Limitations of using real-world, public servers to estimate jitter tolerance of first person shooter games

  • Authors:
  • Grenville Armitage;Lawrence Stewart

  • Affiliations:
  • Swinburne University of Technology, Melbourne, Australia;Swinburne University of Technology, Melbourne, Australia

  • Venue:
  • Proceedings of the 2004 ACM SIGCHI International Conference on Advances in computer entertainment technology
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper quantitively evaluates the relevance of network jitter on player satisfaction and performance in multiplayer online games, particularly in comparison to overall network latency (or 'ping') times. We extend previously published work that showed Quake III players preferred servers less than 150-180ms away. Our modified, public Quake III server logged 20+ latency samples per second per client, from which we derived the jitter (instantaneous latency fluctuations) for each connected player. We found that using real-world traffic resulted in a highly correlated relationship between jitter and latency (jitter being generally one fifth, or less, of the path's latency), making it difficult to derive any independent relationship between jitter and player satisfaction. However, our results do demonstrate that absolute jitter over typical Internet paths is far less significant than the absolute latency for interactive games such as Quake III. This suggests Internet service providers (ISPs) should focus primarily on bringing latency within reasonable bounds.