Abstract
In an epistemic framework due to Aumann we characterize the condition on the rationality of the players that is both necessary and sufficient to imply backward induction in perfect information games in agent form. This condition requires each player to know that the players are rational at later, but not at previous, decision nodes.
Original language | English |
---|---|
Pages (from-to) | 325-345 |
Number of pages | 21 |
Journal | Journal of Mathematical Economics |
Volume | 27 |
Issue number | 3 |
DOIs | |
State | Published - Apr 1997 |
Keywords
- Backward induction
- Common knowledge
- Extensive games
- Subgame perfection