[EM] range and borda really alike? not.

[Warren Smith]

range and borda really alike? not.


>Scott Ritchie:
>Range voting is just a Borda count with a bunch of throwaway candidates

--WDS REPLY:
Range voting and Borda indeed have a lot of similarities,
but they also have some extremely crucial differences.  In fact I would
say that range voting keeps all the virtues of Borda but
loses its major problems.  It sort of combines
the virtues of Borda and Approval.

Let me illustrate with a few examples.

EXAMPLE #1:
BORDA: Bush v Nader v Gore.
In this situation, consider (a) the strategic Borda voter who feels Nader>Gore>Bush,
votes Gore>Nader>Bush.  Namely, Gore top and Bush bottom both are forced
strategically, and then the rules of the Borda system force Nader middle.
Similarly consider (b) voters who feel Nader>Bush>Gore.  They strategically vote
Bush>Nader>Gore.
This causes it to be impossible for Nader to win with strategic Borda voters,
even if the voters in class (a) and (b) combined are 60% of the population.
Gore or Bush always win (except in an exact 3-way tie).

RANGE: Bush v Nader v Gore.
Our voters of type (a) can vote Gore=Nader=99, Bush=0.
Of type (b) Bush=Nader=99, Gore=0.  These are the strategically correct votes.
In this case  if the voters in class (a) and (b) combined are 60% of the population.
then Nader will always win (assuming Bush & Gore are pretty close).  A big difference.

IRV:  Bush v Nader v Gore.
The CRV web subpage shows by explicit numerical examples
that it can often be advantageous for Nader
voters to exaggerate strategically and vote Gore and Bush max and min
(with Nader consequently middle).  See, e.g.
   http://math.temple.edu/~wds/crv/TarrIrv.html .
If the voters in class (a) and (b) combined are 60% of the population and
all act this way (and we again assume Gore & Bush pretty close), 
then Nader cannot win under IRV.
And it is consequently not surprising at all that all three IRV countries
have historically been 2-party dominated.  It is a myth that IRV breaks
the stranglehold of 2-party domination.   In fact both theory and
historical evidence prove that IRV forces 2-party domination, and this is agreed
by the usual political science books that cover the subject of Duverger's law,
e.g. Taagapera & Shugart.  Range, on the other hand, CAN break that stranglehold.

MORAL:
systems in which the votes are rank orderings are too limited in voter-expressivity - 
after you strategically vote Gore=max, Bush=min you by the rules of such systems have no freedom 
whatever to express any opinion about  Nader.   That lack of expressivity is bad.
Range voters can still express any opinion 0-99 about Nader.  That is good.

EXAMPLE #2:
Suppose it is Gore v Bush.  And the evil Democrats secretly sponsor a large
number of Gorelike clone candidates to run, all of whom seem
superior to Bush in democratic eyes (although not in republican eyes)
and all of whom seem slighly inferior to Gore (in all eyes).
result:
   50% of voters:   Gore>Gore1>Gore2>...>GoreN>Bush
   50% of voters:   Bush>Gore>Gore1>Gore2>...>GoreN
(if all voters honest).
This results in an enormous Borda victory for Gore, but it
was caused only by the clones.   With range voting, Gore does not
get amplified by introducing clones, and such ludicrous and too-easy
election manipulation attempts just do not work.

This is another important difference between Borda and Range.  So it is simply 
ridiculous to lump Borda and Range together.  They behave entirely differently.
wds