Should be easy to do in MS, but is not…?
Well the big one that immediately springs to mind is J.M. Hurst. The
man who originally made cycles, envelopes and the simple moving average make
sense (to me at least). Trouble is that his techniques involve “centering”
the moving average so that it fits the data correctly. That means you have to
forward reference, which means that there is no data at the right hand side of
the chart for 0.5 the periodicity (0.5 * the cycle). Sound familiar? His method
involves just extrapolating the cycle forward to the right hand edge “by
eye” and then making the usual assumptions for the future data. Which,
despite all the modern technology (“The Profit Magic…” was
written in 1970) still seems to work!
It’s amazing how much clearer a chart can become when you overlay
a simple fixed-width (or maybe adaptive??) envelope around a *centred* moving
average…and get the periodicity right so you filter out the noise of the
lower-order cycles. It’s even clearer when you have two sets of envelopes…one
for the longer-term and one for a more ‘intermediate’ cyclic periodicity
(what Hurst would call the “trading cycle”). Strange that…how
a model invented in 1970 is still very relevant today. I wonder if that has
anything to do with the fact that it actually makes theoretical sense?
That’s not to say though that his methods can’t be improved
upon with our fancy new computers. Personally I would think there is some kind
of way forward in using Hurst for the longer cycles (big picture) and Ehler for
the shorter ones (entry timing). Adaptive techniques could only help. At the
moment I use pretty much Hurst
‘out of the box’ plus a support/resistance analysis. The actual
entries are a different matter though…I mainly use bog-standard S/R
breakouts (trendlines, pivot levels etc) following signals from my own divergence
indicators. Which is another area where adaptive techniques can be applied...
So, lots of ideas…
But where’s the time Brad, where’s the time???!!
From:
Metastockusers@xxxxxxxxxxxxxxx [mailto:Metastockusers@xxxxxxxxxxxxxxx] On Behalf Of bradulrich33
Sent: Monday, July 04, 2005 5:46
AM
To: Metastockusers@xxxxxxxxxxxxxxx
Subject: [Metastockusers] Re:
Stationarity and Real World application of statistics
That is right on teclogeo...I
wrote this last one before you wrote
yours...you do a better job of saying it...I can't
agree with you more.
and the part where you
mention..."over-complicate or compromise my the
rest of my strategy" is exactly what I
have found Metastock to do...
Please let me know if you know of any
"theoretically sound" ideas that
you think should be easy-to-use in Metastock, but
that are not..
YAHOO! GROUPS LINKS