Article ID: | iaor20132257 |
Volume: | 54 |
Issue: | 2 |
Start Page Number: | 838 |
End Page Number: | 845 |
Publication Date: | Jan 2013 |
Journal: | Decision Support Systems |
Authors: | Alvim Leandro G M, Milidi Ruy L |
Keywords: | stock market, least squares, human factors, high-frequency trading |
Automated traders operate market shares without human intervention. We propose a Trading Team based on atomic traders with opportunity detectors and simple effectors. The detectors signalize trading opportunities. For each trading signal, the effectors follow deterministic rules on when and what to trade in the market. The detectors are based on Partial Least Squares. We perform some trading experiments with twelve BM&FBovespa stocks. The empirical findings indicate that the proposed trading strategy reaches a 77.26% annualized profit, outperforming by 380.07% the chosen baseline strategy with a 16.07% profit. We also investigate Multistock Resolution Strategy (MSR) performance subject to brokerage commissions and income tax. Whenever the initial investment is at least US$ 50,000, the MSR strategy provides a profit of at least 38.63%.