|
PureBytes Links
Trading Reference Links
|
Chris,
Is Ehlers' approach described in recent article, or book? I think you'll
need
to generally detrend the data before processing it. This is just one of
those
conventional steps used in signal analysis.
Alternatively, you could take the differences between each point and average
those. This idea works off the theory that differences approximate the first
derivative, and the first derivative of a sinusoidal function is still a
sinusoidal function, with the phase shifted 90 degrees. However, the sum
will
be zero if you're looking at summing the differences.
> -----Original Message-----
> From: Chris Evans [mailto:evanscje@xxxxxxxxxxxxx]
> Sent: Wednesday, January 07, 2004 8:34 AM
> To: Gary Funck
> Cc: Omega List
> Subject: Re: Cycle Length
>
>
> I agree I do want to just find the minimum ... not a +-1% level
> but I'd rather not detrend .. I want the length to grow if a
> trend is in place (a little like Vima)
>
|