Time

Timing is another important design issue in build­ing agent-based markets. In an agent-based market, timing refers to many things including the time span of the previous information considered by the agents.

Timing refers also to the order in which events occur. In other words, trading in most agent-based financial markets is synchronized by the designer. These models are based on the assumption that all trades take place between two discrete points in time, whereas in reality there are no fixed pe­riods. This is by no means a clear indication that this assumption is not realistic and the need to address this issue in agent-based financial markets is vital. Some examples of models with trading synchronization are Levy et al. (1994), Arthur et al. (1997), Farmer (1998), Zimmermann et al.

(2001a), Markose et al. (2003), and Martinez- Jaramillo and Tsang (2009).

Timing refers also to the frequency with which agents update their strategies. As can be observed from reality, traders update their trading rules over time, and this has an impact on their performance and as well as on the dynamic of market prices. For example, agents in LeBaron (2001b) and Martinez-Jaramillo and Tsang (2009) updated their strategies in the market, based on a fixed periodicity. On the other hand, agents in Markose et al. (2003), Gilli and Winker (2003), and Martinez-Jaramillo and Tsang (2009) updated in an endogenous way.

3.6.