I really do not get this whole "it's too complicated" argument I keep reading about in the American press. I'm an Australian, and we have our General election on October 9. There'll be about 13 million "IRV" votes cast for 150 seats of the lower house, and we should know the outcome of about 145 of them about 4 hours after the polls close. Voting is done with pencil and paper and voters have to rank all the candidates sequentially. Counting is done by hand and about 4% of the ballots will be "informal".

We've been using this system Federally since 1918, and frankly, it works great.

regards,

Darryl

I really do not get this whole "it's too complicated" argument I keep reading about in the American press. I'm an Australian, and we have our General election on October 9. There'll be about 13 million "IRV" votes cast for 150 seats of the lower house, and we should know the outcome of about 145 of them about 4 hours after the polls close. Voting is done with pencil and paper and voters have to rank all the candidates sequentially. Counting is done by hand and about 4% of the ballots will be "informal".

We've been using this system Federally since 1918, and frankly, it works great.

regards,

Darryl

I forgot to mention that I'm also the Greens candidate for one of the seats in the Australian Federal election. Just in case that matters to anyone.

Since this is a labour-oriented blog, I may as well include a link to our Industrial Relations policy (http://www.greens.org.au/policies/society/industrialrelations). The concepts and terminology are pretty different to North American stuff, but someone might be interested I suppose :^)

(and sorry about the double post)

d

The original poster mistakes the purposes of IRV. IRV is not, as he implies, a backdoor means of proportional representation and the direct empowerment of third parties it implies. It is a way to get majority-rules results without the expense and reduced turnout which non-instant runoffs require.

The hypothetical result about which the commentator complains is exactly the result which would occur in a run-off election with 100% repeat turn-out.

Putting IRV up against non-runoff first-past-the-post (i.e., plurality-rules systems) is really comparing apples and oranges.

The real obstacle to proportional representation is not America's voting systems but our embrace of single-member districts, which makes PR an impossibility.

SF progressives were largely responsible for a single-member Board of Supervisors, and civil rights activists have made a 30+ year campaign of getting rid of multi-member at-large systems all around the country.

It's worth noting that single-member leads to significantly greater represenation of blacks and latinos in legislators and in Congress, because single member districts are apportioned and districted on the basis of raw population, not voters, or citizens. PR systems would increase the relative important of voting citizens, and decrease or eliminate the relative importance of non-voters and non-citizens, which are disproportionately black and latino.

There is no animosity between Kerry and
Nader voters, because there are no Nader
voters. People who say they will vote
for Nader are actually going to vote for
Bush and are just trying to make Nader
sound acceptable in an attempt to trick
us. Don't buy it.

Woo hoo! Go San Francisco...

The best solution to the early elimination problem is to use condorcet counting instead of instant runoff. As far as the voter is concerned, there's no difference, but the counting is a little more complicated. Basically, every candidate is compared to every other candidate *individually*, and the candidate that never loses, wins the election. This is the Democrats, in the example above:

R vs G: 51 votes R, 49 votes G
G vs D: 49 votes G, 51 votes D
D vs R: 51 votes D, 49 votes R

In rare situations, no candidate will win every comparison, but there is a simple formula to determine the winner in such cases based on the size of the candidates' victories.

Like the first poster, I think IRV is problematic. Basically, it is an improvement over the current system, but it is such a tiny improvement, perhaps a huge improvement would be a better idea.

The huge improvement, like with the first poster, I would suggest would be a variety of Condorcet.

The mathematics of voting, or "social choice theory", or whatever you want to call it, is quite complex. It was lost to western civilization during the Dark Ages, only to be redone in the late, late 1700s. After, in fact, America had settled on a system.

Their math wasn't perfect. Oh well. Let the National Science Foundation study the problem for a while, and make recommendations to Congress. That was what they did with the apportionment issue. They did a good job on that one. Took them 12 years (although the start and end dates (1929-1941) look very suspicious when you consider we re-apportion each decade, on the decade).

Cheers!