Chart patterns capture the development of crowd emotion and provide potentially high
probability trade ideas with well-defined price targets and exact measures of risk
management. But patterns ? by themselves ? do not necessarily lead to consistent outcomes. The development of chart patterns only alerts traders that one particular type outcome is more likely to occur than another. As price moves towards a selected price point, the trader pays more attention to the stock, ready to place a buy order if prices move a few ticks above that level. In other words, chart patterns signal that trading potential and the probability to take action may exist.
Chart patterns are an invaluable aid to trading, but only when they point the way to high probability outcomes. The key feature of a chart pattern is reliability. Why should we be interested in a chart pattern that works on average only 30% of the time? We?d get a better balance of probability by flipping a coin. As a ?performance filter,? we must trade patterns which provide a 70% or greater probability that the pattern will develop as expected. There are simply too many ways to lose money in the market without going out of our way to trade low probability patterns.
We use this high-probability performance filter to reduce the number of useful patterns to only a handful ? which also makes them easier to identify. The danger with all pattern trading is that we build castles in the sky. Instead of seeing what is really on the chart, we
tend to see what we want to see. Although there are few effective ways to overcome this rose-coloured view ? other than years of practice and strict adherence to our disciplines ? we can improve our pattern recognition if we are more precise about the specific features
we look for.
As an analogy, an automobile has four wheels, an engine, and space for passengers. This broad definition excludes some vehicles and does not help us decide which SUV is best for our family. To make a better buying decision, we need a better definition of the automobile we prefer. The same applies to pattern recognition in charts. I use very exact definitions which reduce the subjectivity and filter out look-alike patterns which have a low probability of success.
Classic Bullish Flag
One of the most powerful and consistently reliable patterns is the bullish flag applied to a daily chart (see figure 1). I use this as one of my main short-term trading techniques. The flag forms when the initial Working with enthusiasm for the stock slows down. There are few sellers as most new stockholders hold onto their recently acquired positions. Buyers
collect stock from long-time stockholders who are selling into the unexpected strength and who are frightened that the rally has failed completely. Gradually, prices move downwards but as you can see in figure 1, they maintain a steady trading band.
The ?flagpole? is the key initiating characteristic of this chart pattern. The flag pattern only occurs at the top of a flagpole. It is created by one to five days of extreme and continuous price action moving in the same direction. There are no significant retracements
during the construction of the flag pole. A flag is not a triangle nor is it a pennant. Like most flags, this pattern has parallel sides. Identification starts as prices start to pull back from the high created following the flagpole. We want two to three points to plot a tentative down-sloping trendline. This is the starting point for the pattern and it usually takes a minimum of three days to confirm.
Once the upper straight edge trendline is plotted, we hunt for the potential parallel trendline to define the lower boundary in this pattern. Simply take the upper trendline and plot it as a parallel line using the most recent low in the emerging pattern. We do not have to wait for two or three lows to develop before we plot the trendline. Instead, we infer the position of the lower parallel trendline and plot it from a single low. We then look for future price action to confirm the initial placement.
Why rush to draw the lower trendline? In a fast moving bull market, this pattern may develop over three to five days. Sometimes there are only two lower points in the pattern and, if we wait for validation from a third, we miss the opportunities for the bullish breakout. By plotting an inferred lower parallel trendline based on the confirmed upper trendline, we give ourselves the advantage of early recognition of the pattern.
When to entry?
The pattern may take longer to develop. When it does, we have more points with which to confirm the placement of the upper and lower trendlines. Aggressive traders buy the stock as it approaches the value of the lower trend line. More conservative traders wait for the pattern to ?prove itself? by breaking out through the upper trendline before buying. I tend to be aggressive, so I buy near the middle of the pattern spread when I am confident that the pattern is for real.
The initial stop loss is placed one tick below the value of the lower trendline. This gives room for price to move within the bounds of the pattern development. If the pattern continues its downward trend, the stop is moved lower. We do not want to be stopped out of this trade prior to full pattern development. In moving the stop lower, it is important to ensure the amount at risk in this trade does not exceed 2% of total trading capital (or whatever your trading plan defines as your maximum allowable loss per trade).
