[EM] IRV letter

Curt Siffert siffert at museworld.com
Thu Apr 22 19:18:07 PDT 2004


I like explaining it in real-world scenarios rather than using numbers 
and rankings, for instance:

~^~^~^~^
Right now in the U.S., our voters are registered in thirds - about a 
third are Republican, about a third are Democrat, and slightly less 
than a third are Independent.

Say you had an election of three candidates - a centrist, and two 
wingers.  They might each have a third of the vote, the centrist 
slightly less than the other two.  The centrist would almost certainly 
be the second choice of those who prefer the wingers, given that they 
would dislike the candidate from the opposite wing.

It's obvious that in this scenario, the centrist would be the consensus 
choice.  However, IRV would eliminate the centrist first.
~^~^~^~^

I think that the approach of listing out numerical examples where odd 
things happen is a difficult approach, because there is never any 
indication of how likely that scenario is.  Putting it in real-world 
terms like the above would help.

As for the partial ballots, it may help to point out that the system 
only looks at a person's full ballot "if necessary", but in most cases 
it completely ignores most of the ranking that a voter gives.  If a 
voter fills out a ballot, wouldn't they want their full ballot to 
count?

I also think that it helps to point out that in effect, IRV gives a 
"bonus" to first-place votes.  While this may make sense intuitively at 
first, point out that it assumes that everyone *far* prefers their 
first choice, disenfranchising those who may like their first two 
choices about equally, but hate the third.  We know the contradictions 
that can happen when people "weight" their choices as in Borda, etc, so 
wouldn't it make more sense to use a system that makes no such 
assumptions?  Condorcet is the only one that looks at preferences, and 
preferences only.

If the goal is to find a consensus choice and to select the candidate 
that the population prefers to any one other candidate, then Condorcet 
is the only way to go.

Thoughts along these lines have helped me advocate Condorcet in the 
past.

Curt

On Apr 22, 2004, at 6:44 PM, MIKE OSSIPOFF wrote:

> Here's my letter that I'll be sending to everywere where I hear that 
> there's an IRV proposal. Of course I may add to and improve the letter 
> as I keep using it. Feel free to use any information in it, or to copy 
> or forward the entire letter, or to quote or use it entire or in part 
> in any medium:




More information about the Election-Methods mailing list