Exploring Trading Strategies and their Effects in the FX Market

Agent-based computational economics acknowledges the distributed nature of trading in financial markets by modeling the markets as evolving systems of autonomous, interacting agents that correspond to the trading parties. Conventionally, the behavior of traders has been described mathematically, and the market system is analyzed at equilibrium conditions. The dynamics of price formation, however, is influenced by the large diversity in the cognitive structures of the traders (e.g. differences in decision making methods, interpretation of available information and learning capacity), their specific circumstances (e.g. attitude to risk, time horizon) and the organization of the specific market in which the traders operate (e.g. market microstructure). Therefore, we propose to study financial markets by using intelligent agents that have rich cognitive structures borrowed from artificial intelligence research for modeling their decision making behavior. This representation allows us to model the decision making behavior of agents in terms of algorithms, that can represent a more diverse set of behaviors than mathematical formulae only. We discuss the role of intelligent-agents in the analysis of financial markets and speculate on the type of agents that can be expected to be suitable for the analysis and simulation of financial markets. We elucidate our thoughts by exposing the outline of a research project that has started recently at our university. As a first step of our research project, we discuss a classification of adaptation that we proposed recently for agents in agent-based computational economics.

2008

This paper investigates market microstructure in an experimental artificial futures market using the data generated by the submitted trading agents in two laboratory experiments. Since our previous analyses reveal that the price series successfully replicate so-called "stylized facts" in some regards, the aim of this study is to check whether such phenomena are also observed at microlevel. Our empirical results confirm that although there are locally mispricing effects of several trading agents, in most cases the market liquidity improves over the rounds. Future perspectives of agent-based computational finance and experimental economics are also discussed.

Proc. of the UK Computational Intelligence Workshop 2004 (UKCI2004), Loughborough, UK, 62-69, 2004

This paper investigates the effect of using varying amounts of training data on the specificity and robustness of a multi-agent based solution for use in trading simulations using historical equity market data. Three separate amounts of training data were used in five experiments to evolve 15 groups of agents under varying conditions. These groups were then exposed to three separate test environments to determine whether differences in performance could be related back to their training environment. The results indicate that larger training data sets lead to more general solutions and overall better performance when tested in environments with varying conditions. This will lead onto future research focusing on agent behavior and decisions in financial markets.

2004

Stock markets strive to provide an ecient trading platform for investors. Trading rules and mechanisms issued to accomplish this dier among stock markets, and are subject to modification over time. Furthermore, market participants assume a broad range of roles and trading strategies. Such vari- ation poses problems to those involved in the study of market dynamics, when developing an artificial

Proceedings of the 1999 Congress on Evolutionary Computation-CEC99 (Cat. No. 99TH8406), 1999

In this paper, we propose a new architecture to study arti cial stock markets. This architecture rests on a mechanism called school" which i s a procedure to map the phenotype to the genotype or, in plain English, to uncover the secret of success. We propose an agent-based model of school", and consider school as an evolving population driven by single-population GP SGP. The architecture also takes into consideration traders' search behavior. By simulated annealing, traders' search density can be connected to psychological factors, such a s peer pressure or economic factors such a s t h e standard of living. This market architecture was then implemented in a standard arti cial stock market. Our econometric study of the resultant arti cial time series evidences that the return series is independently and identically distributed iid, and hence supports the e cient market hypothesis EMH. What is interesting though is that this iid series was generated by traders, who do not believe in the EMH at all. In fact, our study indicates that many of our traders were able to nd useful signals quite often from business school, even though these signals were short-lived.

Journal of Mathematical Economics, 2005

This paper revisits the Kareken-Wallace model of exchange rate formation in a twocountry overlapping generations world. Following the seminal paper by Arifovic (Journal of Political Economy, 104, 1996, 510 -541) we investigate a dynamic version of the model in which agents' decision rules are updated using genetic algorithms. Our main interest is in whether the equilibrium dynamics resulting from this learning process helps to explain the main stylized facts of free-floating exchange rates (unit roots in levels together with fat tails in returns and volatility clustering). Our time series analysis of simulated data indicates that for particular parameterizations, the characteristics of the exchange rate dynamics are, in fact, very similar to those of empirical data. The similarity appears to be quite insensitive with respect to some of the ingredients of the GA algorithm (i.e. utility-based versus rank-based or tournament selection, binary or real coding). However, appearance or not of realistic time series characteristics depends crucially on the mutation probability (which should be low) and the number of agents (not more than about 1000). With a larger population, this collective learning dynamics looses its realistic appearance and instead exhibits regular periodic oscillations of the agents' choice variables.

SSRN Electronic Journal, 2001

We construct a computer simulation of a repeated double-auction market, designed to match those in experimental-market settings with human subjects, to model complex interactions among arti cially-intelligent traders endowed with varying degrees of learning capabilities. In the course of six di erent experimental designs, we i n vestigate a number of features of our agent-based model: the price e ciency of the market, the speed at which prices converge to the rational expectations equilibrium price, the dynamics of the distribution of wealth among the di erent types of AI-agents, trading volume, bid/ask spreads, and other aspects of market dynamics. We are able to replicate several ndings of human-based experimental markets, however, we also nd intriguing di erences between agent-based and human-based experiments.

Progress in Artificial Economics, 2010

This paper investigates the question of the sophistication level, in the mean of behavior and intelligence, one should endow artificial traders in order to obtain, with realistic market microstructure, not only both qualitative and quantitative stylized facts in an artificial market. For this purpose, we introduce an agent-based simulation environment with an architecture close to the Euronext-NYSE Stock Exchange. Series of experiments with different kinds of agents' behavior and trading framework specifications where realized within this environment. The results indicate that only special calibrations provide realistic stylized facts with coherent quantitative levels. We introduce a new type of agents, called in this paper "strongly calibrated agents", with their specific environment design, that provide price dynamics in quantitative and qualitative accordance with real stock market characteristics.

Journal of Economic Dynamics and Control, 2001

In this paper, we propose a new architecture to study arti"cial stock markets. This architecture rests on a mechanism called &school' which is a procedure to map the phenotype to the genotype or, in plain English, to uncover the secret of success. We propose an agent-based model of &school', and consider school as an evolving population driven by single-population GP (SGP). The architecture also takes into consideration traders' search behavior. By simulated annealing, traders' search density can be connected to psychological factors, such as peer pressure or economic factors such as the standard of living. This market architecture was then implemented in a standard arti"cial stock market. Our econometric study of the resultant arti"cial time series evidences that the return series is independently and identically distributed (iid), and hence supports the e$cient market hypothesis (EMH). : S 0 1 6 5 -1 8 8 9 ( 0 0 ) 0 0 0 3 0 -0 that many of our traders were able to "nd useful signals quite often from business school, even though these signals were short-lived.

… of the 2009 IEEE/WIC/ACM …, 2009

This paper describes our experience in building an evolutionary system for agent-based modeling of a financial market. The system uses a type of BDI agents, which are deliberative agents with a mental state defined by a belief base and by a set of desire-generation rules. Beliefs are graded and trust in information sources is taken into account. At any moment, an agent generates a set of desires and selects a consistent subset thereof, whose elements are adopted as goals, to be achieved by executing actions.