Sunday, 10 July 2011

THE CURSE OF THE DOMED HOUSE II (Revenge of the Red Dragon)

The Fire Dragon dance, Hong Kong mid-autumn Festival.
But it may be China bulls getting a roasting in the Hong Kong stock market

this autumn...

In April I introduced my reader to the possibility that stock markets across the world were set for a major top, after discovering a venerable technical pattern in the charts going by the peculiar name of Three Peaks and a Domed House

This humble little sequence of numbers lends technical credence to something many shrewd observers have been warning of for months: that the economic tide for emerging markets like Brazil and China may be about to turn. Since the strength of emerging market economies has been such a crucial prop for the world's recovery - and considering the fact that their stock markets have been leading our own for some years now - it suggests too that carnage in the east would inevitably wash up on western shores not long thereafter.

Back in April, I showed how this rare pattern allowed us to predict a market peak which would likely lead to a subsequent major decline.  Sixteen patterns have been identified in the Dow Jones Industrials looking back to the 1920s.  Each and every one resulted in a heavy drop in equity prices.  My own chart diagnosis suggested a top was likely around mid-May.  So I thought we ought to take another look at the picture as it stands today.  Good thing too, because it turns out there's been a twist in the tale.

EEM, July 8th 2011

The main emerging markets investment vehicle fell into late June, then rebounded: 
this throws up the prospect that instead of peaking in April (yellow labelling), it may 
be another two months or so before it tops out (green labelling).

Here's the critical instrument to watch - EEMishares MSCI Emerging Markets ETF.  This replicates an index which is a weighted average of more than 400 leading stocks across 21 emerging market exchanges.  It has carved out the unmistakeble signature of the Three Peaks pattern since mid 2009, culminating in what appeared just a few weeks ago to be the classic Dome configuration peaking at point 23.

EEM then turned right on cue, falling 9% from its top in late April to a trough in the last week of June.  It has however rebounded strongly in the last several days, raising an awkward question: what if it now breaks up through the April highs?  Would that not invalidate the pattern and leave the way open for yet another leg up in this amazing bull market?

Not so fast, grasshopper.


EEM as shown in my mid-April post, which suggested the chance of seeing another leg lower before a final surge to the peak, re-labelling the pattern.  The recent plunge and rally suggests this new labelling may be correct... leaving only two probable outcomes:
a top by autumn, or the pattern is bust.

In my April post, I discussed the potential for an alternate labelling.  In this scenario, stocks would fall back to test the 200-day moving average, forming an extended third leg down for the so-called first storey.  Only then would they begin to rise towards the Dome.  Therefore, while the pattern would remain intact, the timeframe for reaching the ultimate peak would shift forward a couple of months.  

As the charts now show, prices did indeed plunge to the 200-day moving average by late June, but instead of breaking down they have suddenly surged higher amid multiple signs of internal technical strength.  

A push to new highs now on EEM will confirm this new labelling and imply a peak by late summer/autumn as illustrated below.  A failure to hit new highs, followed by a collapse below the March trough would clearly re-affirm the original pattern.

As long as either one of these scenarios plays out, the minimum downside target will remain 35.28 for EEM, a drop from current levels of almost 30%.


As I showed last time, it is often possible to roughly determine in advance when the Three Peaks pattern will top out.  

Schematic of the Three Peaks pattern -
my roughly scribbled annotations show how it's
possible to time the likely top 
By measuring the length of time the market has taken to rise from point 14 to point 15 - ie. from the last point of the base to the initial point of the first storey - and projecting that time forward from the end of the first storey (point 20), it is often possible to forecast with surprising accuracy when the price momentum will have run its course and should most likely reverse.

To check this more thoroughly, I took another look through the 16 examples of this pattern in the Dow Jones Industrials Average over the past 90 years.  I measured the number of calendar days between points 14 and 15 in each case, then the time from point 20 to the exact top of the market at 23.  The results are noteworthy.

Year of pattern                   Days from point 14 - 15                   Days from point 20 - 23
      1923                    30                                48

      1929                    29                                26

      1946                    27                                24

      1948                    39                                42

      1953                    70                                76

      1957                    58                                19

      1961                   134                               148

      1966                    87                                66

      1969                    48                               99

      1973                   132                               88   

      1976                    98                               105

      1981                   117                               138

      1987                    86                                98

      1990                    25                                82

      1998                    89                                33

      2007                   13                                 16

* Came to within a fraction of the final high within 36 days.

I have shaded the obvious outliers in red.  

