Civil Wars MOOC (#28.3): Revelation Principle Pt. 1

[@MZciggy]

hey William I am a new sub in youre channel, after I watched some of youre youtube intertantional relations vids and I loved them, but could you tell me (short version) what is youre mainly youtube channel about? PS. keep it up :)

[@Gametheory101]

Ignacio Parellada Basically whatever I'm teaching about or have recently written a book on. Happy to have you here!

[@MZciggy]

William Spaniel oh cool, thanks.

[@sfdv1147]

I don't really understand here what you mean by "optimal play". Is it just any outcome or is it a Bayes Nash Equilibrium?

[@InventiveHarvest]

Question: Regarding truth telling in this scenario, suppose I was a strong rebel type. It seems to me I could lie about my preferences, say I was weak type and receive the payoff distributed by the game master. I could the star a war and receive that payoff as well. Yes?

[@Gametheory101]

Well, there are two ways to look at this: 1) When thinking about whether direct mechanisms are incentive compatible, the rules of the game are the rules of the game, period. So in the example, you get the stuff that the game master gives you, and there is nothing you can do to change that. This isn't because we are suggesting that we could replace real life interactions with the incentive compatible direct mechanism. Rather, the revelation principle gives us a way to neatly compact the otherwise (potentially) extremely complicated strategies of any given game. If this seems to only be of academic interest, you are right...so far. However, we can rewrite the revelation principle in a logically equivalent way. The next lecture does this, and we will see the absence of an incentive compatible direct mechanism means that certain outcomes are impossible to arrive at through strategic play, regardless of what the actual interaction is. This is definitely *not* purely of academic interest, as it says that some situations will inevitably end in war. 2) That aside, lying in that manner doesn't make sense. That would be like if strong Putin says "I am weak," and then is given a small portion of Ukraine, say Crimea. Then, the next day, he realizes he could get more through war, starts a conflict, and takes a larger portion, say Crimea and Donetsk. By lying, he doesn't get Crimea twice---he only gets it once. (That is, you can't get a war payoff and a peace payoff.) And if he was better off starting a war than taking the smaller amount of concessions, then he might as well have just told the game master the truth from the start so he could receive his war payoff.

[@InventiveHarvest]

William Spaniel Makes sense. Thanks!

[@mariorpg11]

I actually don't get, what the difference between the real played game and the simulated game from the game master ist.

[@Gametheory101]

Well, the game master doesn't actually play any game. He just takes the information that the players tell him and hands out payoffs based on that information. What guarantees the existence of incentive-compatible direct mechanisms is the fact that the game master can use an allocation rule that duplicates the outcome of the game had they just played it for themselves. So, outcome wise, there is no difference.

[@mariorpg11]

William Spaniel And does this imply, that the gm doesn't have any own interests regarding the game? Is the gm just some kind of abstract conzept we need for understanding the theory or are there any real life situations, where there is such a gm (I cannot think of any right now)?

[@Gametheory101]

mariorpg11 >Is the gm just some kind of abstract conzept we need for understanding the theory? This is the correct way to think about it.

[@bobsmeth1404]

fuiyg+William Spaniel

[@sarasojha]

Bob Smith