Paper-Rock-Scissors: an Exploration of the Dynamics of Players' Strategies

Abstract

In the current study, we designed a Rock-Paper-Scissors game and collected data in an online 2-player experiment. Our results conclude that participants did not play the game optimally nor randomly. We also found that humans behavior are very heterogeneous, and cannot conclusively be described by a commonly claimed heuristic of Win-Stay/Lose-Shift (WSLS). Instead, we found evidence for different types of individuals: ones strategy selection is dependent on outcomes (dependence group) and another one revealed independent strategy selection (independence group). The cluster analyses further showed that the two groups performed differently as how to increase the chances of being unpredictable. The independence group altered their cyclic methods. The dependence group, on the other hand, showed individuals that preferred the classical win-stay and lose-shift heuristics; but also individuals that preferred win-shift and lose-stay heuristics. A plausible explanation of this apparently irrational behavior is that people might follow a more sophisticated reasoning, in the sense that they believe their opponent will expect them to follow the WSLS, and as a best response they decide to do the opposite. More research is needed to follow on this finding. We believe our research contributes to understanding the dynamics of players strategies in an adversarial setting.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 09, 2021
Accession Number
AD1142438

Entities

People

  • Cleotilde Gonzalez
  • Frédéric Moisan
  • Hanshu Zhang

Organizations

  • Carnegie Mellon University

Tags

DTIC Thesaurus Topics

  • Agreements
  • Behavioral Research
  • Chi Square Test
  • Clustering
  • Data Analysis
  • Demography
  • Dynamics
  • Education
  • Game Theory
  • Human Behavior
  • Institutional Review Board
  • Mathematics
  • Military Research
  • Personality
  • Probability
  • Recreation
  • Security
  • Surveys
  • Transitions
  • Universities
  • Web Applications

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Game Theory.
  • Regression Analysis.