Using weekly empirical probabilities in currency analysis and forecasting

Abstract

The Empirical Probability (EP) technique is proposed as an effective support tool to assist agents operating in a global fusion of financial markets. This technique facilitates the identification and prediction of primary, secondary and tertiary trends in addition to the recognition of trend reversals and the detection of changes in trend momentum. The suggested procedure is illustrated by deriving weekly (five-day) non-overlapping estimated probabilities from daily Euro/USD exchange rate data from 04/01/1999 to 31/12/2004 and applying these probabilities to the analysis and forecasting of exchange rate movements. In addition, trend characteristics of the data are used to develop a trading system that not only provides buy and sell indicators but also supplies directional probabilities associated with the signalled actions.