Talk:Bayesian game
Game theory Start‑class High‑importance | ||||||||||
|
Statistics Stub‑class Low‑importance | ||||||||||
|
Mathematics Stub‑class Low‑priority | ||||||||||
|
example PBE
In the example section, the following text appears: "In fact in the second game there is a perfect Bayesian equilibrium where...". Yet, there is only one game shown in the picture. Should the text be "In fact in the second stage there is a perfect Bayesian equilibrium where..." ?
- Not sure what that was about, got rid of "in the second game". Cretog8 (talk) 03:36, 2 June 2008 (UTC)
citations
The signaling example is based on an 1973 article by spence. here's the citation: Spence, A. M. (1973), 'Job Market Signaling', Quarterly Journal of Economics 87(3), 355--74.
would be nice if someone cared to add this citation to the article text... --1-is-blue (talk) 21:54, 24 March 2008 (UTC)
Types
"(while the word type is used, it is not actually defined in any game theory texts, and may be lost to the ages.)" This is kinda a distraction in the article. "Type" is a standard term for game theory, and so it's fine to use it without qualification. (BTW, it's defined In Myerson's Analysis of Conflict, "The initial private information that a player has at this point in time is called the type of the player." Cretog8 (talk) 00:16, 22 May 2008 (UTC)
- Sorry, that's Myerson's Game Theory: Analysis of Conflict.Cretog8 (talk) 00:28, 22 May 2008 (UTC)
As pointed out above, type is typically thought of as the current private information for a player. This means that one player's type can effect another player's reward function. This is in contrast to how the article presents types as a player's reward function. In other words I think this article needs to be changed to better reflect what a type is. Allliam (talk) 09:10, 10 October 2011 (UTC)
The example in "An example" doesn't require PBE
This game can be solved using deletion of dominated strategies and/or Nash equilibria. As such, it doesn't strike me as a very useful example of Perfect Bayesian Equilibrium in action. First, 2 can infer that 1 will never play U because the payoffs from D dominate the payoffs from D. 2 will always play U' because she knows that 1 will never play U (no matter how much 1 wants 2 to play D'). Second, by converting this game to normal form, the game is easily solvable, either by deletion of dominated strategies or by identifying the Nash equilibrium. A more useful example would demonstrate how PBE refines both Bayesian Nash Equilibrium and Nash equilibrium as solution concepts. I wish I could contribute further by thinking of a game that does this, but nothing's coming. Ideas? Abu Casey (talk) 23:56, 17 March 2011 (UTC)