In the first 'outlier' - where points 14-15 were separated by 30 days and points 20-23 were separated by 48 - there was in fact a peak 36 days after point 20 which came within a gnat's whisker of the ultimate top.  So if we conveniently discount that as an outlier we see that, in 11 of the 16 cases, the number of days in the first time window bore remarkably close relation to the number of days in the second.

Could this be simple coincidence?  Of course.  But remember that the principle which underlays this method is used all the time in technical analysis - ie. the length of a move leading into a consolidation (here, the first storey) often roughly equals the length of the move coming out.

If, for the sake of this argument, we take a leap and assume that a causative correlation does exist and applies in this case, it may only require the application of a couple of further tools to allow us to pinpoint any forthcoming top in EEM.

The maximum margin of error in those 11 cases was +- 25% (usually less).  So if the time from point 14 to 15 was 30 days, you could expect to see a top 30 days after point 20, plus or minus 25%: in other words, 7 days earlier or later.  

You would thus have a two week window in which to take action and, as I illustrated in April, using a simple momentum indicator such as the MACD could then narrow things down even further, making a decision to exit (or sell short) relatively straightforward.

Please note that we're dealing with just sixteen cases here, so drawing firm conclusions and betting the farm on the basis of these stats would be extremely unwise. Prospective short sellers should consider instituting a stop-loss order slightly above the point identified as the probable peak, preventing major damage in the event this pattern fails to follow through. With this caution in mind, let's apply what we've discovered to our current situation.

ishares MSCI Emerging Markets 

Calendar days from point 14 to 15 = 77

77 - 25% = 57
77 + 25% = 96

If the original scenario holds and point 20 was back on March 16th, projecting forward 57 and 96 calendar days gives us the following time window in which to expect a top:

May 11th - June 19th

Of course, we're now deep into July and can see that there was no new high set in that time window - the intraday peak was reached on May 2nd.  This leaves us with two realistic possibilities: either May 2nd was indeed the top and this proves to be one of the outliers, or the new labelling is correct and we have a new high to watch for later this summer.

If it is an outlier, this will be confirmed by a close under the March 16th closing low of 44.60 - a level which would mark the first significant sequence of lower highs & lower lows and confirm a longer-term downtrend.

Until that happens, however, we must work under the assumption of the second scenario, which is that point 20 in the pattern occured not in March but on June 24th.

Calculating a reversal window for EEM

With June 24th as point 20 in the sequence, we project forward 77 days and come to an idealized 'reversal date' of 9th September.  To create a probability window for the most likely turning point we add and subtract 19 days (77 x 25%), giving us the following dates within which to look for a top:

19th August - 28th September 2011

By using sell signals from the standard momentum indicators such as stochastics and MACD within this time period, an investor should then be able to exit or short the market very close to the absolute top.


If EEM rises above its May highs this summer, expect it to peak and begin a huge decline between mid-August and the end of September.  If it continues to climb after that time period, there is no alternate labelling possible and the probability will increase substantially that this is a failed chart pattern.

If we do see a reversal by then, we should anticipate a minimum drop to the level indicated by point 10 in the sequence, amounting in this case to some 30%.* As our historical examples show, however, the decline often extends well below the projected minimum.

*If you are considering making a trade, please bear in mind that there can be no guarantees in this business.  Ideas and recommendations are posted here entirely in good faith, but any risks you choose to take are solely yours.  My lawyer insists I now refer you to the disclaimer on the top right of this page.  Phew, thanks.

What will the catalyst for such a huge move actually be? We can speculate, but major market tops almost never have an obvious cause except in hindsight.  This is self-evidently true, otherwise everybody would try to get out at exactly the same time.  Look back at the headlines in March 2000 and October 2007 and, sad to say, you'll find no red flags waved, no great warning bells rung. Gradually, stealthily, the market just rolled over.

For what it's worth I've posted my own view on the situation in China in these pages many times (you can get the gist here if you're interested) and it seems more than likely that renewed crisis in the Eurozone will provide the overall theme and explosive charge.  

But by the time the outcome has become obvious to everyone, much of the market damage will probably already have been done.  We want to be out of the market, or short, at the moment when risk v. reward is most heavily in our favour.  Hopefully, the ideas presented here will help us achieve that.  

From that moment on, the pattern either confirms - or it doesn't.  If it doesn't, sensible money-management should mean we lose little.  If it confirms, the rewards may be tremendous.  All we can state with confidence is that, of 16 similar patterns identified in the Dow Jones Industrials Average since the 1920s, all 16 resolved in the same way - down. 

You can be sure I'll update you with my own strategy as the story unfolds.