Re: [EM] 05/15/02 - Josh calls for more math:

[Alex Small]

> * Most of the lower choices are not informed choices.

When we discuss races with 3 or 4 candidates, we're talking about 3 or 4
contenders.  Contenders are by definition known to many people.  I presume
that in just about any method, the little-known obscure candidates will
have no chance of winning (except maybe Borda).

To take two specific methods, why would anybody give an approval vote to a
candidate from an obscure party like Natural Law or whatever?  How would
that lower choice hurt the favorite?

In Condorcet, that obscure candidate would probably lose all pairwise
contests (except against other obscure candidates).  How exactly will John
Hagelin undermine George Bush and Al Gore with Approval or Condorcet?  And
if you don't know who John Hagelin is, obviously that lower choice is not a
threat.

> * Lower choices are used to harm earlier choices by helping some other
> candidate.

Let's look at 3-level approval.  If somebody has a majority of "preferred"
votes he wins.  Similar to IRV.  If not, nobody is eliminated, but
additional votes are distributed to candidates based on how many people
rated them "acceptable."

Admittedly, a vote is being given to your second choice, even though your
favorite is still in the running.  However, other people can now give votes
to your favorite.  Nobody is eliminated.  Favorite still has a chance.
Would you rather see favorite eliminated when your vote is given to
somebody else, or would you rather see him still have a chance?

Which case is better is not clear-cut.  Some people dislike the idea that a
lower choice is benefiting while favorite remains in the game.  Others like
the idea that favorite still has a chance.  Since there's no clear
philosophical advantage, we have to look at strategic questions, among them
monotonicity, IIA, etc.

> * The lower choices are not netural.  Most of them are for the lower
> candidates.

Your point being?

> * Most choices are not made for the candidate, but instead the choices
> are
> made for the party of the candidate.  Change the candidate of the party
> to some unknown and the voter will vote for him anyway.

How is this different from any other method?  Many people that I talk to
seem convinced that any candidate from a different party must be bad.  They
are convinced that all Republicans want to take food from children, or all
Democrats want to rob them blind.  Will their minds suddenly open if they
have an IRV ballot in front of them?

Alex

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em