[EM] Re: IRV bill on Washington State Governor's desk

[James Green-Armytage]

Russ, here is my reply to your recent posting...
>
>I did not mean to suggest that others should not discuss "complicated" 
>methods. 

	That's good to know.

>I am merely suggesting that, for their own good, they should 
>avoid setting themselves up for disappointment by imagining the public 
>will be open to complicated methods for major public elections.

	Yes, well, that's nice of you, I suppose... but I suspect that most of us
are aware that large scale election methods are extremely hard to change
in general. The fact that we still use the electoral college to choose the
president may be evidence enough for that.
>
>IRV is deceptively simple from the voter's perspective. The count at 
>each "round" is identical to a conventional plurality election. If your 
>top choice loses your vote transfers to your next choice. 

	Yes, IRV is pretty easy for most people to understand, because it is very
similar to the two round runoff.

>(The fact that it's non-monotonic and non-summable 
>would never occur to 99% of voters, and even if they are told they won't 
>care.)

	I don't think that either of these criteria are extremely important
(especially summability). The most significant deficiencies of IRV in my
opinion are its failure of the Condorcet and Smith criteria.
>
>I was being a bit flippant about the "margins vs. wv" issue. Let me take 
>another shot at it. Many voters will intuitively reject wv in favor of 
>margins. You can call them naive if you wish, but the point is that such 
>"naive" intuition will create a major impediment to adoption. The issue 
>will never get resolved to the point where ordinal-only Condorcet can be 
>adopted. If public acceptance of Condorcet/margins ever reached a 
>critical mass, Mike Ossipoff would start throwing stink bombs or 
>threaten to light himself on fire in protest, for example.

	Self-immolation should not be necessary in such a scenario. Margins has
very severe strategy problems, and thus it could not withstand the kind of
scrutiny that it would receive if it was a serious consideration for a
large-scale election system.
>
>But the issue will never get anywhere near that far. The whole Condorcet 
>ordinal-only method is so complicated that public interest will never 
>reach a critical mass. And when it becomes clear to the public that many 
>variations exist and even the "experts" do not agree on which is best, 
>even receptive members of the public will simply throw up their hands 
>and move on to other issues -- like who the major-party candidates 
>should be in the next election.

	You seem to be stating matters of opinion as if they were matters of
fact, and speaking for large numbers of people whom you can't accurately
speak for.
	Again, the acceptable level of election system complexity varies entirely
from situation to situation. (And you rarely specify exactly what
situation you are talking about. Is it U.S. presidential elections,
perhaps?) Highly complex voting systems have been adopted for large-scale
use. STV (which is significantly more complex than WV, and arguably more
complicated even than cardinal pairwise) is a perfect example of this. The
important thing is not necessarily how many rules there are, but that
every additional rule has an extremely well-defined purpose, i.e. a
demonstrably valuable purpose that couldn't be adequately fulfilled by a
simpler method.
>

my best,
James