Reinhard Selten's paper "Reexamination of the Perfectness Concept for Equilibrium Points in Extensive Games" revisits the concept of perfect equilibrium in extensive games. The paper argues that the earlier definition of perfectness, which excluded disequilibrium behavior in unreached subgames, was insufficient to address all issues related to unreached parts of the game. Selten introduces a new concept of perfect equilibrium, which is always subgame perfect but not vice versa. He demonstrates that every finite extensive game with perfect recall has at least one perfect equilibrium point. The paper also introduces an "agent normal form" as a more adequate representation of games with perfect recall compared to the normal form. Selten defines extensive games with perfect recall, where players can remember all past choices, and discusses the limitations of games without perfect recall. He introduces strategies, expected payoffs, and the normal form, and restates Kuhn's theorem on realization equivalence and payoff equivalence in games with perfect recall. The paper then delves into subgame perfect equilibrium points, where a strategy combination induces an equilibrium point in every subgame. A numerical example is provided to illustrate the limitations of subgame perfect equilibrium points, showing that some intuitively unreasonable equilibrium points are not excluded by this definition. Selten proposes a model based on slight mistakes, where players have a small probability of deviating from their strategies, and defines perfect equilibrium points as limits of equilibrium points in perturbed games. He proves that every perfect equilibrium point is subgame perfect and discusses the decentralized nature of perfect equilibrium points, showing that their existence can be decided locally at information sets.Reinhard Selten's paper "Reexamination of the Perfectness Concept for Equilibrium Points in Extensive Games" revisits the concept of perfect equilibrium in extensive games. The paper argues that the earlier definition of perfectness, which excluded disequilibrium behavior in unreached subgames, was insufficient to address all issues related to unreached parts of the game. Selten introduces a new concept of perfect equilibrium, which is always subgame perfect but not vice versa. He demonstrates that every finite extensive game with perfect recall has at least one perfect equilibrium point. The paper also introduces an "agent normal form" as a more adequate representation of games with perfect recall compared to the normal form. Selten defines extensive games with perfect recall, where players can remember all past choices, and discusses the limitations of games without perfect recall. He introduces strategies, expected payoffs, and the normal form, and restates Kuhn's theorem on realization equivalence and payoff equivalence in games with perfect recall. The paper then delves into subgame perfect equilibrium points, where a strategy combination induces an equilibrium point in every subgame. A numerical example is provided to illustrate the limitations of subgame perfect equilibrium points, showing that some intuitively unreasonable equilibrium points are not excluded by this definition. Selten proposes a model based on slight mistakes, where players have a small probability of deviating from their strategies, and defines perfect equilibrium points as limits of equilibrium points in perturbed games. He proves that every perfect equilibrium point is subgame perfect and discusses the decentralized nature of perfect equilibrium points, showing that their existence can be decided locally at information sets.