgeable
  with the method if they do not get enough quality information (i.e.
  how many 'approval' sub-votes should be made?).

'Equal information' has no analogy to equal richness or equal power.

Limitations on influence is another key quality possessed by STV/IRV
  that politicians presumably look for and which the Approval Voting
  method can't deliver to the ["supremely rational"] pickers of systems.

Even the one winner Approval Vote does not limit voter power when
  power is considered in a way that is done with signal spectrums,
  [energy in signals or waves]: the root of the integral or sum of
  squares of the values in the wave (waves correspond to weights).

To restate that: even when there is one winner only, the Approval
  Voting method gives a lot more power to groups of voters that use a
  lot of preferences, as compared to those that use one preference.
  If there is a clear best choice, then a group having sincere or
  uneducated voters, those that use less preferences, can lose
  influence in affecting the final choice of winner.

It was correct to say a paper (voter) can only have the power to undo
  1.0 of a vote by any other particular person. But so what?, when that
  voter is part of a group that has 10 times the power _per member_
  of another particular group. How could the argument that each person
  has the same power as another be correct when group A that casts X
  approval sub-votes has many times the power of another group.

The disempowered group is presumably the one that would seem to be
  likely to contain a vote from the mayor or president. Don't the
  media get to film them as they cast their vote and ask them who
  they will be voting for. The could be easy targets to questions
  from the TV media like "I had to vote today too. Tell us why you are
  not exercising for the good of your party, your full powers under the
  new Approval Voting system?. You said you were voting for yourself,
  and that is not, is it sir, the best that can be done to cause
  candidate C to lose if you lose.". What would an A.V. advocate
  recommend in a circumstance like this?.

This is another matter where STV passes a strict test perfectly
  (a requirement that undue influence of voters be limited). It seems
  to me to be neither monotonicity nor truncation resistance. The
  Approval Voting method fails. If Benjamin Franklin could comment
  on the Approval Voting method, what might he say?.

Perhaps that the method was worrying because it gave a possibility
  of scientists having too much power, and then, if they had that,
  who else has too much power. A scientist could be a keen observer
  inside a voting room in trying to assess from short cropped
  hairstyles, snuffles, nazi emblems and negroid skin colour, and
  so on, whether that voter will be improperly promoting a narrow
  interest once he walks into the booth.

  So, I guess a select committee would want to investigate that sort
  of possible corruption of a new Approval Voting system prior to
  settling upon that a system.

  Perhaps this could be done: require that group g get less than
  R(f) times the power the group should have, where f is the fraction
  of the public that is contained in group g.

  R is a function decided upon by the committee that embodies their
  tolerance of the level of corruptness of the democratic process
  that is permitted under an Approval Voting system. It should be
  accessible under a Freedom of Information Act.

  Those would be decisions made in USA, for the people, except that
  Mike wrote something about IRV being favoured over the Approval
  Voting method.




________________________________________________________________
Errata: A wrong statement I made can be corrected to this:
   (y loses => x loses) = (x wins => y wins).