The chart pattern we saw before the last market sell off in technical analysis terms was that of a broadening top; further descriptive adjectives included a "bearish rising 'broadening' wedge pattern". This mouth full calls for a fuller explanation!


I discuss in my (Essential Technical Analysis) book about some pioneering research done some years back by Dr. Andrew Lo, an MIT professor and researcher in the area of the financial markets. He and a group of his grad students made a comprehensive study of whether technical price "pattern" had future predictive value. Assuming the characteristics of a chart pattern could be defined objectively enough for a computer search of historical stock market data (for individual stocks). If so, the resulting outcomes could be studied as to whether the trend subsequently went in the direction predicted by that pattern according to mainstream or 'conventional' technical analysis.

Of the chart patterns Dr. Lo studied, he found that there were 5 such technical patterns that yielded "statistically significant test results"; i.e., they had a level of predictive value that was greater than any random chance occurrence.

I have written about some of these patterns having 'proven' predictive outcomes in prior Trader’s Corner articles. The 5 . They are:

1.) The Head & Shoulders top or bottom pattern - 7/9/10

#'s 2, 3, and 4 haven't had specific prior articles done by me yet.

2.) Double tops

3.) Rectangle Tops

4.) Rectangle Bottoms

The 5th pattern of 'proven' predictive value in the historical data analyzed for stocks is:

5.) The broadening bottom

I'll write about both broadening tops and bottoms, although the MIT group couldn't prove the TOP pattern was as predictive of a future outcome as the broadening BOTTOM pattern tends to have.

There are as I said, both what's termed broadening bottoms AND broadening tops. The broadening bottom formation is the one that Dr. Lo's MIT group found to have a predictable outcome; i.e., the (broadening)'bottom' pattern predicts an upcoming trend reversal from down to UP often enough to be proven statistically.

Before moving on, my first chart is of the recent 'broadening' TOP pattern:


When there's a downside price objective reached that is implied or 'measured' by a particular chart pattern, AND the major stock indexes gets to an oversold extreme AND bearishness gets extreme also, we have one of the best bottom 'signals' going:


The broadening formation is often a reversal pattern because, unlike the compression of price swings implied by a triangle formation, occurring when buyers and sellers get more and more finely matched, the broadening formation is 'decompression' in a way. When price volatility increases as is the case when you have a series of higher highs and lower lows, it is because buyers and sellers of a stock get less certain as to what the correct valuation (price) of a stock or index should be. This dynamic often accompanies a reversal in the direction of the trend.

One peculiarity about a broadening formation: it is the SAME pattern in either instance. It can be labeled an expected broadening BOTTOM when the trend that preceded the pattern is DOWN and the and a possible broadening TOP when the preceding trend was UP as was seen above in the S&P 500 chart. The MIT study implied that the broadening pattern, where the prior trend was DOWN (before a broadening formation was traced out), as useful in predicting an upside reversal ahead. A pattern you might want to trade off from. Here of course a picture from my historical chart bank is worth a 1000 more words:

Now, the megaphone shape looks the SAME as the pattern of the recent intermediate TOP. The difference is that the trend that PRECEDED the recent broadening (top) pattern, which is also a formation described as a 'wedge') was UP. This suggested a possible downside reversal based on the broadening price swings, once the lower trendline was pierced. Moreover, the objective implied by a bearish rising wedge pattern is for a decline back to where the wedge first formed.

The 'measuring' objective for the broadening bottom as seen immediately above is different. The widest part between the two trendlines highlighting the broadening bottom, is added to the 'breakout' point to yield a 'minimum' upside objective. And in this Nasdaq stock example, the minimum objective was considerably surpassed.

The broadening bottom is where the price trend that preceded the pattern is downward and a sideways series of price swings develop that form a series of higher upswing highs and lower downswing lows. Meaning that there is one upward and one downward sloping trendline. There should be at least 2-3 prices swings that result in the 2-3 lows or more and 2-3 highs (or more) that allow drawing of the trendlines. The subsequent outline to this pattern is also triangular in shape but that shape takes the shape of a megaphone.

