As
can be seen from the orginal data and the Sine Regression Analysis that
we derived from it, this method will clearly "detect" a periodic cycle in
the data if one exist, and in this case, one clearly does exist at one year,
after the U.S. made the transition to a summer only peak, in 1985.
While this may look like the Holy Grail to locate cycles in chart
data, of any type...like stocks, I can assure you that's not the case.
From our own experience, it's clear that this method only returns
conclusive, or even "firm," results about 10% to 20% of the time, and even
then we, at PTR, will not "act on" a cycle unless it's confirmed by all
four of our methods: 1) sine regression, 2) Fast Fourier Transform, 3)
Digital Filter rejection, and 4) our Visual Elliptical fitting.
While only a small portion of this analysis turns out to be useful, it
has confirmed or disproved many "so-called" defact-o cycles in the U.S.
stock and bond markets.
For the PRICE-TIME Review
Andrew J. Quiggly
Editor
Copyright(2005) USA
PriceTime LLC
