I promised a follow-up on my article Wave within Waves. It seems the idea has merits. The one question not answered was: How many different timeframes should I implement to get a reliable reading. My first implementation used 3 different settings, as these were the oscillators I always had on screen. I now expanded that to 10 different settings based on the Fibonacci number series and added a second independent oscillator, which I setup the same way.
This is what I got (crude oil July 09 60min):
and here is a crude oil July 09, 5min chart
Even if both oscillators forming my new Wave oscillator are independent from one another, both are scaled to use a mid-line and the scale seen on the indicator is fixed, so crosses of the lines are not just arbitrary due to screen resolution or chart size.
I'm now looking working with the assumption, that readings above the mid-line represent uptrends, readings below represent an underlying downtrend.
Color-changes at extremes might signal a trendchange.
Color-changes near the mid-line might represent an entry opportunities with the trend.
Crosses of the oscillators might confirm the color change signals.
Divergence of the blue line against price near extremes may support a trendchange signal signaling exhaustion of the previous selling or buying momentum as prices are running into support or resistance levels.
I will have this setup on screen for a while to see, if these assumptions are correct. If yes, I will try to implement an ATS, which should impartially tell me wether these assumptions are correct or not. I won't do that in the first place, as an ATS requires very strict rules. And there is a tendency to tweak an ATS until it produces the desired results. First I need to get a feeling for the setup, I need to understand, if this setup produces good signals or not. Then I can start looking to implement the rules into a mechanical ATS.