|
Abstract
|
This study presents a laboratory experiment investigating the relationship between game complexity and individuals' higher-order rationality. According to Level-K theory, individuals' rational choices are assumed to be independent of the decision environment, with subjects always expected to identify the optimal response based on their beliefs. However, as games increase in complexity, accurately determining the best response becomes more challenging. We argue that certain game characteristics, such as complexity, can influence individuals' higher-order rationality.
Using a within subject design, where participants played against computer-controlled opponents, we rule out the belief-bounded subjects and focus purely on ability-bounded subjects. Our experiment used a modified Ring Game structure where we can modify the complexity of games and we examined three dimensions of complexity, namely the structure of the game, the content of the payoff matrix, and the salience of specific strategies.
Preliminary analysis of our experimental data suggests that game complexity does indeed affect individuals' higher-order rationality. Increased randomness in the information available was found to lower individuals' strategic levels, while greater salience in iterative reasoning raised them. Additionally, prior experience in game theory significantly influenced subjects' levels of rationality, with more experienced participants achieving higher levels.
This paper contributes to the understanding of behaviour in both strategic interactions and individual decision-making processes. Our finding suggest that higher order rationality is content dependent, complexity of game influences individuals’ strategic reasoning, a phenomenon not fully accounted for by Level-K theory. Our future research will investigate how game complexity shapes belief formation in strategic decision-making.
|