The chart at the top of this page deals with the
long term trends, and the page at this link
<DJIA: Near Fibo. 5/25/04>
considers only the
near term "expectations." Needless to say, the two are connected at
the hip, and I urge everyone to read both if they have the time.
The long term
chart of the DJIA index, below, shows an excellent example of another
important tenet (doctrine) of our Fibonacci Trend Theory. When placing
a Fibonacci trend we want to "locate" a trendline that touches as many
of the points where actual price crosses over, touches, or bounces off
a Fibonacci number line and thereby keep the "open gaps" to a minimum.
An "open gap" is those places where actual price "did not"
touch the "hypothetical" Fibonacci trendline and the Fibonacci number
line at the same time. This "gap" is illustrated on the top chart
by the area at Fibonacci number line 1597, in 1977. If we are "assuming"
that the actual regression mean line is our "hypothetical" Fibonacci trend
line, then we can see that we have two large open areas, or "gaps," during
1940-1960 and then again from 1965 to 1985, where price (index value)
"does not" touch our mean trendline AND the Fibonacci number lines
"at the same time."
While a "perfect" Fibonacci trendline should not have
any "Fibonacci gaps," that is "fairly rare" occurrence in actual practice.
That is to say, in theory, there "should be" a straight line (in semi-log
scale only) that connects the lowest "average" price to the highest "average"
price and which passes through, or otherwise touches, very Fibonacci
line between that high and low, AND the "time period" that elapses between
each "crossing" MUST BE equal or at least "close to" being equal.
Now, if, and only if, the time "steps" are the same,
and "usually" a whole number in years, and the price passes through the
Fibonacci line numbers at those time "steps," do we have a CONTINUOUS
MATHEMATICAL FUNCTION that is an "exponential growth function," and a
Fibonacci summation series that is also a exponential function.
Notice that, in the graphic, the "time steps" for the main
trend (yellow dots from 1900-2020) are "about" the same (about 8.8 years),
and the time steps of the three "spur trends" (blue dots for 1900-1926,
1932-1968, and 1974-2008?) have "about" the same time between each step for
it's own trend. Therefore, they are all continuous functions, exponential
growth curves, and a Fibonacci summation series. By the way, another
tenet of our theory is that that these "time steps" will, eventually," regress
to a whole number (integer) in years, like 1y-2y-3y or ect...with a reasonable
degree of tolerance (leeway) assumed.
Unfortunately, the real world
is not as nice and clean, or accurate, as the hypothetical basis of our
theory, so in actuality we have to find the trend that best meets the
most criteria while not expecting perfect patterns. The underlying
mathematical basis of the Fibonacci Trend Theory is the exponential growth
function, and the underlying basis of this exponential growth function
is the compounding interest equation.
However, we cannot expect a stock or index to "always" follow
it's "hypothetical" long term trend, and it would be totally erroneous
to assume that price (index value) "must always" touch that trend at the
time when it crosses the Fibonacci number lines. The reality is that
this does happen so often that we, at PTR anyway, have come to expect it;
eventhough, we know better then most that you have to "assume" some "range,"
or tolerance, when dealing with this form of analysis.
While this manual "fit" to locate the "likely" future
mean trend, which may or may not be the same as the existing regression
mean trend, and it looks very good, I could be totally wrong about this,
and price could continue to trend up for many more years.
While I believe that is "highly unlikely," I do not have any
crystal ball that will actually predict the future, and I firmly believe
that the future IS NOT preordained to some exact specification, or if it
is, then GOD is the only one who knows what it will be and he is not confiding
in me...as far as I know!
Before I leave this analysis, I want to point out a
few historic facts about the "deka-boundry" lines at 100, 1000, 10000.
As you can see, on the top chart, the DJIA price (index value)
was "hung up" in the area around these lines for many years after it initially
reached those lines in a bull leg up. The index was then "hung up"
around 100 from 1922 to 1942 (20 years), around 1000 from 1966 to 1982 (16
years), and now it has only been "hung up" around 10000 from 1999 to 2004
(five years).
While anything is possible, and no where is it chiseled in
stone that stock market history "must" repeat itself, I "highly suspect"
that this is another piece of circumstantial evidence that supports the
bearish point of view. On the flip side of this thesis, who is
to say that we cannot make another run up to a double top, or even "minor"
new highs, without actually "pulling away from" this 10,000 deka boundary.
While it's possible that we could even have a blow out rally up through
10k that then turns into a huge crash back down, to repeat the action near
100 in 1926 to 1942, I suspect that senerio is also very-very remote.
By the way, let me point out
some key observations related to this chart that have a strong bearing
on our overall analysis. For the long term bias, I'm very close to
switching our wave count from the preferred to the alternate, for THIS
INDEX ONLY, which is a major switch from being intermediate and long term
bearish to intermediate term bullish and long term bearish.
The reason for this is illustrated on this chart, as
I will explain. As you can see, the period from 1932 to 1968 covered
36 years up, and passed through seven Fibonacci number lines (55-987),
which we "count" as SIX (6) segments. The rally up from 1974 has also passed
up "through" seven Fibonacci number lines (from 610-10946), another SIX
(6) SEGMENTS, but it has only covered 32 years to date (six "steps"). If
we assume that we will get another 36 years to 10,946 (in which the index
could go above and come back down too...again.), then we get some "excellent"
ratios and measurements in both price (index value) and time.
