Subject: Eliades on Market Top?
From: Peter Eliades <cyclese@earthlink.net>
Date: 1997/07/13
Message-Id: <33C9BE37.3EA3@earthlink.net>
Newsgroups: misc.invest.technical,misc.invest.futures
Conventional technical analysis suggests there is almost no way the
market could yet be at a top. The number of daily 52 week highs on July
3 hit a staggering 537, the greatest number since 1982, and even though
the number is not anywhere as impressive as the 1982 figure because of
the dramatic increase in the number of issues traded on the NY Exchange,
it is a very high number even on an adjusted basis. Most technicians
will tell you that market tops rarely coincide with peaks in the number
of 52 week highs. They are right. Price highs after peaks in 52 week
highs usually take anywhere from several months to a year or longer.
The McClellan Summation Index, a companion tool to the McClellan
Oscillator, is a marvelous technical indicator and currently receiving
much publicity because the Summation Index is hitting all-time highs as
recently as yesterday, July 11. There is little question the Summation
Index is at a very high level, but it also requires adjustment because
of the greater number of issues traded on the NY Exchange. Our own
adjusted McClellan Summation Index on Friday was high, but nowhere near
a record. It was, in fact, only the 702nd highest reading in the past 57
years. Nevertheless, it would be rare to see the stock market reach a
top coincidentally with the Summation Index at a new recent high,
although it has occurred before.
So much for just a few of the reasons we could well be wrong about a
market top. Now we’ll present you with some of our perhaps not so
conventional technical reasons why Tuesday and Wednesday of this past
week could well have marked a final top in this greatest of all bull
markets.
Let’s start with market sentiment. We have always had a
proper reverence for sentiment indicators, but we also have some real
bones to pick with the conventional use of most sentiment indicators. We
believe sentiment readings should somehow incorporate the underlying
market movement that accompanied the sentiment readings. A put-call
ratio could have different interpretations depending on whether it was
registered during a market rally, decline, or consolidation. We have not
yet been able to do the research we hope to ultimately do in that area
We are fast accumulating data on what we believe will ultimately be one
of the best sentiment indicators available although the history is
limited to the past four years or so. It is the asset data from the
Rydex Group of funds. Rydex is the only fund we know that actually has a
fund inversely indexed to a market index, in this case the S&P 500. It
also has two other funds positively corellated to stock market indices,
the Nova Fund indexed to the S&P 500 with a beta of 1.5 and the OTC Fund
indexed to the NASDAQ 100. There are other funds in the Rydex group such
as a precious metals fund indexed to the XAU index but for market
sentiment purposes, we believe the Ursa, Nova, and OTC funds are almost
ideal. Currently the bullishness evidenced by money managers and
individuals in the Rydex funds is unprecedented. To give you some
reference points, at the late May 1996 market top, a top that led to an
almost 10% decline on the Dow and the S&P 500, the ratio of Nova plus
OTC divided by Ursa i.e bulls divide by bears, was 1.45 on May 20. On
July 2, 1996, at the secondary top before the large decline, the ratio
reached 1.62. Around the February 18, 1997 closing high on the S&P 500,
the highest bullish reading was 3.16 on February 12. Despite the runaway
bull market since December 1994, sentiment ratios never reached 3.0 or
higher until 1997 except for one other day, and even as late as May of
this year, there had never been a ratio higher than 4.0. On Friday, July
11, the ratio closed at a stunning 4.64 and the 10 day moving average
was 4.16. That works out to be 82.3% bullish and 17.7% bearish in
conventional sentiment terms. It is almost inconceivable that bullish
numbers could get much higher. You should be aware that the majority of
managers who use the Rydex funds have a bearish bent to begin with, and
are there primarily to be able to use the Ursa Fund. The fact that they
have been almost totally converted is a very ominous sign for the stock
market. We believe it is marking that long elusive market top and is
doing so with numbers we may not see again in our generation.
