Market micro-structure analysis by multiagent simulation in X-Economy -- comparison among technical indices

We are developing an agent and server library referred to as X-Economy, by which we can execute multiagent simulations and network games for financial and economic systems. To this end, we analyzed the characteristics of network games in a financial context and compared them with traditional ones. X-Economy has also provided a new research direction in market micro-structure analysis. We executed several kinds of multiagent simulations for technical traders (indices) and obtained non-trivial suggestions regarding the relationship between the market randomness and the effectiveness of technical indices. For instance, the performance of complex technical indices seemed to deeply depend on the characteristics and nature of a market when a market became complex, i.e. it moved far from the Wiener process.