academia | advice | alcohol | American Indians | architecture | art | artificial intelligence | Barnard | best | biography | bitcoin | blogging | broken umbrellas | candide | censorship | children's books | Columbia | comics | consciousness | cooking | crime | criticism | dance | data analysis | design | dishonesty | economics | education | energy | epistemology | error correction | essays | family | fashion | finance | food | foreign policy | futurism | games | gender | Georgia | health | history | inspiration | intellectual property | Israel | journalism | Judaism | labor | language | law | leadership | letters | literature | management | marketing | memoir | movies | music | mystery | mythology | New Mexico | New York | parenting | philosophy | photography | podcast | poetry | politics | prediction | product | productivity | programming | psychology | public transportation | publishing | puzzles | race | reading | recommendation | religion | reputation | review | RSI | Russia | sci-fi | science | sex | short stories | social justice | social media | sports | startups | statistics | teaching | technology | Texas | theater | translation | travel | trivia | tv | typography | unreliable narrators | video | video games | violence | war | weather | wordplay | writing

Wednesday, April 25, 2007

A vote guaranteed to elect a loser

Out of a field of dozens of candidates, French voters may have already eliminated the most popular, thanks to their clunky runoff system. The Herald Tribune wrote before the first round election:
The French know they do not have the luxury of voting their hearts the first time around. In the first-round ballot in April 2002, the protest vote went to Le Pen, and in a shock, he knocked out the Socialist candidate for second place in the first round and faced the incumbent Jacques Chirac in the runoff.
"I'd like to vote for Besancenot - a simple mailman who speaks to the little person - and there are a lot of people like me," said Azzedine Hamet, a 25-year-old unemployed metal worker, referring to Olivier Besancenot, the 32-year-old Trotskyite candidate. "But I feel the burden of 2002. To vote like that is to throw away your vote."

In a number of polls, [Francois] Bayrou has a better chance than Royal of defeating Sarkozy in the second round. That means some voters on the left who are determined to defeat Sarkozy are contemplating whether to sacrifice Royal and support Bayrou.

We all know that didn't happen. As a result, though French people in general might collectively be happiest with centrist Bayrou, they now are set to pick between socialist and conservative candidates.

This despite the fact that the day before the election, a poll by TNS-Sofres showed Bayrou beating both Sarkozy and Royal in head-to-head races, and in effect beating every single candidate running for president of France in a head to head race. This is the candidate who should be elected. This from an article that, like other misguided media coverage, starts "Nicolas Sarkozy continues to lead all presidential contenders in France..." The colossal lede that the most popular candidate is likely to lose is buried in the fourth paragraph, as usual.

Of course, the United States' system is even worse. John McCain was a more popular candidate than George Bush in 2000, but the primary system kept him from millions of swing voters who never had a chance to vote for him. (For the record, McCain is on the wrong side of nearly every important political issue and isn't the bipartisan savior we're looking for, but he sure is better than Bush.) The way Congressional seats are now allotted in the Republicans' favor, Democrats can--and have--get more votes for House and Senate candidates than the Republicans do, but win fewer of the seats. And the ills of our electoral college have been enumerated plenty.

It's crazy that in this age of self-satisfied democracies, citizens are content with electoral systems that don't choose the most popular candidates or parties. Voters in Kiribati and Sri Lanka rank candidates, and a form of instant runoff voting is used in each case; I like the Borda count method (with a set number of candidates to rank) and other Condorcet methods, which guarantee that if there is a candidate like Bayrou who would win head-to-head against any other candidate, he or she will be chosen.

Voters should be able to focus on who they want to elect, and free to ignore questions of voting strategy and other such machinations. No voter in a democracy should have to vote against their first choice for reasons of game theory.

Labels: ,

Anonymous Anonymous on Thu Apr 26, 12:39:00 AM:
Borda voting seems like a reasonable idea, until voters get strategic, at which point it falls apart.

As for Condorcet methods, they are all deeply flawed, because of certain problems inherent to ordinal voting methods, like not satisfying independence of irrelevant alternatives. Plus, the promise of Condorcet methods, which is to always elect a beats-all winner, when one exists, are flawed because

A) The DH3 pathology often prevents Condorcet winners from winning, and instead picks a very very BAD winner, in practice.

B) They are predicated on the idea that if a majority of voters prefers A to B, then A is a better candidate; but this is easily disproven by Condorcet cyclic ambiguities like this one

34% A > B > C
32% B > C > A
34% C > A > B

Which ever one of these candidates you decide is "best", you can remove on of the other two candidates, and suddenly that "best candidate" loses the majority vote.

The superior measure of a voting method is social utility efficiency - or, average voter satisfaction with the result. By that metric, Range Voting, where voters simply score the candidates and elect the one with the highest average, is hands down the best method.

The simplest form of Range Voting, called "Approval Voting", just uses a 0-1 "range". It's effectively identical to the way we vote now, except that you can vote for as many candidates as you want (or vote "against" them by not voting for them). This method is far superior to Borda and Condorcet, and much simpler.

Clay Shentrup
San Francisco, CA
Anonymous Anonymous on Thu Apr 26, 01:20:00 AM:
Ankaŭ, se vi ŝatas krei lingvojn por neveraj landoj, ci povas komenci kun ĉi tiu.
Blogger Ben on Fri Apr 27, 06:47:00 PM:
I also like approval voting, and I appreciate its simplicity compared to range voting, which would take more explaining than Borda ranking.

But I think range voting, including approval voting, are almost as problematic when it comes to strategic voting as Borda is. All of these are better than one-single-vote, of course.

On the site you link to, Range voting's strategic voting problems are dismissed with:

"the mild-opinioned voters [whose voting backfires and leads to the election of a less-desired candidate] have little basis for complaining, since their preferences were only mild and since they voluntarily chose to express them".

But couldn't the same be said of strategic Borda voters, as are described in the DH3 case?

In all voting systems, there will be cases in which a voter might achieve a better result by voting dishonestly than honestly. The question is, how do we minimize such cases, while keeping the system simple to understand? I think Borda, on balance, fits these needs better than range or approval voting.

An added benefit of some of these alternatives is that they would make it harder to see clearly who is leading in polls, since the polls would be necessarily farther-removed from the actual ballott decisions facing voters. Borda, in particular, is simple to follow but complicated enough that strategic voting would be harder to think through than it would be with approval voting.

I do concede that the cloning problem is damning evidence against Borda. In this light, maybe approval is better overall.

At any rate, we can agree that Bayrou should have won, and that Borda, approval or range voting would probably have elected him (and that instant runoff would not).

By the way, I love the page on funny elections at .