Perfect Bayesian Equilibria and Cross-Pair Independence from Common Actions

Preprint Aug 2026

Agnostic Sequential Rationality

Agnostic sequential equilibrium (ASE) is a refinement of sequential equilibrium that does not force on the players a single, arbitrary belief system. In addition, whereas sequential equilibrium assumes the players'beliefs to be fully consistent (a notion that is based on perturbations of strategies), ASE employs a nove...

Igal Milchtaich · 0 citations

Preprint Aug 2026

Rational Learning One Step Off the Path

Which social norms are self-correcting under rational learning? We show that conduct sustained by false beliefs cannot persist if a single departure from prevailing behavior generates evidence against those beliefs. We study the overlapping-generations learning model of Fudenberg and Levine (1993), in which finitely li...

Zihao Li, Minghao Pan · 0 citations

Preprint Sep 2026

The Robustness of Sequential Implementation under Information Perturbations

Bayesian monotonicity is a necessary condition for full implementation in Bayes-Nash equilibrium. Under the sequential equilibrium refinement, however, multi-stage mechanisms can expand the scope for implementation. We show that this additional implementation power is fragile. Whenever a multi-stage mechanism sequentia...

Ian Ball, D. Fudenberg, Stephen Morris · 0 citations

#machine learning Preprint Aug 2026

The concentration game: Bayesian updating, regret, and information

A two-player zero-sum repeated game between a learner and nature whose value identity generates Bayesian updating and an exact accounting of exponential-weights regret at once is given, and supplies the comparator-class variational form that a wide class of concentration phenomena share.

Akshay Balsubramani · 0 citations

Open access Aug 2026

Two-person zero-sum games without expected utility preferences: a proposal

A solution concept for two-person zero-sum games is proposed with only the axiom of Independence assumed of players’ preferences. Each player is shown to admit a set of admissible strategies assuring minimum guarantees. Rationality requires agents to reject non-admissible strategies in further considerations. Knowledge...

Siddharth Chatterjee · International Journal of Gam... · 0 citations

Preprint Sep 2026

Partially Observed Mean Field Games Without Perfect Recall: Optimality Conditions and Equilibria

This paper studies partially observed mean field games without perfect recall (WPR). The representative agent observes a noisy signal, but the control at time \(t\) uses only \(\mathcal G_t^I=\sigma(y_t)\), a generally non-nested information family. The conditional population law instead uses the observation filtration...

Xuan-Ping Zhang, Xiao Zhang, Wang Yao · 0 citations