This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Economics 101, UCLA Fall 2010 Jernej Copic Lecture 4, 10/5/10. Game Theory: Mixed strategies. Introducing Pareto optimality. When Grunf left, No1 was still scribbling furious notes on random pieces of paper. Among the more notable ones, there were some rather unpronouncible phrases, composed mainly of vulgarities, and there was also a write-up about a game. Ironically, its title was Prisoners Delight. A prisoners delight. This is like a prisoners dilemma, but the payo ff s are di ff erent when they are silent, they get a large payo ff . The reason is that Pai Mei and the monk have bonded so that they care for each other even if the tea ceremony they would have to endure when they are both silent is very long, dull, and tedious, remarkably enough, each one prefers going through that to seeing his companion rot in jail for a very long time and then going berserk with revenge upon release. The payo ff s are all the same as in the prisoners dilemma between Pai Mei and the monk (see Lecture 3), except that if they are both silent, then the monk gets a payo ff of 6 utils and Pai Mei a payo ff of 5. Okay, so now, to figure out the equilibria, lets ask the question, when would neither of them have an incentive to change their behavior. 1 If Pai Mei is silent and the monk is silent, then none of them could benefit by changing their behavior - if nothing else, each of them is getting the most he can get. So this is an equilibrium! If Pai Mei is silent and the monk rats, then Pai Mei would in fact prefer to change to rat; also the monk would then prefer to be silent. 2 If Pai Mei rats and the monk is silent, then again, either would have wanted to change their behavior, so this doesnt work. 1 Exercise. Write down this game in a matrix form. 2 Exercise. Confirm this by looking at the matrix. Note, when for instance Pai Mei makes his consideration of whether he couldve benefited by changing his behavior, he cant a ff ect the behavior of the monk. So he compares his payo ff s for the case when the monk rats. 1 If they both rat, then... Oh, then again none of them could benefit by a change. For example, Pai Mai, if he instead remains silent, he is getting a payo ff of 1 rather than 1, so he wouldnt want to do anything di ff erently. No1 was still staring at the Prisoners Delight, and he was determined he would get EVERYTHING right this time. Something was bothering him. What about if Pai Mei decides for some reason to flip a coin between his two possible actions? I remember from a book that this is called a mixed strategy , and they wrote it like 1 2 silent + 1 2 rat . Hmmm... hold on, hold on... Then, lemme compute how much the monk gets out of playing either of his two actions. Okay, so if he decides to be silent, he is now facing a lottery. With probability 1 2 he will find himself in an outcome where Pai Mei was silent (and the monk is silent) in which case he gets a 6, and with probability...
View Full Document
- Spring '08
- Game Theory