As I mentioned in a recent Charts in Play, flag formations are one of my favorite chart patterns to trade. These triangles generally represent continuation patterns or pauses in a major trend.

Therefore, when you are looking to trade or invest based on flag patterns, you have already made a decision about the direction of the major trend.

When a flag is formed as an interruption in the major uptrend, it is often referred to as a "bull flag." The formation of a flag during a downtrend is, therefore, known as a "bear flag."

In addition to simplifying the major trend analysis, one can also use Fibonacci analysis on the flag formation to give you an objective entry-level protective stop, as well as an initial profit target. This allows you to clearly assess the risk and reward of each trade.

The flags can take a wide range of shapes, as some are quite simple and others are quite complex. Often in a major trend, you will see a number of flag formations that can occur over a several years within a major up- or downtrend.

Click to Enlarge

During the 2007-2009 bear market, there were quite a few "bear flags" evident on the daily chart of the Spyder Trust (SPY).

In 2008, there was a classic bear-market rally from the March 2008 lows that ended in May. The SPY had topped out on October 10, 2007 at $157.52 (point 1), and the first wave of selling took SPY to a low of $140.66 in late November.

The ensuing rebound was quite sharp—this is typical in bear markets—as SPY gained 8.9% in just 11 days before the decline resumed. The next wave of selling was much more severe; the SPY dropped from $152.89 to a low of $126 (point 2) in late January 2008.

The 17% decline pushed sentiment to extremes. By February, most newsletter writers were convinced that a new bear market was underway. The SPY retested the January lows in March (point 3), setting the stage for a bear-market rally.

Flags are often formed in reaction to extremes in sentiment. By the time the bear flag was completed in May, many of the February bears had turned bullish on the market.

By the middle of April, SPY had formed an upward trading channel (lines b and c). With the peak in May, a more pronounced flag formation was evident. The Fibonacci retracement resistance levels calculated from the October high (point 1) and the January low (point 3) allowed one to define the levels where the rally could be expected to fail.

The 38.2% retracement resistance at $137.60 was tested in April, but SPY was unable to close above it. Two weeks later, SPY did close higher, setting the stage for a rally to the 50% retracement resistance at $141.22. (For more, see Fibonacci Analysis: Master the Basics.)

Typically, bear-market rallies will fail near or slightly above the 50% retracement resistance. If the 61.8% retracement resistance at $144.81 was exceeded on a closing basis, it would suggest that one’s analysis of the major trend is wrong.

Once the 50% retracement resistance level was first reached on May 2, the focus was clearly on the short side as a broader flag formation (lines a and c) was then evident.

Now, there are two main ways to trade these flag formations. One is to identify a selling zone, using a stop above the 61.8% resistance levels. Or, alternatively, one can wait until the flag formation is completed.

The first setback from the test of the 50% retracement resistance lasted five days and held the uptrend (line c). As SPY again turned higher, the target selling zone was either the prior high at $142.37 or the midpoint between the two resistance levels, which was $143.01.

In determining your selling level, one must of course compare it to your stop level, the potential profit target, and then decide the amount of risk you are willing to accept. Selling at the prior high of $142.37 would have made it quite likely that your trade would be filled—or, alternatively, you could sell closer to the halfway point between the two retracement levels.

Since markets often stop at round numbers a bit under $143.01, a point like $142.88 would have made sense. SPY formed a doji on May 19, 2008, as it made a high at $144.30.


Click to Enlarge

As for the stop, how much room above the 61.8% level should you give your position? I will often look for a chart point above the 61.8% resistance level.

Alternatively, one can use a stop 0.5% above the 61.8% resistance level, which would put the stop at $144.81 + ($144.81*0.005) = $145.54. This would make the risk at $145.54 - $142.88 = $2.65. Instead, if one sold at $142.37, the risk would have been $3.17.

Now, what are the likely downside targets? The most conservative target would be a test of the low at $126, but I generally look for a move to the 127.2% retracement level, calculated using the rally from point 2 to point 4. This level was $120.88.

Using either target level, the risk-reward was clearly favorable. The risk was at 2.65 or 3.17, and the reward if the lows were tested would be 143.88 -126 = 17.88. So the trade would be risking 2.65 or 3.17 for the potential to make 17.88. Of course, looking at the 127.2% target, the reward was significantly greater.

On the other hand, you could wait for the close below the uptrend (line c) that occurred May 20. The opening on the following day was $139.43, and a stop 0.5% above the prior high would have been roughly $145.12.

This would have meant a risk of $5.69. But with the potential to make $17.88, the risk reward was still favorable.

The 127.2% target was exceeded on July 14, and then SPY began a six-week rebound.

Click to Enlarge

Now let’s look at a bull flag example. This chart of the SPDR Gold Trust (GLD) covers the action in GLD from August through December 2007. GLD had just rallied 31.7%, from the low of $63.47 in August to a high of $83.64 in November.

In just seven days, GLD dropped to a low of $76.11 (point 3) and reached the 38.2% support of the rally from point 1 to point 2. Most traders, after missing a sharp rally, are too eager to get back in and buy on the first bounce.

The problem with this is, as often happens in a flag formation, the second decline will result in lower lows. Therefore, stops just under the first low will get hit, leaving traders on the sidelines at a level that in hindsight will turn out to be an excellent entry point.

In this example, GLD managed to rally back above the $82 level and filled the prior gap before the rally stalled. On the following decline, the low in GLD was at $76.98, which was well above the prior low.

After this low, a triangle formation was clearly evident (lines a and b). In my experience with this type of triangle, a break below the lows at point 3 would be quite unlikely.

