Some Stocks Foretold Their Big Declines

Some of this year’s biggest stock market losers gave warnings that could have saved investors at least some of
their losses.

We’re talking about the arcane science of technical analysis, the study of chart patterns to predict a stock
or index’s next move. In some of the cases, patterns served as a warning that the stock market was topping. In
other cases, in which the stock in question later dropped on news, technical analysis served as a warning that
something was amiss. A few insiders and followers knew, and they were selling. That is technical analysis’s
greatest benefit: as an equalizer for those who aren’t in the know.

We’ll break down the most common technical patterns into two categories: reversal patterns, which mean a stock
or index’s trend is reversing; and consolidation or continuation patterns, which are a breather before a stock
or index continues in the initial direction before the consolidation occurred.

Reversal Patterns

The Head and Shoulders is probably the most common reversal pattern. It is essentially two higher highs (the
left shoulder and head), followed by a lower high (the right shoulder) that shows that the rally is running
out of steam. The pattern is not considered complete until the neckline is broken by 3% on a closing basis (a
good rule of thumb for all pattern breakouts). America Online and JDS Uniphase were two stocks that formed a
head and shoulders top (note that AOL also later formed a symmetrical triangle, which we will get to under
consolidation patterns).

An inverse head-and-shoulders pattern, the reverse of the topping pattern, can signal a bottom (see this
United Technologies chart).

The head-and-shoulders pattern predicts a move equal in size to the pattern (measured from neckline to head)
from the neckline break; in the case of AOL, this meant a 30-point move from 70, or to $40 a share.

The Wedge is another reversal pattern. It is essentially a three-week to three-month rally (or sell-off) with
converging boundary lines, signaling that the rally or sell-off will run out of room eventually. It is common
for bear market rallies to be rising wedges. The Nasdaq 100 had two rising wedge rallies within a larger
rising wedge, signaling to investors that the rally wasn’t going to last.

The predicted move out of a wedge is to the point where the wedge began. Also, rising wedges usually precede
rapid moves, while falling wedges predict slow ones. The Nasdaq’s recent bottom was a falling wedge,
predicting the index’s subsequent slow recovery.

And the most maddening falling wedge of all belonged to Amazon.com, which broke out of a falling wedge at 35
and retraced all the way to the apex of the pattern at 19 before recovering (also note the complex
head-and-shoulders top in the stock).

Two other reversal patterns are worth mentioning. They are normally rare, but appeared often enough this year
to signal to technical analysts that something was amiss. The Broadening pattern, a series of three higher
highs and two lower lows, appeared in Nokia and Juniper Networks.

Such patterns can sometimes have a flat top or bottom, and in the case of Nokia, can sometimes go on to make a
fourth high. The pattern has a particularly fascinating history; it occurred so frequently in leading stocks
leading up to the 1929 crash that the appearance of broadening patterns struck fear into technical analysts
for decades afterw

ards. While the pattern is similar to other chart patterns in the sense that the size of the
subsequent move is dictated by the size of the pattern, the circumstances that create broadening patterns are
somewhat worrisome, as they require markets that are out of control, whipped about by the slightest news or
rumor. Such conditions usually only occur at tops.

The second rare pattern is a kindred pattern to the broadening formation, called the Diamond pattern, which is
a broadening pattern that narrows into a symmetrical triangle before making its move. The Nasdaq 100’s top may
have been a small diamond pattern.

The Dow has been tracing out a massive diamond pattern for a year and a half, and has broken it both to the
upside and the downside. Time will tell if the pattern is anything more than just a large trading range.

Consolidation or Continuation Patterns

Most patterns fall under the heading of consolidation or continuation patterns, meaning they are just a
breather between further moves in the same direction. However, these patterns can sometimes appear at tops or
bottoms and signal a reversal.

