[Election-Methods] Re : A Better Version of IRV?

[Kevin Venzke]

Hi Forest,

--- En date de : Dim 6.7.08, fsimmons at pcc.edu <fsimmons at pcc.edu> a écrit :
> There is a lot of momentum behind IRV.  If we cannot stop
> it, are there some tweaks that would make it more liveable?

My idea is to cripple it:

1. The voter ranks a first preference and a second preference. No more,
and possibly no fewer.
2. Eliminate the 3rd place candidate and transfer preferences to the
1st and 2nd place candidates only.
3. Elect the one of these latter two candidates with more preferences.

The desired effect is that the top three candidates should become very
obvious prior to the vote, so that a "third frontrunner" has the best
chance of winning possible, without fear of losing out on lower rankings
due to not being as visible as the two stronger frontrunners.

The main (and most unfair) way this works is by causing you to have wasted
your vote when you don't pick one of the top three candidates as your
first preference. This creates the incentive for all players to determine
in advance which three candidates are viable, and which should drop out
to avoid spoiling the outcome.

Asking (or especially requiring) two preferences hopefully reinforces
in the voter's mind that we would appreciate it if he voted for two of
the three frontrunners.

IRV is still not very good at picking a compromise from the three (though
these mechanics could be used with other methods instead), but I think
the "third" candidate would have something of a better chance due to his
status as "one of the three" that can usefully be voted.

I'm sure this method is unacceptable. But I'm very concerned that even
the technically best single-winner methods may be doomed to lead to
two-frontrunner elections, unless something is installed to prop up the
visibility of a third choice.

Kevin Venzke


      _____________________________________________________________________________ 
Envoyez avec Yahoo! Mail. Une boite mail plus intelligente http://mail.yahoo.fr