|
PureBytes Links
Trading Reference Links
|
I can't seem to let this one alone, so I apologize in advance (this is long),
but for those who hadn't noticed I wanted to chronicle some major recent fib
swings. I agree that minor short term swings may be self fulfilling, but I
feel that it is unlikely that very large, emotionally extreme pivots in the
cash indexes, are caused by a few traders with calculators.
Einstein said: "As far as the laws of mathematics refer to reality they are
not certain and as far as they are certain they do not refer to reality." OK:
fib analysis is not something I have coded, nor would I trade it on its own,
nor do I recommend anyone else trade it, and certainly I do not claim that
there is a fib rational for every swing. Indeed, most of my trading is not
based on fib patterns. Nonetheless, I find the golden ratio to be
fascinating, and I find it curious how people resist acknowledging its
persistent appearances in the market. I do not believe long term fib
relationships in the markets are coincidental or self fulfilling. FWIW, most
of my fib analysis has been based on price, rather than time, percentage of
winners, whatever else.
I've outlined most of the intermediate Nasdaq 100 swings since 1994 at my
site:
http://smarttrades.com/SmartCharts1.htm
I've also highlighted many of the swings since the March 2000 high in the NDX
as well as the DJI swings since 1992:
http://smarttrades.com/SmartCharts3.htm
If you do not want to see a sales pitch, do not click the links on those
pages. I sell my various services, but if you stay on the above pages you'll
be safe.
Below I've chronicled some of those corrections:
In 1987 the DJI hit a high near 2742. The 1932 low (an important low by most
accounts) was 42. A 38.2% retracement of that 55 year rally is 1710.6. The
actual print low (according to my CQG quotes at that time) was 1706.9 or
about .2% under a perfect 38.2%. Keep in mind this was in an environment
where the Dow had just lost over 10% in the last hour of trading and panic,
not cool calculation, ruled the day. Fib critics may want to use theoretical
lows, rather than the actual lows, to prove their point of view. <g>
The 1990 Dow peak near 3010 preceded the Persian Gulf war. The 1984 low (the
secondary kickoff low to the bull, which BTW corrected about 38%) is near
1079. A 38.2% retracement of that six year swing is 2273. The actual low was
2354. That one is off by 3.3%; I'd call that a miss. I'll blame it on the CNN
news team that broadcast live from Baghdad: those quick reports turned the
market "too" early. It was the first war to begin and "end" (in the mind of
the market) in about an hour.
In June 1996 the Nasdaq 100 (NDX) hit 704. I propose the NDX was becoming the
larger focus of the investment world about that time, and therefore better
represents the markets emotional swings. The end of the 1994 correction was
near 350. A 38.2% retracement of that rally was 568.8. The actual NDX low was
572.7 in July 1996 (off by .7%).
The October 1997 NDX peak was 1153.8. A 38.2% retracement of the rally from
the 1996 low of 572.7 was 931.8. There were two distinct lows that followed.
The October 1997 low was 926.9; the January 1998 low was 933. Both lows were
off a perfect 38.2% by .5% or less.
The July 1998 NDX peak was 1485.9. A 38.2% retracement of the rally off the
1994 low at 350 is 1052. The actual NDX low in October 1998 was 1063.2. That
one is off of perfection by a whopping 1%.
The March 2000 peak in the NDX was 4816. A 38.2% retracement of the rally
from the 1994 low (350) would be a drop of 1706 points. Funny that 1706
number was the 1987 print low in the Dow. I have no idea what to make of
that. Anyway, 4816 - 1706 = 3110. The April 17 low in the NDX was 3107.4.
That was the emotional (tax day) and momentum low (by many measures) of the
decline. That would be .08% off a perfect 38.2% (that's 8 hundredths of one
percent). The secondary low in May was 2897. A 38.2% retracement of the rally
from zero (it's a stretch) is 2976. The print low was 2897.2, off by 2.7%,
but it held 2976 on a daily close basis SO FAR (the low close was 3022.5) and
could only manage one hourly close (2940) under 2976. The weekly closing low
(so far) is 3101: .3% below the 3110 calculation.
So of nine intermediate term corrections, seven of them retraced previous
major swings by 38.2% +/- 1%. All nine corrections starting in 1984 have
retraced within 3.3% of 38.2%. The most emotionally dramatic declines had
turns within .2% of ideal 38.2% retracements. Given that these examples are
all in cash indexes, how would it be that the individual participants, in the
individual stocks, were all watching the same fib retracements and combined
to create a "self fulfilling" rally from such precise levels? In 1987, were
traders calculating fib levels from the 1932 lows (aside from me ;-)? I
believe the vast majority of players were unaware of these retracements as
they were happening. Most traders are not looking back to 1994 now, much less
1932 (in 1987), to calculate fib support. Sometimes "if it walks like a duck,
it is a duck."
