The 'Hot or Not' Solution

A mathematical – but controversial – idea for fixing the flaws in voting

This originally appeared at Newsweek.com as a Steven Levy "The Technologist" column, on 17 Dec 2007.

In the history of U.S. elections, the fall of 2000 is notorious for the debacle that occurred in the country's attempt to elect a president that year. But if a compelling new book is to be heeded, an even more significant election development occurred in the month before America went to the polls that November: the launching of the "Hot or Not" Web site.

I'll explain. According to William Poundstone, 52, author of "Gaming the Vote: Why Elections Aren't Fair" (to be published early next year by Hill and Wang), George W. Bush reached the White House as a beneficiary of our reliance on a demonstrably flawed system of registering voter preferences, known as plurality voting. In this scheme the winner among multiple candidates running for a single office is the one who gets the most votes. On the surface this sounds pretty reasonable, but it turns out that with alarming frequency this can result in a victor who is the least favorite candidate of an actual majority of voters. The problem is the "spoiler," a third candidate who draws support from what otherwise would be the favorite of the electorate. (Often such spoilers are financially supported by those who despise that candidate's politics, utilizing an "any enemy of my enemy is my friend" strategy.)

Editor's note: actually, the strategy of supporting a spoiler is more accurately described as "any friend of my enemy is my friend."

A clear example of the spoiler problem came in 2000, when the combined vote for Al Gore and Ralph Nader – cast by people who presumably preferred either to the eventual victor – was significantly higher than Bush's vote. Poundstone analyzes previous campaigns and concludes, "There have been 45 presidential elections since 1828. In at least five, the race went to the second-most-popular candidate because of a spoiler. That's over an 11 percent rate of catastrophic failure. Were the plurality vote a car or an airliner, it would be recognized for what it is: a defective consumer product, unsafe at any speed."

In trying to figure out how we may improve our system, Poundstone discovers that it's not so easy to come up with a system that expresses the wishes of the electorate and is resistant to being gamed. Casting a glum shadow over the book is his account of how in 1948 Kenneth Arrow, an economist who would later win a Nobel Prize, threw down a mathematical gauntlet that seems to prove that no elections system could be devised without serious flaws. It is known as "the Impossibility Theorem."

It asserts that no matter what you come up with – schemes like our current plurality voting, schemes where you rank candidates in order, or schemes that match each candidate against another – a system will have vulnerabilities and problems that can produce a result that does not reflect the will of the people. Quite a bummer and, indeed, Arrow's Impossibility Theorem has done much to discourage serious efforts to improve what we have. (If the outcome is doomed, why bother?)

Nonetheless, Poundstone himself winds up casting an unequivocal ballot for one system that he's sure can help us choose our leaders in a better way. It's called range voting.

That's where Hot or Not comes in.

When two young geeky types– James Hong and Jim Young – cooked up a Web site in October 2000 that asked visitors to rate pictures of young men and women on a scale of one to 10, it became an instant Internet phenomenon, and remains popular. (Poundstone reports that more than 12 billion votes have been cast, four times the total in all American presidential elections.) It is simple, effective, and able to handle multiple contenders with ease. Poundstone sees no reason why it should not be used to choose our leaders.

If the 2000 election had used range voting, for instance, instead of having to cast a single vote for either Gore, Bush, or Nader, voters would be able to rate each of those candidates on a scale of zero to five. The candidate with the highest ranking wins. A stalwart Democrat would probably rate Gore a 5 and Bush a 0. A rib-rock Republican would vote the reverse: 5 for Bush, zilch for Gore. But a green-oriented voter who rates Nader a 5 may also rate Gore a 4 or a 5. This ends the risk of "throwing your vote away" – and giving the election to someone whom the majority of the voters don't want to see in office.

Poundstone's choice aligns him with a mathematician from Cleveland named Warren Smith, who stands as the most passionate advocate of range voting. Smith, 43, who runs an information-packed Web site on the subject, has used all his mathematical chops to compare systems and claims that range voting is demonstrably superior – he's quoted as saying that switching systems would "be a larger improvement to 'democracy' than the entire invention of democracy." What's more, he insisted to me, it's totally constitutional, and our current voting machines can be altered to handle the new system. Smith thinks that range voting can be particularly effective in primaries, when voters must choose among a long slate of candidates. "It's in the party's own interest to switch to range voting," he says. "There would be a much better chance that the best candidate would win, and then the party would do better in the general election." Plus, the popularity of range voting on the Internet – not just Hot or Not but innumerable sites that ask people to rate restaurants, movies and books– has made people comfortable with the idea.

Will we ever change from plurality voting? Some groups are working hard to come up with alternatives. Advocates of a system called instant-runoff voting (IRV) have gotten some municipalities (San Francisco) to adopt their system, which asks voters to select, in addition to their preferred choice, their second and even third favorites, which can be used in case no candidate wins a majority. (Poundstone's book notes flaws in IRV, notably a scenario in which the least-preferred candidate among three could win the election.) As for the possibility of range voting being adopted, I'm not so sure that citizens will necessarily think that effectiveness in choosing hunks and hotties will tilt them toward choosing leaders in the same way. Poundstone, though, is optimistic about the long run. A switch to range voting in, say 50 years, "is something I would say is conceivable," he says.

In a sense, the battle between those defending our current systems and those who are urging change is emblematic of many problems that have proved intransigent. Those who seek provable, data-driven solutions are frustrated by a resistance to change – and inertia bolstered by special interests that feast on the dysfunction. It's enough to drive a mathematician insane. "I find it maddening when people say that Nader was an evil man for running against Gore," says Warren Smith. "What's evil is the voting system. It just drives me nuts." Poundstone's book raises a big question: how mad do the rest of us have to get before we change a system that just isn't working?


Return to main page