Digital Signal Processing Filters and Cycle-Based Trading
Summary
The article introduces digital signal processing concepts for trading, including cycle period, frequency, amplitude, and phase. It explains how low-pass, high-pass, and band-pass filters emphasize or suppress different cycle lengths, and how stacking filters can improve selectivity while adding lag. It also describes estimating a dominant cycle and using a filtered price series with a leading trigger line to generate crossover trades.
Examples apply these ideas to currency pairs and an equity index across several timeframes, and show frequency spectra as a way to inspect signal components. The reported backtests vary by market, timeframe, and strategy version, with results that are mixed rather than consistently positive. The article notes that the dominant-period estimate lags by about ten bars and that its crossover behavior can fail when prices trend. Spectral patterns may also be random, so the examples illustrate methods to investigate rather than establish a robust trading edge.
Key ideas
- Low-pass, high-pass, and band-pass filters select for different cycle lengths in a price series.
- Stacking filters can sharpen cycle selection but may introduce additional lag.
- A dominant-cycle estimate can guide the center and bandwidth of a band-pass filter.
- A filtered signal and trigger line can generate entries when they cross, though trending markets can undermine the expected turning points.
- Spectral analysis decomposes a signal into cycles, but apparent patterns may be random.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.