Zero-Sum Fictitious Play Cannot Converge to a Point
Jaehong Moon

TL;DR
This paper demonstrates that fictitious play in zero-sum games can fail to converge to a single equilibrium point, revealing fundamental instability in the dynamics.
Contribution
It identifies specific geometric conditions under which fictitious play does not stabilize at a single equilibrium in zero-sum games.
Findings
Fictitious play can fail to converge to a point in certain zero-sum games.
Two geometric conditions (A1 and A2) prevent pointwise convergence.
Conjecture that these conditions alone may cause non-convergence broadly.
Abstract
Fictitious play (FP) is a history-based strategy to choose actions in normal-form games, where players best-respond to the empirical frequency of their opponents' past actions. While it is well-established that FP converges to the set of Nash equilibria (NE) in zero-sum games, the question of whether it converges to a single equilibrium point, especially when multiple equilibria exist, has remained an open challenge. In this paper, we establish that FP does not necessarily stabilize at a single equilibrium. Specifically, we identify a class of zero-sum games where pointwise convergence fails, regardless of the tie-breaking rules employed. We prove that two geometric conditions on the NE set (A1 and A2) and a technical assumption (A3) are sufficient to preclude pointwise convergence. Furthermore, we conjecture that the first two conditions alone may be sufficient to guarantee this…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