A breakout from the broadening bottom exists when prices advance about the highest high in the formation; i.e., 'within' the megaphone. The measuring implication for the height of the next significant move or leg is the same regardless of whether the breakout direction is up or down. The measuring implication for a broadening bottom is for a minimum objective equal to the height of the formation at its widest point added to the breakout point or price, which is the price point where there's a penetration of the upper trendline.

The price objective rule of thumb for broadening formations in general is not as reliable as for the Head & Shoulders or triangle patterns. If the breakout is to the upside, when any subsequent gain after the breakout tends to run around 20-30%. At least some studies suggest this is an 'average' gain after such a breakout.

While the lower line of the broadening bottom is usually formed by LOWER relative lows, I was struck by the similarity of the pattern that was traced out back when in the XAU, the Philly Gold & Silver Index, as seen in my next chart example. Of course the long-term breakout was strictly to the UPSIDE; with XAU trading at 216 today, those were the days when you could buy it at 80!

Back when I wrote about the XAU chart above, I said that "while there is a need to clear its weekly down trendline around 80 and to exceed the prior weekly closing high at 86, it appears to me that XAU is going to do this and 86 will be the 'breakout' point." My prediction and commentary was based on the broadening formation. As a sideways move was possibly forming a bottom, the next move was likely to be UP.

As I said there is little to differentiate the actual broadening formation in the top or bottom varieties from one another EXCEPT for the direction of the preceding trend. The preceding trend, when down, implies the formation of a broadening bottom. When the broadening formation forms after a preceding advance, particularly a prolonged run up, will often turn out to be a broadening top.

If not readily apparent from the chart pattern being traced out as to whether the preceding trend was up or down, we can define the trend direction by whether prices are trading below a 200-day moving average.

Another example of a broadening TOP

My next chart has the same upward sloping trendlines traced out by a series of HIGHER highs from rallies, coupled with a downward sloping trendline connecting lower downswing lows. This results in the same megaphone shape as the bottom formation. However, the trend that preceded this formation was up, so at the time I assumed that the resolution of the pattern was going to be a breakdown below the lower trendline. One example (JNPR), which is seen next, was taken from my technical analysis book. (The other at the time I saved the INTC chart was an example that I went looking for among some of the big tech bellwether stocks.)

The breakout direction is the important thing. We have to be prepared for a pattern 'failure' also. A pattern failure is when the outcome reverses from the expected or usual. A pattern failure for the chart above would have been if the chart formation turned out to have led to the same result as a continuation pattern and the breakout was to the upside, in the direction of the prior trend. In the JNPR chart show above, I've used the SAME 'measuring' technique (for assessing the minimum point move) as used in the examples of broadening bottoms.

NOTE: This was different than when the broadening top also formed the prototypical wedge pattern which starts with a narrow starting point for the two trendlines that widen out. In that example I looked for a move BACK to the initial apex or starting point of the triangular shaped wedge.

The 'breakdown' point in the broadening top is a move to below the lowermost trendline, such as seen with the Intel chart (in 2000) shown below:

There is more price volatility than usual being shown in a broadening formation as it's not the most common thing to see big price swings in BOTH directions; i.e., big moves that result in higher highs and big price swings that result in lower lows. The direction of a next price leg after such increased volatility is never going to be completely predictable.

What is needed is to watch carefully for a breakout at either trendline; the price objective on a breakout either way is the same as suggested by a measurement of the height of the formation added to or subtracted from the breakout point depending on the direction of that thrust.

The broadening pattern is not as common as other types of reversal patterns such as double tops or the rectangle bottom or top. Even the Head & Shoulder's pattern probably is more common, but I haven't tried to quantify all that. While the broadening formation is not all that common, it is distinctive and typically forms over a long enough period that you have time to consider the pattern's reversal prospects.

And, as the work done by Dr. Lo would suggest, if the trend preceding the broadening formation was strongly UP, it's more likely to signal a significant top, such as seen above with INTC. The stock ended up in the $13 area in the year following the end period shown on the chart above.