A ‘Harmonic’ Approach To Currency Strength
Before I start: this is experimental – no actual trading setup!
About a week ago I came across John Ehlers’ latest book Cycle Analytics For Traders. Highly recommended for anybody who isn’t scared of some math on a higher level. I just couldn’t resist to read it and code some tiny bits instead of being wise and work on my EA 
.
Today I finished to wrap his Roofing Indicator in a function, so now I can integrate it as a one-liner into other projects. This lead to the Currency Strength indicator shown in the screenshot, which is based on a normalized Ehlers’ roofing algorithm.
I exchanged Ehlers’ original normalization algorithm (‘Automatic Gain Control’ AGC) for a simple division by price, because I found that the AGC added too much distortion to the original roofing signal, which is counterproductive after reducing spectral dilation effects by the roofing algo before.
Basically a roofing filter is a band-pass filter with a wide bandwidth. Default cutoff periods (Ehlers’ suggestions) are tuned to 10 bars low-pass and 48 bars high-pass, which means that faster as well as slower swings will be more or less suppressed.
The pic shows a display of 4 roofing filter based currency strength indicators tuned harmonically, so every indicator instance will display signals of one octave only. From top to bottom, the roofing filters bands are 4 – 8 bars, 8 – 16, 16 – 32, and 32 – 64.
Ehlers states that everything below 10 bars can’t be analyzed properly because of noise, but packaged in a cluster indicator it seems to work pretty well as an early warning for potential swing reversals. Has to be tested thouroughly! Maybe noise effects tend to annihilate each other statistically when summed up over several pairs per currency.
The 32 – 64 bars octave seems to be too slow to provide suitable signals, so maybe a combination of the upper three bands might be interesting to explore further.
All clusters are showing values for EUR, GBP, USD, CAD, JPY, AUD, NZD. At the moment, I’m skipping CHF by default in higher timeframes.
Maybe this one will serve as the default cluster indicator for the multipair EA I’m currently working on. Looks pretty promising, yet has to be tested.
Info about Ehlers’ book can be found here: http://eu.wiley.com/WileyCDA/WileyTitle/productCd-1118728513.html Regarding the roofing algorithm, there are several white papers and presentations freely available (I don’t have the links at the moment – sorry!)
Direct link to a free chapter provided by the publisher: http://media.wiley.com/product_data/excerpt/13/11187285/1118728513-10.pdf
Cheers, simplex
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
