[RT] Gen: FFT Phase Space

[Ronald McEwan <rmac@xxxxxxxx>]

This is the combined sine wave pattern of the FFT analysis I ran on the
AD data. I am using non linear analysis (via phase space plots) to
explore the dynamics of a time series that may not be immediately
available using traditional technical analysis. This is becoming, for me,
a very practical way of determining why one entity may be more tradable
then another. (also when and how to trade it). This also involves using
the "Largest Lyapunov" exponent . The Lyapunov exponent is the measure of
the dynamics of an "attractor". The measure of the loss of predictive
ability of an indicator going forward in time. An "attractor" can be seen
as the equilibrium level of a time series. Pivot points are a type of
attractor. A "Strange Attractor" is a level of equilibrium that is never
intersected. One analogy is the attraction of a mean reverting time
series to its moving average. Once price reaches the moving average
(attractor) the state of equilibrium is not quite achieved, or
maintained. The price action continues until another state of equilibrium
(attractor) is achieved, and so on. The detection of "strange attractors"
in a time series can be exploited in the development of trading models.
If they can not be detected in some form, I would rethink trying to trade
the security, or at least refine my model for trading this. A high tech
way of saying stick to what works best for you.

Ron McEwan   


Attachment Converted: "f:\eudora\attach\fft1.gif"