摘要:The paper discusses the role of memory in an asset pricing model with heterogeneous beliefs. In particular, we were interested in how memory in the fitness measure affects the stability of evolutionary adaptive systems and the survival of technical trading. In order to obtain an insight into this matter, two cases were analyzed: a two-type case of fundamentalists versus contrarians and a three-type case of fundamentalists versus opposite biases. It has been established that increasing memory strength has a stabilizing effect on dynamics, though it is not able to eliminate speculative traders’ short-run profit-seeking behaviour from the market. Furthermore, opposite biases do not seem to lead to chaotic dynamics, even when there are no costs for fundamentalists. Apparently some (strong) trend extrapolator beliefs are needed in order to trigger chaotic asset price fluctuations.