Therefore, the strategy was to either buy on a setback towards the trend line support or the closing support in the $78.20 area. The initial upside target was the 127.2% target at $85.75, which was calculated using the decline from point 2 to point 3.

GLD traded as low as $77.80 in mid-December, before it again turned higher. A tight stop at 0.5% under the point 3 low at $75.72 could have been used; or, alternatively, a much wider stop under the 50% support at $73.50.

In the first example, the risk would be $78.20 - $75.72 = $2.48, while the risk on the wider stop would have been $4.70. The potential reward would have been $85.75 - $78.20 = $7.55. This would have made the risk-reward good with the tight stop but less than 2:1 using the wider stop.

Of course, this is based on the first target of $85.75—but the explosive nature of the prior rally made higher targets very likely. The 127.2% and 161.8% targets were hit within a month of the triangle’s completion.


Click to Enlarge

The historic bear market that followed the top in the Japanese Nikkei-225 has lasted a couple of decades, and has included many bear flag formations.

I included it to demonstrate how they form in all markets, over any time periods. (I have some chart books from the 1970s and earlier that have great examples of flag formations. In the original Edwards and Magee, there are countless examples from the 1930s and 1940s.)

From the December 1989 high of 38,957, the Nikkei finally bottomed in April 1990 with a low of 27,251. After a 10,000-point slide, the rebound was quite sharp, and the 38.2% resistance level at 31,722 was exceeded by early May.

The Nikkei traded above the 50% retracement resistance at 33,104 several times over a two-week span in late May and early June. The 61.8% resistance was at 34,485.

The short-term support (line b) was finally broken, as the Nikkei declined to test the 38.2% support (line 4). The following rally allowed one to define a longer-term flag formation (lines a and c). The three-week rally took the Nikkei back to the 50% retracement resistance.

How might you have traded this? Of course, the impetus for most was looking for a spot to get out of longs.

For those who were looking to short, selling at the 50% retracement resistance with a stop 0.5% above the 61.8% resistance level was a reasonable strategy after a 10,000-point decline. (Occasionally you will see a flag that exceeds the 61.8% resistance, but it will not exceed the 78.6% resistance before it is completed.)

Thus, a sell at 33,104 with a stop at (34,485 + (34,485*.005) = 34,658 would have had a risk of 1,554 points. If the Nikkei just tested the lows, the potential reward would have been 33,104 - 27,551 = 5,853 points, for a risk-reward of 3.76:1. The 127.2% target was at 25,658, which made the risk-reward look considerably better.

Of course, one could have waited for the flag formation to be completed, once support (line c) was broken on July 23, 1990. The Nikkei opened the following day at 31,834, so a stop 0.5% above the prior high of 33,186 (33,352) would be standard. This would have been a risk of 1,518 points, with a minimum potential reward of 5,853. Both the 127.2% and the 161.8% targets were met in late August.

Next: Bull Flags In Amazon and Apple?


Click to Enlarge (AMZN) made a low of $160.39 on March 18, 2011 (point 1), and then rallied to a high of $206.39 in early May (point 2). During the rest of May and into June, AMZN made a series of lower lows (line b).

Those that bought after either the first or second decline were likely stopped out, as the final correction low (point 3) was at $181.59. This was just below the 50% support at $183.49, but held well above the 61.8% retracement support of $178.08.

From the late May high, the upper boundary of the flag formation (line a) was drawn. This resistance was broken one day before the low, but not on a closing basis.

Buying at the 50% support with a stop 0.5% under the 61.8% level was a risk of $183.49 - $177.18 = $6.31. The upside target would have been either the prior high of $206.39 or the 127.2% retracement resistance at $213.44 that was based on the decline from point 2 to point 3. Therefore, the potential rewards ranged from $22.90 to $29.95, making for a favorable risk-reward on the trade.

Apple (AAPL) is a stock that consistently gets lots of media and investor attention. In early July, AAPL overcame the five-month downtrend (line c) and accelerated to the upside. It eventually hit a high of $404.50, as their earnings beat all expectations, causing a buying frenzy.

Those that bought on the earnings were quickly disappointed. Nine days later, AAPL hit a low of $353.02. This was a 12.7% decline from the highs. The 50% retracement support at $357.50 was broken before AAPL once again rallied sharply back to $384. At this point a flag formation (line e and f) was evident.

The 50% support and line f were again tested before AAPL was able to rally above its prior highs, reaching $392.08 before last week’s decline. This created a broader flag formation (lines d and f). A convincing close above $392.08 will complete the larger flag formation.

Those who bought the secondary test of the 50% support at $357.50 could have either used a stop 0.5% under the prior low or 0.5 % under the 61.8% support at $346.40. In the first instance, the risk would have been approximately $6.25, while using the wider stop would have had a risk of $12.80.

The potential upside targets were either $404.50 or the 127.2% retracement resistance target of $418.90. The potential reward was therefore $47 or $61.50. Therefore, with either stop this strategy had a decent risk-reward profile.

The same approach can, of course, be used on weekly or even hourly flag formations. I am a strong advocate of scaling out of positions, and after a position is established I will typically place an order to close out half just under the 127.2% retracement target. If there is a significant reference point on the chart, I may put it even lower.

Also, once a triangle has been completed, I will often raise my initial stop to further limit the risk.

I hope this discussion will help you look at charts in a slightly different way. It is often very hard to identify major bottoms or tops, but once they have been completed, flag formations can help you find many good risk-reward entry points within the major trend.