That Perennial Western Malady, the Revolt of the Individual Against the Species

Atom Feed

Creative Commons License

Blogger Profile

Get Firefox!
This site appears best in the FireFox web browser. Faster, Safer, and Free.

The Electronic Frontier Foundation

Institute for Justice

I Proudly Fly the Gadsden Flag

Bill of Rights Enforcement Defend the Bill of Rights

eXTReMe Tracker

Tuesday, November 16, 2004

Instant Runoff Sucks

Posted October 24, 2004 on the forum. Information from

Our voting system is inherently flawed. The main reason we are stuck under this two party system is not unjust ballot and debate access, but because our voting system dictates, mathematically, that two parties share all power. I haven't talked to or heard of anyone not in support of the above statement besides high level Democrats and Republicans and the partisan hacks that run the mass media.

For the last few months I've been slightly in favor of Instant Runoff Voting, in which voters rank their candidates by preference. For instance, of the eight presidential candidates on the Florida ballot, Florida voters would choose their top choice and as many alternate choices as they wished. After all votes have been cast, the officials will total the candidates' first choice votes. If one receives a majority, nothing else needs to be counted and that candidate is elected. If no candidate receives more than 50%, then the candidate who received the fewest number of first choice votes is thrown out, and the people who chose him as their first choice have their second choice candidate counted in the recount. The process continues until one candidate has a majority.

Supporters of the IRV method argue that it will eliminate the "wasted vote" problem. However, in the past week I've learned that this is only true as long as the third party candidates have a very small chance of winning. As soon as the pundits begin predicting that a third party candidate will make a substantial showing, we'd be put back in the same "strategic vote" boat we're in now. For example, we can pretend that in 2004 there will be three candidates on the ballot, Bush, Kerry, and Badnarik, and that one will be elected using the IRV process. If the analysts predict that Badnarik could take upwards of 20% or so, many of the people who planned on voting Badnarik but favor either Bush or Kerry strongly over the other, would be worried that giving Badnarik their first choice vote would lead to their second choice being thrown out. If Bush received the fewest number of first choice votes, but a larger number of second choice votes than Kerry or Badnarik, the Badnarik supporters have "spoiled" the election in Kerry's favor.

As evidence, Australia has been using instant runoff voting for a number of years, yet they are still stuck under the same two party system they were before.

Recently, I've found out about Condorcet voting. Condorcet voting takes each candidate in the election, and pairs them up into as many one-on-one combinations as possible. In the example above, one could vote Badnarik over both Bush and Kerry and still be able to vote Bush over Kerry. No candidates are thrown out, and points are spread about all candidates, as opposed to only the "first choice" candidates. The strategic vote is effectively eliminated. I haven't found any compelling arguments against Condorcet voting. It is the fairest and most just method I've come across.

For a more in depth discussion of Condorcet voting, as well as a few other voting methods: The site is for advocating "approval voting," which is interesting as well.

Every supporter of election reform and every third party sympathizer needs to research this issue and become vocal about it. This is the issue we need to address, not ballot access, not debate access criteria, not campaign finance restrictions. If you support IRV, you need to reconsider your options. Every single unit of energy spent advocating IRV is wasted. IRV offers no advantage even to our current system.

UPDATE: February 01, 2005

This is for clarification. I fully support Approval Voting as an immediate solution to the problems caused by the current system. Approval Voting allows the voter to vote for multiple candidates for a single office. For instance, I would have voted only for Michael Badnarik, but my family could have voted for both Badnarik and President Bush. Implementation would not require any new equipment, and the alterations to the existing ballot design would be limited to changing the wording of the instructions from "choose one" to "choose as many as you wish." It can be implemented immediately and costlessly, and is a huge advancement over Plurality Voting and much better than Instant Runoff. Approval Voting is not as accurate as Condorcet, but the potential financial costs and the learning curve that would be associated with switching to Condorcet make Condorcet desirable as a long term goal. Approval Voting is a practical solution with definite positive results that can be implemented TODAY.


At 5:54 AM, Anonymous Anonymous said...

Great work!
[url=]My homepage[/url] | [url=]Cool site[/url]

At 5:54 AM, Anonymous Anonymous said...

Great work!
My homepage | Please visit

At 5:55 AM, Anonymous Anonymous said...

Nice site! |

At 8:51 PM, Anonymous Anonymous said...

Approval Voting (which is the simplest form of Range Voting) is better than Condorcet.

Condorcet voting is based on the intuitive, but flawed notion that if more than half the voters prefer X over Y, then X is better than Y for society.

Approval Voting has a substantially better social utility efficiency than Condorcet methods, meaning you as a voter have a higher expected value with the election result. You're more likely to be more happy. That can be objectively calculated, believe it or not.

Also, Condorcet methods suffer severely from a type of strategic voting flaw called the DH3 pathology, that makes them fall apart in practice.

I got into voting methods in the summer of 2006, when I "invented" Condorcet voting, then found out online that it had already been invented. I thought Range Voting seemed pretty silly at first, but then I spent a great deal of time looking at the science to it, and it's hands down superior to all rank-order methods (Range Voting is cardinal, not ordinal).

Call me if you're unconvinced by any of this. I'll spend as much time as necessary to prove to you that you're wrong.

Clay Shentrup
San Francisco, CA


Post a Comment

<< Home