Now let’s discuss our RSI 316 indicator. In our last posting
to this newsgroup in mid-June, we suggested that a rare reading of 60+
on a daily Dow chart with a 316 period RSI could well mark the final Dow
high based on prior history of the indicator. We will not go into the
details of that analysis here but it now appears that the Dow will not
reach the 60+ level seen in 1987 prior to the August top. The 1929 top
also did not reach the 60+ level at the top. We should explain to those
of you who did not read the first post in June that there have been only
three prior general time periods in this century when the Dow reached
60+ on the RSI 316 indicator. Two of those three periods, 1928-29 and
1986-87, led to the greatest crashes of this century. We have now
experienced the fourth such period with 60+ readings in early 1996 and
early 1997. All such prior periods this century led to final market tops
within, at most, 529 calendar days from the first reading of 60+ on RSI
316. Friday, July 11, marked calendar day 519 from the first RSI 60+
reading in this time period, also the first such reading since August
1987. It occurred on February 8, 1996. It was interesting to us that in
looking at the NY Composite Index this past week, it virtually reached a
60 reading on July 3 with a reading of 59.97. We should also note that
the highest reading on the DJIA this century was 61.7. On Friday, July
11, the Value Line Composite Index (arithmetic) closed at an RSI 316
reading of 61.76. Unlike shorter, more conventional RSI periods such as
9 and 14, the very long period RSIs used in association with secondary
stock indices often top out coincident with price tops rather than after
momentum divergences. The current reading on the Value Line points to
one of the most overbought readings on a market index seen this century.
That’s not all, folks! There is the Coppock Curve (or
Coppock Guide) so-called Killer Wave. For those of you not acquainted
with the Coppock Curve, it is constructed by adding an 11 month rate of
change to a 14 month rate of change on the S&P (or any other index),
then calculating a 10 month weighted total of that sum. It has been one
of the most spectacularly accurate long term buy indicators ever
devised. Its last buy signal was given in January 1995. A buy signal is
generated when the Coppock Curve turns up from below the zero level.
It’s that simple. The beauty of the Curve is that it tends to move above
and below zero very smoothly with little or no hesitation. The Coppock
Curve was not designed to give sell signals, but an analyst with A.G.
Becker, Don Hahn, discovered long ago that if the Coppock Curve turned
up by more than 10 points within two months of a bottom above the zero
level, it was called a killer wave. Up to now, there had been only three
killer wave signals given over the past 35 years. They came in November
1968, September 1972, and July 1987. A double Killer Wave signal has now
been given in both February and May 1997. The three prior signals led to
three of the largest declines of the past 35 years. The current signal
now calls for an equally ominous decline. The timing of the decline is
not a pinpoint affair, but as you can see from the dates of prior Killer
Wave signals, they preceded the final tops: 1)just days later (December
1968), 2) 3 1/2 months later (January 1972), and 3) one month later
(August 1987). Three instances are not, of course, statistically
significant but the suggestion from those prior three occasions is that
the top should be seen between June and September 1997. The past three
Killer Waves preceded declines averaging 34.4%. We believe that would be
a minimum type decline expected from the next top.
The next technical point is also from our own work. It is a
simplistic but, we believe, elegant indicator. The indicator is derived
by looking at DJIA monthly closes this century. We began this study on a
weekly basis but moved to monthly, but let’s backtrack a moment to give
you our thinking on the matter. Here is a quote from our October 18,
1996 newsletter. “One of the more simple techniques of measuring market
cycles is to calculate how many days (weeks, months, etc.) the market
index you are following was up over a specified period of time. We have
done that often on a daily chart with varying degrees of success. We
experimented over the past few days with weekly data........We looked at
a Dow weekly price chart going back almost 60 years. The longest
uninterrupted advance seemed to be the period from October 1962 to
February 1966. It consisted of 172 weeks. Over that 172 week period, the
Dow Industrials were up for the week a total of 113 weeks...since 1943,
there have only been 45 weeks with readings of 111 or higher.[In the
last 40 years, the first] readings of 111 or higher occurred in the
original test period of 1966. It began on the week ending January 14 and
continued the next two weeks. On February 4, 1966, the reading reached
112. At the market top week, the week ending February 11, the reading
reached the record 113. The next two weeks saw readings of 112 and 111
respectively and then the string ended. As the string was ending, the
market was registering a high that would not be significantly surpassed
for over 16 years. Perhaps of equal significance is the fact that the
magical and rare reading of 111 or higher would not be reached again for
over 20 years. The next occurrence was the week ending August 21, 1987.
The 1987 pre-crash high on the Dow was registered two market days later
on August 25.” That’s the thinking behind the indicator. Moving now to
monthly studies, the two monthly time periods we have found to be
significant in the past have been 40 months and 100 months. Monday, June
30, 1997 marked the end of another month.It was an up month. Only one
other time this century has the DJIA been up more than 68 months out of
the past 100 months. That occurred in December 1961. Just short of 6
months later, the Dow was down over 28% from the December 1961 close.
