Trading Agents Competing: Performance, Progress, and Market Effectiveness

  • Authors:
  • Michael P. Wellman;Shih-Fen Cheng;Daniel M. Reeves;Kevin M. Lochner

  • Affiliations:
  • University of Michigan Artificial Intelligence Laboratory;University of Michigan Artificial Intelligence Laboratory;University of Michigan Artificial Intelligence Laboratory;University of Michigan Artificial Intelligence Laboratory

  • Venue:
  • IEEE Intelligent Systems
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

Held annually since 2000, the annual Trading Agent Competition provides a forum for designersto evaluate programmed trading techniques in a challenging market scenario. Using three years oftournament data, the authors attempt to evaluate competitors' trading competence and progress in the2002 tournament and over time. Although it's difficult to assess absolute individual performancemeasures, we can directly analyze relative and overall market performance measures. This articlequantifies the TAC travel market's effectiveness in terms of allocative efficiency and finds improvementwithin and between tournaments. Comparing these results with alternative allocation benchmarks lets uscalibrate this efficiency and identify opportunities for further gains.