We’ll begin with the Rectangle, which is a pattern with both a flat top and a flat bottom, hence its name. The
predicted move out of the pattern is, once again, the size of the pattern from the breakout point. The most
perfect example this year probably belonged to Ariba, which surprisingly formed a rectangle at both a bottom
and a top.

Triangles are another common pattern. Like the rectangle, the Symmetrical Triangle is a neutral pattern, not
giving away direction until the pattern is broken. However, it too is usually a continuation pattern. The
symmetrical triangle is probably the least reliable of major patterns, giving false breakouts to either side.
Commerce One was another stock with a failed symmetrical triangle breakout that has so far held the lower
boundary line of the pattern (see the red line extending that lower line).

The predicted move out of a symmetrical triangle is the widest measurement of the pattern from the breakout
point, but notice that Commerce One has never hit its 88-90 price target, based on the 36-38 point pattern
from the 52-53 initial breakout point.

Right-angle Triangles are more reliable. They can be either Ascending Triangles (bullish), with a flat top, or
Descending Triangles (bearish), with a flat bottom. Priceline.com gave investors a chance to get out at 32
when it broke a descending triangle. The 18-point size of the pattern, from 32 to 50, predicted a move to 14.
Notice that patterns with a horizontal boundary line tend to break on the side of flat resistance, which is
more easily exhausted than a rising or falling resistance line.

Finally, Flags and Pennants are small patterns (three weeks maximum) that tend to appear in the middle of
rapid moves. They are sharp rallies against the prevailing trend, often fooling investors into thinking the
trend has changed. The best example of a bear flag was in Lucent Technologies, which fooled investors into
thinking the bottom had arrived in the high 20s and low 30s.

The difference between the two patterns is that a flag has parallel boundary lines and a pennant has
converging boundary lines. The Dow formed a bear pennant right before its plunge to 9650 in October.

Flags and pennants also form in upside moves, again in the form of sharp rallies against prevailing uptrends.
Notice in the Ariba chart above that the stock formed

a bull flag exactly at the mid-point between its bottom
and top rectangles in mid-July. The size of a move out of a flag or pennant is predicted by the size of the
“pole” that preceded its formation, the distance from the previous consolidation.

Sometimes stocks will combine patterns, giving investors multiple warnings of an impending move. Yahoo gave
investors three chances to get out above 100: at a broadening pattern in July, a rectangle in August (which
shouldn’t have broken to the downside but did), and a bear flag in September.

A few important points to keep in mind in tracking chart patterns. Volume tends to decline within patterns as
they develop (with the exception of broadening patterns), and then pick up after the pattern is broken. Don’t
trust a pattern that doesn’t develop on declining volume. Also, always begin tracing a pattern from a high or
low point; these are reaction moves within or against a trend. And always wait for a close that is 3% beyond
the boundary of the pattern before considering the pattern broken (2% in the case of an index). And finally,
the entire predicted move of a pattern is completed only about two-thirds of the time. These patterns
represent significant pent-up energy, so there’s no substitute for staying alert. Also, be careful using chart
patterns on very low-priced stocks, as many Internet stocks have become; many don’t have enough room to form
reliable patterns. Very low-priced stocks can, however, form dormant or rounding bottoms, and the bottom may
be considered in place when the stock turns up or breaks out of its trading range.

So what looks interesting now? Juniper should be avoided below 180, since the broadening top gives it downside
potential to 120. Amazon should have put in a bottom with its falling wedge, but probably won’t climb in a
hurry. Ariba may be forming a head and shoulders top.

eBay may be forming a head and shoulders bottom.

And BroadVision may be forming a bear flag, with a possible retest of 17 in store.

Again, wait for patterns to be broken by 3% on a closing basis before considering them complete. Until then,
anything can happen.

For more on technical analysis, visit internet.com’s twice-daily market analysis, either at http://www.internetstockreport.com/midday/ and http://www.internetstockreport.com/close/, or at http://www.afterhourstrading.com/column.html.

News Around the Web