There is also another first that occurred on June 30. The century old
records of 30 months up out of 40 and 69 months up out of 100 was
achieved in the same month. It is a record extreme based on these two
extreme readings being reached in the same month. Let us put these
current readings in perspective for you. The only times this century the
market has been up more than 67 out of 100 months prior to the past year
were October 1961 through February 1962 (all readings of 68 except for
December which was 69) and January, February, and April 1966.
Similarly, the only times this century the market has been up more than
29 out of 40 months prior to the past year were February 1937 and April
through August 1961. In 1937, just in case you don’t remember, the Dow
topped out 9 market days after the February close and proceeded to
decline 16.5% intra-day in the next 3 1/2 months, 40.8% in the next 7
1/2 months, and 50.2% in the next year.
We have come this far- let’s go just a bit further. Let’s
just make the mad assumption that, against all odds and certainly
against all sentiment, the market is at or fast approaching a major
major top. We went back and examined the most important market tops of
the century making sure to include the tops that were signalled by the
Coppock Killer Wave formation, namely December 1968, January 1973, and
August 1987. We were curious to see if there was a consistency in the
time of the final leg up that led to the major tops. We discovered there
was indeed an impressive consistency. We used the major tops of 1929,
1937, 1968, 1973, and 1987 to measure the time span of the final rallies
leading to the tops. The results were: 1929--78 trading days,
1937--63 trading days, 1968--65 trading days, 1973--57 trading days,
1987--67 trading days. The average of the 5 resolutions is 66 trading
days and the extremes were only only 9 days shorter (1973) and 12 days
longer than average (1929). So far the DJIA intra-day top occurred on
July 9, 1997. One look at a chart makes it obvious the current rally
began from the low of April 14. Friday, July 11 marked the 62nd trading
day from the bottom. We are in the heart of the average period to a
final top using trading days. Using calendar days and throwing out 1968
because there was no trading on Wednesdays during the period in
question, we arrive at an average of 89.25 calendar days as the length
of the final rally leading to a major top. Saturday, July 12, was the
90th calendar day from the April 14 intra-day low.
There is much more of interest, but we have probably been
presumptuous already, posting this long a message. The bottom line of
the message is there is some decent evidence we have seen a top already.
Imagine the irony if we don’t see a close above 8,000. It is surely
taken for granted now. But the level is far less important than the
timing of the overall picture. The above indicators are screaming that
some kind of a serious top is imminent and it could well lead to one
hell of a bear market. As to a crash, it depends on your definition
thereof. I believe the chances are real that 7-8 weeks after the final
top, a secondary top will be seen that will lead to crash-like behavior.
But, you say, crashes don’t happen when people are predicting and
debating their possibilities. HA! Go read the September 5, 1929 edition
of the NY Times and see if a crash was not being debated just two days
after the market top and just weeks before one actually occurred.
Peter Eliades
Editor-Publisher Stockmarket Cycles
Subject: Re: Eliades on Market Top? From: YCGN15B@prodigy.com (Arch Crawford) Date: 1997/07/21 Message-Id: <5qubvp$2hgo$1@newsxfs02-int.news.prodigy.com> Newsgroups: misc.invest.technical Peter, Great Post!! You're absolutely right about how difficult it is to put up decent posts when so many creeps throw mud for no other reason than to see their own typing!! There area few top analysts lurking around these boards and I'm sure most of them don't want their good names besmirched by the careless dogs. You have a darn good record (one of the best) ranked by Third parties, and I am one other who believes that we may have just seen THE TOP or at least an important TOP. The drop during the next few days will give us some idea of the parameters coming up. If the markets can get moderately Oversold and start another good bounce, we might still be in the ballgame. Advances-Declines, New Highs-New Lows and 5 and 10-day TRIN are acting pretty well whereas 25-day Put/Call (CBOE Totals) are the lowest since summer of 1990 (the last time we had a decline of more than 10%)! Momentum of every sort stinks here. Most things topped in MAY for momentum's sake! and have made one or two more lower highs since then. That suggests an immediate down for NOW! If New Lows (NYSE) expand above, 40, we got problems. If they stay over 40 for 3 days running, it may be more serious. If New Lows actually rise above New Highs for 3 days in a row, IT'S A BEAR MARKET. We'll see. Best of Luck & Light. Arch Crawford