Towards a non-linear trading strategy for financial time series

A new trading strategy based on state space reconstruction techniques is proposed. The technique uses the state space volume evolution and its rate of change as indicators. This methodology has been tested off-line using eighteen high-frequency foreign exchange time series with and without transaction costs. In our analysis an optimum mean value of approximately 25% gain may be obtained in those series without transaction costs and an optimum mean value of approximately 11% gain assuming 0.2% of costs in each transaction.