Consensus, Condorcet(0), and Condorcet(1/2)

[Mike Ossipoff]

Of those 2 possibilities for a 2 alternative vote,

1. Electing A even though A loses to B 46 to 54, or
2. Electing A even though A loses to B 34 to 46, ...


..#1 is worse. It violates majority rule in a bigger way, in
a sense that #2 doesn't. In #2, though B should win, no one
can say that a true majority has said anything.

Putting it another way, #1 violates all 3 of the basic democratic
principles that I named, while #2 violates only 1 of them.

***

Those basic democratic principles that I named were:

1. If a majority indicate that A is their favotite, then A
   should win.

2. If the number of voters indicating that they'd rather
   have A than B is greater than the number indicatng
   the oppopsite, then, if we choose A or B, it should be A.

3. If a majority indicate that they'd rather have A than B,
   then if we choose A or B, it should be A.


[I've listed those principles in the order of the rarity of
getting a winner by that princple, with the one most rarely
giving a winner being listed 1st]

***

The justification for Condorcet(1/2) is a justification in
terms of a points count rather than a preference count.
Though it can be argued to have a system in which someone
can give bad-points (or good-points) in any way they want
to, that's a point count, not a preference count. 

So that's what Condorcet(1/2) implements: a standard based
on assignment & counting of points--a points method standard.

We've discussed the advantages & disadvantages of point
assignment methods. They let us more truly express the
intensity of our preferences--but that turns out to be
a bad disadvantage disguised as an advantage, because it
gives us a dilemma about whether to give Middle a full point
vote over Worst, or to give Best a full point vote over
Middle. We can't do both in a points system, though we can
in a pairwise-prefernce counting system. And, as I've said,
Condorcet ensures that when other pairwise-count methods no
longer do.

Of course the point-system disadvantage named in the preceding
paragraph is another statement of the LO2E problem--a problem
present in all but the best sw methods.

***

Mike





--