[EM] Slashdot: An Analysis of Various Election Methods

[Dave Ketchum]

On Sun, 3 Oct 2004 10:54:36 -0700 (PDT) bql at bolson.org wrote:

> On Sun, 3 Oct 2004, Bart Ingles wrote:
> 
>> I started out on this list in 1998 as an IRV supporter, but now see it 
>> as a step in the wrong direction.  Back then I believed IRV had 
>> properties I considered important (e.g. resistance to low-utility 
>> winners in worst-case scenarios), but have long since learned that I 
>> was mistaken.  Thus it turns out that my support was due to ignorance.
> 
> 
> I don't think IRV is a step in the wrong direction. It's still vastly 
> superior to the present single vote system. It solves the simple 
> problems correctly, like any of the systems we consider do. It's just 
> all those "interesting" problems we like to conjure up that trip things.


IRV is better than the ugly ones most of the time.  Since it uses the same 
basic ballot as Condorcet, it even has the voters studying what we want 
them to learn.

Enough use and I feel certain its flaws will raise the question of WHY NOT 
go straight to Condorcet, rather than having to propose a second move to 
get away from IRV:

I feel CERTAIN enough use of IRV will demonstrate the following pattern:
      35% are SOLID backers of some "important" position, such as about 
abortion and many of the other issues that get us hot under the collar.
      65% - a SOLID majority - disagree.
      Among the 65% there is a new thought that is gaining momentum as an 
improvement.  Backers of the new thought rank its candidate first, and the 
candidate of the basic majority belief second.
      Let the new thought candidate get to 32% and IRV will still see this 
candidate lose, letting the 65% majority win.
      New candidate get a few more votes to 33% and old candidate with 32% 
loses in IRV; 33% loses to 35%; and we have UNHAPPINESS.

With Condorcet the matrix is a reasonably compact summary of the voting 
for the public.  Also, with multiple polling stations, such as for 
electing a governor, matrices can be summed for convenience of all concerned.

With IRV there is dependence on the voting pattern in individual ballots.

> 
> When it eventually comes to getting these things implemented, my 
> compromise solution is to propose a law that allows the elections 
> official (county clerk, secretary of state, etc.) to chose from a list 
> of approved methods. In exchange for a little "chaos" I think we'll get 
> definitive practical results with none of the results being terribly bad 
> (and there's always recall elections if someone truly suckful does 
> somehow get elected).
> 
> Brian Olson
> http://bolson.org/

-- 
  davek at clarityconnect.com    people.clarityconnect.com/webpages3/davek
  Dave Ketchum   108 Halstead Ave, Owego, NY  13827-1708   607-687-5026
            Do to no one what you would not want done to you.
                  If you want peace, work for justice.