A statistical analysis of the trading agent competition 2001

ABSTRACT

We present a statistical analysis on the data from the Second Trading Agent Competition (TAC01). Our goal is to study the effectiveness of this competition as a testbed for the evaluation of trading strategies in artificial markets. Our results suggest that in TAC01 no agent performed significantly better than all the others, from a pure statistical viewpoint. Instead, it reveals groups of agents which performed significantly better than others. Thus, our results suggest that although this competition may not give a quantitative evaluation of the agents' trading strategies, it can still provide some useful qualitative evaluation.

REFERENCES

• 1: Glantz, S. A. and Slinker, B. K. 2001. Primer of Applied Regression & Analysis of Variance, 2nd ed. McGraw Hill, New York.
• 2: Strada, A. 2002. Autonomous agents for electronic auctions. M.S. thesis, Politecnico di Milano. Dipartimento di Elettronica e Informazione. (In Italian).
• 3: Wellman, M. P., Greenwald, A., Stone, P., and Wurman, P. R. 2001. The trading agent competition 2001. http://auction2.eecs.umich.edu.
• 4: Michael P. Wellman , Amy Greenwald , Peter Stone , Peter R. Wurman, The 2001 trading agent competition, Eighteenth national conference on Artificial intelligence, p.935-941, July 28-August 01, 2002, Edmonton, Alberta, Canada
• 5: Peter R. Wurman , Michael P. Wellman , William E. Walsh, The Michigan Internet AuctionBot: a configurable auction server for human and software agents, Proceedings of the second international conference on Autonomous agents, p.301-308, May 10-13, 1998, Minneapolis, Minnesota, United States [doi>10.1145/280765.280847]

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.3 INFORMATION STORAGE AND RETRIEVAL
      H.3.5 On-line Information Services
          Subjects: Web-based services

Additional Classification:
  D. Software
  D.2 SOFTWARE ENGINEERING
      D.2.4 Software/Program Verification
          Subjects: Statistical methods

  H. Information Systems
  H.4 INFORMATION SYSTEMS APPLICATIONS
      H.4.3 Communications Applications

          Nouns: Internet

  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE
        I.2.11 Distributed Artificial Intelligence
            Subjects: Multiagent systems

General Terms:
Economics, Management, Measurement

Keywords:
E-Commerce, agent, auction