AlanCrowe comments on Why Support the Underdog? - Less Wrong

35 Post author: Yvain 05 April 2009 12:01AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (86)

You are viewing a single comment's thread. Show more comments above.

Comment author: AlanCrowe 05 April 2009 12:46:28PM 16 points [-]

At the Go club, some-one asked about using red, green, and blue stones instead of using black and white. The chap who is doing a PhD in game theory said: the two weakest players will gang up on the strongest player, *just like any truel".

I was surprised by the way he spoke immediately without being distracted from his own game. Study long enough and hard enough and it becomes automatic: gang up on the stronger.

Now humans have an intuitive grasp of social games, which raises the question: what would that algorithm feel like from the inside? Perhaps it gets expressed as sympathy for the underdog?

It might be possible to test this hypothesis. A truel is a three player game that turns into a duel after one player has been eliminated. That is why you side with the weaker of your two opponents. The experimental psychologist setting up his experiment can manipulate the framing. It the game theory idea is correct, sympathy for the underdog should be stronger when the framing primes the idea of a follow on duel.

For example if you frame America versus bin Laden as the battle of two totalising ideologies, will the world be dog-eat-dog Capitalist or beard-and-burka Islamic, that should boost underdog-sympathy. If you frame America versus bin Laden as pure tragedy, "Americans just want to stay at home eating, bin Laden really wanted to stay in Mecca praying, how came they ended up fighting?", that should weaken underdog sympathy.

I'm not sure how to set up such an experiment. May it could be presented as research into writing dialogue for the theatre. The experimental subject is presented with scripts for plays. In one play the quarreling characters see rivals they are discussing (e.g. Israel v Palestine, etc) as expansionist, in another play the quarreling characters see the rivals they are discussing as fated to fight then go home. The experimental subject is probed with various lines of dialogue for the characters, which either sympathise with the underdog or the overdog, and asked to judge which seems natural.

The hypothesis is that it is the characters that anticipate a follow on duel whose dialogue feels natural with sympathy for the underdog.

Comment author: andrewc 05 April 2009 11:56:18PM 3 points [-]

Interesting idea: we support the underdog because if push came to shove we'd have a better chance of besting them than the top dog? There's a similar problem I remember from a kids brainteaser book. Three hunters are fighting a duel, with rifles, to the death. Each has one bullet. The first hunter has a 100% chance of making a killing shot, the second a 50% chance, the third a 10% chance. What is the inferior hunter's best strategy?

Comment author: Larks 16 August 2009 08:07:58PM *  0 points [-]

The normal answer (fire away from either) only works if we assume the other hunters are vindictive, rather than rational. If we assume they behave rationally, then the third hunter should target the best.

Comment author: Broggly 03 November 2010 02:06:45PM 0 points [-]

Sure, if you're acting simultaneously If you're taking turns and you kill the best, then the mid-strength hunter will immediately fire on you. However if one of them shoots the other, then you'll have the first shot against the remaining one.

Comment author: Larks 03 November 2010 04:04:32PM 0 points [-]

Yes, you're right. Larks@2009 hadn't studied any maths.