The critics are sure to scoff and say the calculations are based on
subjectively selected swing lows. Sure, but that may be important information
as well. If you know a decline has stopped on a dime after correcting 38.2%
of a 55 year rally it may offer clues as to which rally is being corrected
and therefore indicate the magnitude of the next rally. Moreover, there is
fractal symmetry here as well: larger rallies corrected 38.2% relative to
1994 lows (NDX), while smaller internal rallies, within those larger moves,
corrected 38.2% from their own, higher starting points. The relationship of
each intermediate rally to its corrective phase since 1994 was nearly
identical and that relationship was the golden ratio.
Critics will say I've ignored 1966 through 1982 when there are few if any
38.2% corrections of major swings. True, but that was a bear market. The
absence of clear shallow retracements offered clues to the fact it was a
bear. Moreover, the 1982-83 breakout of that 16 year ledge was evidence of a
major change to a bull market psychology from the choppy deep retracements of
the previous 16 years. BTW the 1984 decline-retest was about a 38.2%
retracement from the 1982 low.
Another pragmatic use is that it may give a trader some emotional perspective
to an otherwise mechanical method: if your method goes long the NDX near
3000, and you know "even a bear market bounce could rally to 4100," that
knowledge may help you to pull the trigger. So when everyone on CNBC is
saying "bear market" at NDX 2970, Omega list posters are predicting "1600 is
the next fibo support" and three out of the last four of your system's "buy
signals" have failed, you might be able to keep perspective and play your
method under duress.
After sitting on this report for a week, and reviewing swing charts going
back into the twenties, I've come to the conclusion Fibonacci retracements
are more pervasive during and immediately following parabolic moves.
Moreover, they seem to exist in clusters when measured from a particular
focal point or acceleration point within that parabolic. This explains, to
some extent, why many believe so strongly in Fibonacci in the markets, while
others insist there is no evidence they exist at all. If you analyze "all
the data" in raw form there appears to be very little consistency for
retracements from swing to swing. If you focus on parabolic (and emotional)
moves there are fib retracements everywhere, or so it seems to me. Judge for
yourself, I'd be curious what you think. Here again are the links:
I've outlined most of the intermediate Nasdaq 100 swings since 1994 at:
http://smarttrades.com/SmartCharts1.htm
I've also highlighted many of the swings since the March 2000 high in the
NDX, as well as swings from the 1991 DJI low:
http://smarttrades.com/SmartCharts3.htm
This is by no means meant to be the last word on fib analysis. It is only an
effort to review, in very simplistic terms, the major retracements from 1982.
Bryan Tharp suggested on the list "it may be worthwhile to create a code that
measured the swing high to swing low in bars." Indeed such a tool would
facilitate ratio research as well. It would also be nice to compute the
points in a swing as well as the time. Although I do not have the time,
programing skill, or the inclination to follow up, I'd be curious about a
study that would measure long term price and time swing relationships (ie the
total number of points in upswings compared to the total number of points in
down swings, and the relative time in the swings: that is what Robert Rhea
did in 1936). In the Elliott Wave Principle, Prechter quotes Robert Rhea's
study of 36 years of price swings: "So in 1936 Robert Rhea discovered,
without knowing it, the Fibonacci ratio and its function relating bull phases
to bear phases in both time and amplitude." I'm not aware that anyone has
followed up since.
Bill Wynne
TradeWynne@xxxxxxxxxxxxxxx
WARNING: Trade at your own risk! Commodity trading can be risky and is not
appropriate for everyone!
NO CLAIMS ARE MADE THAT THESE TRADES HAVE BEEN OR ARE LIKELY TO BE PROFITABLE.
NO CLAIMS ARE MADE THAT THESE TRADES HAVE BEEN MADE.
NOTICE: This information is believed to be reliable, but is not guaranteed as
to its accuracy or completeness. This information is subject to change
without notice. Neither the information, nor any opinion expressed herein
constitutes a representation by Smart Trades Inc., or a solicitation for the
purchase or sale of any commodity futures, or options of any kind. Those
using the information herein for trading purposes are responsible for their
own actions and no claim is made that the recommendations will be profitable
or that they will not result in losses. Smart Trades Inc., its officers,
employees, or members of their families may have a position in the markets
contrary to the information contained herein.
|