|
PureBytes Links
Trading Reference Links
|
BobR,
Last year I did some work with the PFE indicator. I am intrigued by it,
with its advanced math and its unique sawtooth pattern. When it gives a
signal, there is no squinting at it to determine whether or not it has
turned down or up. The problem is to filter out the false signals.
Overlapping time frames is a good approach to that end.
The PFE comes with a default lookback period of 9 bars (length of 9), like
a 9 bar MA.
It is easy to program this as an input so that you can change the length
of the PFE lookback period. I have done so and attached a chart of the
OEX to illustrate.
Visible as data1 are 3 minute bars of the OEX, while a hidden data2 has 5
minute bars.
The Red PFE is the standard lookback of 9 (3 minute bars.
The Yellow PFE superimposed is also the standard, but calculated on the 5
minute bars of data2.
The White PFE is my variable lookback, with its input set to a 15 bar
period. It is calculated on data1 or the 3 minute bars. Thus it has a
total lookback period of 45 minutes (3 min. x 15 bars).
Now notice that this is the same lookback period as the Yellow PFE (5 min.
x 9 bars).
The plot differences are explained by the greater number of bars in the
White PFE.
This gives it more to work with and thus creates a more characteristic
sawtooth pattern.
If this is of interest to you, let me know.
Neal
----------
> From: bobrabcd <bobrabcd@xxxxxxxxxxxxxxx>
> To: realtraders@xxxxxxxxxxxxxx
> Subject: GEN - PFE indicator
> Date: Wednesday, February 11, 1998 11:26 AM
>
> Has anyone done any studies of combining PFE, polarized fractal
efficiency
> ratios, of different time frames for timing purposes. Attached are two
> charts. One is a combo of daily and hourly. The other is of 3 min, 5,
15
> min periods. A convergence of the pfe's seems to mark a trend change.
On
> the daily and hourly we now have both in "ovebought" so we need to watch
> for a downturn of the hourly.
>
> BobR
>
Attachment Converted: "c:\eudora\attach\OEX.gif"
|