This are as follows: 1) 1932 to 1966 (the Elliott high)
is 34 years and 34 years added to 1974 yields 2008. 2) The gain in
% from the 1932 low to the 1966 high was "about" 2400%, and if we advance
the 1974 low, at 570, by 2400% we get a final high target near 13,680-13,900,
which is not so far away that price can't be sucked back down to 10,000
again (like 1933-1938?). Also, we would have a major cycle trough
"setup," with 1932 to 1974 being 42 years, and 1974 plus 42 years would
project the major cycle low out at 2016, which is not a four year cycle
low, but could be a final higher low made to a "deep" low in 2010 and/or
2014.
Our targets for both counts are shown on the top chart, which
all have the same index value, at Fibonacci number line 6,765, but with
time being a real toss up at any of the next four year cycle lows (about
2006-2010-2014), or that 2016 higher low senerio I just discussed.
One more, and "last," point
for those who fully understand Elliott Wave Theory. That alternate
count, which we "may," or may not soon make the preferred count, would
be formed as follows: The 1932 low was the grand super cycle low,
a wave two as (((II))). Wave one of three up, ((1)) of (III),
was the rally up from 1932 to 1937, and wave two of three, ((2)) of (III),
was the decline down from 1937 to 1942. Then we have: wave three of
three, ((3)) of (III), running up from the 1942 low to the 1966 high, and
then wave four of three, ((4)), cycle degree, of ((III)), super-cycle degree,
declining from 1966 to 1974.
After that, wave one of five going up, (1) of ((5)),
now subdividing into a 5th wave extension, runs from 1974 to 1976. The
wave two of five decline is from 1976 to 1978, and then wave three of five
subdivides again and runs from 1978 to 2000. After that, we have
the wave three of five top in 2000 as (3), intermediate degree, of ((5)),
cycle degree, and then the wave four declines down, as a simple A-B-C,
into a major low in October of 2002.
From that low, the DOW is either still in wave four sideways
and down to a 2006 low, as a W-X-Y from the 2000 high, or we are already
heading up into 2008 as wave five of five of of super-cycle three up, (5)
of ((5)) of (III). While it "looks" more like we are starting into
wave five up, rather than still working that wave four down from 2000, I
have very serious doubts about this whole scenario, but expect it to become
much more likely if the NeoCons can steal another election this fall.
As I see it, the key point for this alternate count is that the 1974 to
1976 rally as wave one of five, (1) of ((5)), and we are trying produce
the final wave five up, (5) of ((5)), to match that wave one now. If
you look closely at the chart, you can see that if the trendline placed
across the 1976 high to that 2000 high were projected down and parallel,
in a channel, it "could" be confirming this count; eventhough, it
"looks like anything going out past the end of this year, or maybe late
2005 at the most, is going to be outside the channel unless this baby is
going to explode to that 13,000-13,900 "target" in a hurry, which we find
very hard to believe given that the "strong" two year cycle in semi-conductors
has either just turned down hard or is doing so now.
On the flip side of that alternate senerio (wave count)
, is our preferred senerio, which is explained in a special introduction
to all subscribers after they sign up for our service, but before they
reach PTR's main menu. By the way, if you like that alternate wave
count...don't! It's still highly unlikely since the highest rate
of change came in 1986 and 1987, which means that period was "very likely"
a wave three of three, ((3)) of (III), and not any wave three of five, (3)
of ((5)) of (III).
If this is proven true, probably by election day or shortly
thereafter, or the spring of 2005 at the latest, then this rally, either
still in progress or about to terminate, "is "very likely" to be a Bear Market,
"counter trend," rally, and the Big Bear will be back in control before
too long; eventhough, we can't currently "see" what pattern it is trying
to develop.
Oh, by
the way, even the "alternate count has an alternate, and it calls for top
here in late 2004, near election day, by the late spring of 200, or by 9/11
of 2005 at the latest. This top would "most likely" be either a "double
top" to the 2000 highs, or a "very slightly" higher high.
It would also be a "final wave five of super-cycle wave three
coming up from the 1974 low, and as wave nine (9) of a long subdivision
in wave three of three ((3)) of (III), which run from the 1982 low to the
2000 high. That way, wave ((5)) would have began at the 10/2002 OR
3/2003 low and would run up for two to two and one-half years (30m) to match
the wave one ((1)) of (III), which run up from 1974 to 1976 and lasted
"about" two years. Furthermore, If you look at the chart above, you
can see that a final 5th wave (actually wave 9 of 9), would "fit" inside
the parallel channel up from 1974 to 2000 a lot better than anything going
out as far as 2008.
In summary, while we do not have an actual Crystal Ball to tell
us which one of our three scenarios are "the real deal," if any, at least
one of them "should be" eliminated by the end of 2004, the late spring
of 2005, or 9/11/2005 at the latest, and my "circumstantial evidence"
still points more toward the Bear rally thesis than the bull struggle
up into 2008 thesis; eventhough, the margin of difference is still small
and confusing.
Oh, one more "thing," for the longwave Fibonacci
Trend(tm): while the Bull Market up from a 1938 low to a 1966-1968
high would have only crossed five segments (six lines or 89 to 987), verses
the six segments (seven lines) from 1974 to 2004 (570-10,946), the time
would be 30 years in both cases and that is both a well know cycle for PE
ratios and W.D. Gann's "master cycle."
For the P
RICE- TIME Review
Andrew J. Quiggly
Editor
All content is copyright(2004) PriceTime
LLC
|
