Is the El Farol more efficient when cognitive rational agents have a larger memory size?

Recently, there has been an increasing interest in applying cognitive modelling approaches towards the understanding of the behaviour of agents in complex systems. In this paper we employ a cognitive modelling approach in a representative instance of such systems namely in the El Farol bar problem. We establish a relationship between agents and strategies, by defining the degree of belief of agents in the correctness of strategies, and devise a concept of rationality inspired in cognitive emotion theories, by conceptualizing the belief threshold by which agents must change its current strategy. We investigate whether a larger memory size increases market efficiency. We test this by systematically evaluating fifteen different treatments in which we gradually increase the memory size of cognitive agents. Interestingly, we found that the benefits of having a larger memory size are somehow limited. The results did not provide us with sufficient evidence to state that the El Farol bar is more efficient when cognitive rational agents have a larger memory size.