Super-Nash performance in games

Mehmet S. Ismail

Papers from arXiv.org

Abstract: Since the 1990s, artificial intelligence (AI) systems have achieved 'superhuman performance' in major zero-sum games, where winning has an unambiguous definition. However, most economic and social interactions are non-zero-sum, where measuring 'performance' is a non-trivial task. In this paper, I introduce a novel benchmark, super-Nash performance, and a solution concept, optimin, whereby every player maximizes their minimal payoff under unilateral profitable deviations of the others. Optimin achieves super-Nash performance in that, for every Nash equilibrium, there exists an optimin where each player not only receives but also guarantees super-Nash payoffs, even if other players deviate unilaterally and profitably from the optimin. Further, optimin generalizes and unifies several key results across domains: it coincides with (i) the maximin strategies in zero-sum games, and (ii) the core in cooperative games when the core is nonempty, though it exists even if the core is empty; additionally, optimin generalizes (iii) Nash equilibrium in $n$-person constant-sum games. Finally, optimin is consistent with the direction of non-Nash deviations in games in which cooperation has been extensively studied, including the finitely repeated prisoner's dilemma, the centipede game, the traveler's dilemma, and the finitely repeated public goods game.

Date: 2019-11, Revised 2023-09
New Economics Papers: this item is included in nep-gth and nep-mic
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (3)

Downloads: (external link)
http://arxiv.org/pdf/1912.00211 Latest version (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:arx:papers:1912.00211

Access Statistics for this paper

More papers in Papers from arXiv.org
Bibliographic data for series maintained by arXiv administrators ().