Stocks & Commodities V. 33:1 (16-22): Whiter Is Brighter by John F. Ehlers

Whiter Is Brighter by John F. Ehlers

Relax, It’s Just Noise

White noise, pink noise — does it make a difference? It sure does, and here is how you can use noise theory to create an indicator with zero lag that works both as a countertrend oscillator and as a trend identifier.

Noise spectra are often described by color. Just like white light, white noise contains all frequency components having equal power. When I surveyed analysts’ description of market data, it led me to the conclusion that they describe it as pink noise. Pink noise is noise with memory, and I use this noise-with-memory concept to derive a market data synthesizer.

HERE'S WHAT I DID TO MAKE IT USEFUL

I turned the data synthesizer inside out to derive a white spectrum from the observed data. While simple, it is profound, because whitening the data negates the phenomenon I have called spectral dilation. Just because the derived white noise spectrum has uniform amplitude at all frequencies, it does not preclude the waveform from containing information. I then show how to extract the information in the waveform using a SuperSmoother filter. Since the derived waveform has a uniform power density at all frequencies, a range of optimum indicators can be created. These optimum indicators can be used on intraday data, in short-term trading strategies, and in longer-term trend followers.