We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.GT
< prev | next >
new | recent | 2001

Change to browse by:

cs
cs.AI
cs.MA
econ
econ.TH

References & Citations

DBLP - CS Bibliography

listing | bibtex

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Computer Science > Computer Science and Game Theory

Title: Fictitious Play Outperforms Counterfactual Regret Minimization

Authors: Sam Ganzfried
(Submitted on 30 Jan 2020 (v1), revised 23 Jul 2020 (this version, v5), latest version 18 Dec 2023 (v8))
Abstract: We compare the performance of two popular algorithms, fictitious play and counterfactual regret minimization, in approximating Nash equilibrium in multiplayer games. Despite recent success of counterfactual regret minimization in multiplayer poker and conjectures of its superiority, we show that fictitious play leads to improved Nash equilibrium approximation over a variety of game classes and sizes.
Comments: Fixed a bug in the 5-player CFR implementation from prior version and reran the 5-player experiments
Subjects: Computer Science and Game Theory (cs.GT); Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA); Theoretical Economics (econ.TH)
Cite as: arXiv:2001.11165 [cs.GT]
  (or arXiv:2001.11165v5 [cs.GT] for this version)

Submission history

From: Sam Ganzfried [view email]
[v1] Thu, 30 Jan 2020 03:47:09 GMT (8kb)
[v2] Fri, 14 Feb 2020 18:55:50 GMT (8kb)
[v3] Wed, 8 Apr 2020 00:06:01 GMT (9kb)
[v4] Mon, 18 May 2020 22:01:27 GMT (9kb)
[v5] Thu, 23 Jul 2020 19:02:20 GMT (9kb)
[v6] Wed, 1 Feb 2023 07:41:09 GMT (9kb)
[v7] Mon, 27 Mar 2023 03:47:17 GMT (9kb)
[v8] Mon, 18 Dec 2023 19:42:29 GMT (577kb,D)
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Link back to: arXiv, form interface, contact.