The Welsh are whining about Rhydian's defeat in the X-Factor. However, although his defeat was a travesty, it was the result of an efficient process, as this paper (pdf) explains.
Had the X Factor been a pure singing contest, Rhydian would have won by a mile; who couldn't love this? And that's the problem. The other contestants would have realized their inferiority straight from the first show, and would therefore have been less inclined to try to improve their singing. This would have made for boring shows.
However, making the shows into a partial popularity contest, by allowing public votes which aren't necessarily strongly related to singing ability, improved the show. Lesser singers than Rhydian, such as Leon, were incentivized to try harder. They figured: "Hey, I'm quite popular. If I put in some effort, I might win this." And Rhydian, seeing their effort and improvement, responded: "This isn't the walkover I expected. I've got to raise my game."
The upshot is better singing, and a better show, all round.
This has implications beyond Rhydian's defeat. It matters for the design of incentives generally, and for equality.
For example, imagine a firm seeking to promote someone from within. If there's one outstanding candidate, all other workers might figure: "there's no point me trying hard - he'll get the job." It might, therefore, be efficient for the company to limit the pay-off to being outstanding - say by introducing a little (only a little) randomness into the promotion, or having second and third prizes - as this would increase aggregate effort.
Winner-take-all tournaments can, then, be inefficient.
Does this apply, then, to presidential elections?
Posted by: Dave Cole | December 17, 2007 at 02:40 PM
Possibly, because of the entrenched power of the main parties. Because everyone expects a Democrat or Republican to win, they have little incentive to set up third parties, or engage in political activity outside the party system. The upshot is an under-supply of political activity (perhaps).
Posted by: chris | December 17, 2007 at 04:35 PM
But is the X-Factor purely about the singing? If they all sang the same songs, it might be, but they don't. Contra Messrs Cowell and Walsh, I think Danni Minogue did a great job with Rhyd: songs like 'Go West', which he clearly would not have picked himself, suited his talent. However, I think he blew the final with a dull religious number and by singing with Catherine Jenkins (who I thought had an appalling, screechy voice). I love 'Somewhere' and he sang it very well, but I don't think it was what that audience wanted to hear.
It's what you do, not the way that you do it.
Posted by: Backword Dave | December 17, 2007 at 07:23 PM
second and third prizes
But how can you work this into job applications, where it's first past the post?
Posted by: jameshigham | December 18, 2007 at 02:24 PM
Surely the lesson is to surround people with those who are almost their equal, i.e. success is within range, then all will try harder.
If you put the less able into a room with the very able, the less able are likely to give up, as you suggest, and so out the window goes potential opportunity an in comes poverty of ambition as they give up on themselves.
This is why I believe streaming/setting helps people and mixed ability classes (not neccessarily schools) less so.
Posted by: Roger Thornhill | December 22, 2007 at 11:08 AM
«But how can you work this into job applications, where it's first past the post?»
In some countries jobs in large scale organisations are advertised in batches of N, and the first (according to some point metric) N of the M applicants are hired.
However let's not joke with theory -- in practice nowadays all (those few) non-entry-level positions the hiring decision is made by the affected manager, and many of these will hire the least qualified candidate who still meets the job specs exactly.
Posted by: Blissex | December 23, 2007 at 09:26 PM