Head & Shoulders: a Close-up

A Cinderella among the Nobles

Trading methods are not only different and, most of the time, personalized ways of looking at the market or the developments that take place on it. In most cases, they turn into a sort of living entities that have their own histories and lives. Unlike some of the famous methods, shrouded in myths and invented by some celebrated, charismatic, and enigmatic personalities to whom they owe their popularity, the well-known Head & Shoulders method was neither concocted by any controversial trading guru nor is it promoted by the less lucky in trading offspring of the latter, who opted for the marketing of their great grandfather's dubious legacy having considered the other option of setting up a diners for the local townies round the corner. And this is what gives another reason to look more closely at why the technique is so popular with so many traders.

How It Works, or You Still Need a Head on Those Shoulders

Basically, the head & shoulders technique is rather simple as compared to most other methods. You don't need to painstakingly select any inputs to train your software, measure the planetary influence of Saturn or count the number of successful UFO landings on it. It's all pretty much down to earth, and, probably, this is one the method's major merits: you can clearly see the advantages and disadvantages of what you entrust your hard-earned bucks to.

The Head & Shoulders pattern is comprised of four equally important parts: the left shoulder, the head, the right shoulder and the neckline. The starting point for the formation of the figure on the whole, and the left shoulder in particular, becomes a sharp rise in the trend-line followed by a fall, a downward advance, and the largest volume in the pattern. The peak volume is accounted for by the growing number of buyers, joining in as the upward move continues.

Now we have the left shoulder. The head is formed by an upward and shorter advance toward the highest peak in the formation, a fall, and a peak volume for this part of the figure. The right shoulder starts with a high-volume sell-off and an upward trend. This upward trend breaks lower than the head peak to cause the volume to fall to the lowest point in the pattern as the public starts awakening to the reversal in the making. The main element in the pattern is the neckline, drawn between the lowermost points of the shoulders. The pattern is complete when the neckline, being a support line, is pierced by the right shoulder. The "lifespan" of the pattern is normally from several weeks to several months, and it's unlikely to be suitable for use over the longer haul. One the face of it, getting the hang of the whole thing is like a cinch: definitely no nerdy background is a requirement. But as a thinking trader you are aware that everything can be so simple only in that software with the one magic button, fools spend their whole lives looking for. Therefore, what are the pitfalls and how to avoid them?

Identifying the Familiar Silhouette

A really big problem about the Head & Shoulders method is that if you really want that to happen, you'll see signs of the pattern occurring everywhere. A sign of a true occurrence of the pattern is a very sharp rise in the price and a distinctly large volume this rise is accompanied with.

Also, one should always bear it in mind that for this specific pattern volume is of greatest importance. A very low volume on the right shoulder means that a reversal is happening. A true reversal is also accompanied by the price falling far enough below the neck-line. The market tries to overcome the previous high, fails to do so, and, thus, experiences the deepest drop for the entire pattern. Most probably, it is largely this simple, "trustworthy", and entirely logical explanation that the technique owes the trust many traders put in it to. Please note, that if the price trend-line is simply bobbing along the neckline, you may well expect an upward move.

Another helpful sign is that the peaks of the two shoulders should take roughly the same amount of time to develop and be roughly of the same height.

Overall, one can say that the possibility of error while recognizing the true occurrences of the pattern, especially with novice or less experienced traders, is pretty high. The chances can be approximately estimated as 50/50. That is why, to reduce the related risks, it would be expedient to use this method as an "auxiliary" method along with the other methods you use until you feel more confident about your pattern recognition ability. In any case, another result that coincides with the result of your forecasting results will not do any harm.

Can Moving Averages Be Used to Forecast Price Direction?

In a single word: No.

I often hear of or read the comment that analysts use moving averages to predict price direction and even the next close. I started my career as a technical analyst way back when I was working in Hong Kong and my treasurer sent me on a technical analysis seminar where I learned about golden crosses and dead crosses. In particular, stated the presenter, use a 20 days and 60 average to represent a 1 month and 3 month average.

"Great!" thought I, "this is going to make my life easier. I can't wait to get back into the dealing room to try this out."

In those days the fashion was to draw your own charts to get a better "feel" for the market and thus I used to draw the averages on large charts stuck to the wall.

What a waste of time…

Every time the 20 days average crossed above the 60 days average price reversed lower.

OK. Then throw average out of the window… Since that time I have never used averages in my analysis again.

That was around 1989-1990 just as real time technical analysis systems started appearing in trading rooms. Over the 1990's I became interested in building systems and it was around the mid 90's that I kept hearing this statement about how moving averages predict price.

Well, using a programmable technical analysis software I set up the software to write a report for me that summarized the close to close profit or loss generated from using the underlying moving average direction. Hence, when the moving average was rising I would then measure the profit or loss generated from close to close over the next day, and vice versa when the average was declining. Which average did I use? For the sake of simplicity I used the market favorite, the 20 period moving average used by the fund management industry.

The following are the results for six currency pairs showing the average profit & loss and the percentage of winning trades:

	USDJPY	EURUSD	USDCHF	GBPUSD	GBPJPY	AUDUSD
% Winning Trades	50.06	49.55	48.42	48.93	48.64	49.13
Average Profit	0.62	0.0053	0.0079	0.0066	1.01	0.0035
Average Loss	0.60	0.0050	0.0072	0.0062	1.06	0.0033

Note that in all cases the average profit is not any higher than the bid-ask spread and mostly lower. This doesn't really give me any confidence in using a moving average to predict the next day's close to close movement.

The moral of the story? Don't expect too much from your indicators. There is nothing magic about any of them in providing anything but circumstantial supporting evidence of a trade. All indicators, whatever the creator's claim, are lagging indicators since they are based off historic prices and therefore reflect what has been and not what will be. They can provide you with information about the general characteristic of price action at that time but the real information you require is all in price and it is in the understanding of price and trading mentality that will help you pick better trades.

Andrews's Pitchfork: Thinking Approach

Trading approaches are much like people: some aren't worth a dime but have a lot going for them: their genius creator's legendary bios, intricate, if bizarre, theories behind them, commercial websites, full of promotional puffery, and hordes of steadfast followers, who keep the former afloat. Others are rather modest but talented and really able to help trade profitably. No doubt, one of such talentedly developed methods was invented by Dr. Alan Andrews in the beginning of the 1960-s to become one of the simplest, most efficient and reliable technical analysis techniques. So what makes Andrew's Pitchfork a valuable analysis technical tool and what are the benefits and risks one should be aware of while applying this method?

As we have already mentioned, the method is quite simple. It shows the areas of resistance and support using three lines: a median, starting from the most recent contract's low or high, and two more lines, being the "tines" of the pitchfork. The upper "tine" originates from the peak of the first upward move. The lower "tine" originates from the low of the downward move that follows the first upward move. The pattern looks like a triangle. The idea is to buy at lows and sell at highs. Simple, isn't it?

You may ask: "And what's the big deal? Does this give a better guarantee than the other commonly used methods?" The answer is "No". Andrews's Pitchfork is about as precise as most other methods (our estimate would be around 60%). But the precision level, ensured by Andrews's Pitchfork, is, actually, not the reason for which the method is valued so highly or should be used. What makes the method a great trading tool is the inestimable possibility of being able to analyze one part of the pattern using the other parts. In other words, you can analyze the different trends that make up the pattern as parts of the same figure, for example, you can most effectively and very easily forecast the size of the second bullish trend based on the length of the first bullish trend.

We will agree with those who will say that Andrew's Pitchfork should not probably be used as the bulk, determining method of your trading combination. However, its importance as that of an extremely valuable "tactical" method should not be overlooked.