How far can we let the flag drop before the trade is abandoned? There are no hard and fast rules here. Exit signals include a drop below the value of the lower trend line. In most cases I find this pattern develops fully over five to ten days. As we move beyond ten days, the pattern loses its vitality. It develops a lower probability of success as the top of the trend line continues to fall; the potential returns are also diminished as the projected exit target also drops.
It?s also important to note that the flag patterns come in faster markets. This is not a slow pattern that takes weeks to develop. For example, flag patterns on weekly charts have a much lower level of reliability.
When to exit?
The bullish flag is a pause pattern which is formed while the market is redeveloping strength. To set the potential target we first measure the height of the flagpole. This is why the flagpole is such an integral part of a flag pattern. The flagpole starts with the day signaling the beginning of the fast rise. Usually this is very clear. The distance from the base of the flagpole to the top of the flagpole is calculated in cents.
This same value is then projected upwards from the point of the breakout above the flag pattern. This sets the minimum target which, in many cases, is exceeded. When the breakout occurs it is often very rapid and it is not uncommon for prices to gap above the top of the flag trendline. This is a disadvantage for conservative traders as they may end up chasing prices.
Bullish flag patterns develop quickly and in most cases, price moves rapidly towards the target projections in three to five days. This is not a slow trend. It is a continuation of the initial rally. But in some cases breakouts drift sideways, which indicates that the pattern has lost momentum. I abandon these trades, taking a smaller profit because the probability of success is reduced. The trades are still profitable but they have less chance of reaching the profit target.
The closer the chart display is to the ideal chart pattern, the higher the probability the trade will develop as anticipated. This is not always the case, as a recent personal trade with Singapore listed China Aviation Oil will demonstrate (see figure 2).
The flag was identified three days after the flagpole was completed. The flagpole starts with the first large range day, and the price projection gives a target of $1.59. The buy was activated at $1.35 providing an anticipated return of 17.8%. This is acceptable for a threeto-five day trading opportunity. Note that the lower edge of the flag swings off a single point. No further confirmation occurs.
The breakout from the pattern was powerful and the stop was immediately moved to the value of the upper edge of the flag at $1.38. This is an end-of-day stop based on a close below $1.38. Usually these breakouts continue uninterrupted, but day two saw a pullback with an intraday low at $1.38. Day three saw a resumption of the uptrend and an increase in volume. Had day three continued the decline of day two, the trade would have been closed at a small profit on the grounds that the breakout momentum had vanished.
Five days after the breakout, prices opened above the target level. I exited the trade at $1.66 which delivered a 23% return. It was not the best exit possible, as it was made from a hotel room in Hong Kong. (Although patterns provide us with good ways to manage the trade, this is not always possible in practice. Managing short term trades while travelling is not always a good idea.) Given time to access time and trade information, an exit around $1.85 would have been possible with a profit of around 37%.
I have used this example for several reasons. Firstly, and most importantly, real trading is not the same as perfect textbook examples. This trade highlights an important disadvantage of patternbased trading. Second, this pattern is best suited to short-term position trading ? when the impact of emotions is running high ? as these trades often overshoot their targets. Lastly, in a bullish market this calls for better intraday management of the exit, using your best judgement about the balance of buying pressure during the day. This management can add many extra points to the trade and, where possible, it is worth taking the time to do this.
Patterns capture emotions but they do not capture trends. China Aviation Oil continued to trend upwards, hitting highs of $2.05. Trading this strong trend calls for different trading techniques. Pattern-based trades are designed to take advantage of shorter-term opportunities and there is no point is mulling over regret because we missed the run of the main trend.
I prefer to use patterns for short-term position trades based on end-of-day charts. This method offers a reliable, high-probability trade with good returns and easy trade management ? with an added benefit of being able to effectively define and manage risk.