Game Theory 101 (#66): Ex Ante and Interim Dominance

gametheory101.com/courses/game-theory-101/
While dominance is straightforward in complete information games, incomplete information complicates matters. This lecture explains the two different types of dominance in Bayesian games: ex ante dominance and interim dominance. Ex ante dominance focuses on the strategies of a player (before he has learned of his type) while interim dominance focuses on the strategies of a type. Interim dominance implies ex ante dominance but not vice versa.
As with complete information games, you should immediately eliminate any strictly dominated strategies, regardless of whether they are ex ante or interim dominated. Eliminating weakly dominated strategies can remove Bayesian Nash equilibria, so you still need to be careful with that.

[@PunmasterSTP]

My brain is pretty much on fire at this point. I have to say that the concepts introduced in the last few videos have really upped the challenge.

[@Gametheory101]

Yeah, it's surprising how much more difficult it gets when even a little uncertainty is introduced.

[@PunmasterSTP]

@@Gametheory101 For sure. I plan to keep watching your videos and studying until I get a good grasp on these things. In the...middle of my time doing that, I might check out some more puzzles. Thanks for posting all of this great stuff!

[@davide816]

i don't understand why type a should care about the choices of type b, type a will play R only if it is convenient for him. I write this because we made the assumption that every player knew his type.

[@dominiclebron6073]

theres a big vision illustrated @ 5:39 -> definitions of 'ex ante' & 'interim' graspable terms, nonetheless deep ideas with mathematical logic to back the claims made in game theory. professor vergas then concludes we should always being looking at these 'actors' - making judgements (using %) what type of game the other player plays (spaniel has been using P.D. & Stag Hunt- even tho there are a couple other games?) we can ALWAYS eliminate strickly dominated equilibria. eliminate emotion in some of these social decisions, replace w/ mathematically logical better payoffs. stay on top of yo game.

[@user-in4nk9jj1p]

I'm a little bit confused by the notion of "strategy" here. Could you please clearify what is a strategy here? To the best of my knowledge, a strategy of a player in a Bayesian game is a mapping from its type space to its action space.

[@raissanumerianodubourcqdan2155]

Why in this game we could remove left for P2 when P1 is PD right away? I thought that since que cannot say Right>Left for P1 type SH we could not remove it beforehand as we saw on previous videos

[@PunmasterSTP]

Sorry I'm not quite sure what you mean, and that's also probably because I'm a bit confused with things myself. If you could reference a specific timestamp in the video and provide additional information, I'll gladly take a shot at answering your question. Also, I hope that you have been doing well in the time since you watched this video!

[@colleenleigh9132]

sow good

[@aaaqqq10]

may i know what is the difference between AI and Game theory. . and which one is better for Cyber security

[@jacobbaca6606]

ha

[@ivoriankoua3916]

Too late but go with , AI when you'll be doing Multiple Agent Reinforcement Learning you'll have to deal with Game Theory.