Market Time DataTM Improving Technical Analysis and Technical Trading

The purpose of this paper is to demonstrate that by changing the underlying data used in technical analysis and technical trading systems the performance of these techniques can be greatly improved. We present two techniques – one for real-time (intraday) data and one for standard daily (end-of-day) data. For high frequency data we present a dynamic sampling technique that can generate a time series that ranges from 5 minutes to many hours in our time scale, Market Time. We present empirical evidence that this modified series is superior for technical analysis and trading. This is based on correlation studies, directional forecasts, and profitability of trading rules for FX rates. In the daily data realm a different technique is used then for intraday data, but it is based on the same first principles. Using the FxDx currency index as a test vehicle, the new daily data series is evaluated and shows increased performance and reduced drawdown. The average increase in performance is greater than 1% per year over the FxDx benchmark. Finally, two other trend following systems are presented that show performance increases when using Daily Market Time DataTM.