Optimism Without Regularization: Constant Regret in Zero-Sum Games

This paper studies the *optimistic* variant of Fictitious Play for learning in two-player zero-sum games. While it is known that Optimistic FTRL -- a *regularized* algorithm with a bounded stepsize parameter -- obtains constant regret in this setting, we show for the first time that similar, optimal rates are also achievable *without* regularization: we prove for two-strategy games that Optimistic Fictitious Play (using *any* tiebreaking rule) obtains only *constant regret*, providing surprising new evidence on the ability of *non*-no-regret algorithms for fast learning in games. Our proof technique leverages a geometric view of Optimistic Fictitious Play in the dual space of payoff vectors, where we show a certain energy function of the iterates remains bounded over time. Additionally, we also prove a regret *lower bound* of $\Omega(\sqrt{T})$ for *Alternating* Fictitious Play. In the unregularized regime, this separates the ability of optimism and alternation in achieving $o(\sqrt{T})$ regret.

Authors

• John Lazarsfeld
• Georgios Piliouras
• Ryann Sim
• Stratis Skoulakis

Keywords

No keywords are indexed for this paper.