Jason Mitchell
Jason Mitchell is a full time trader and popular writer on the practical aspects of trading. His tools and strategies are developed to be easily applied to the markets. This article introduces constructing and trading flag patterns on charts.

CONSTRUCTING AND TRADING FLAG PATTERNS

       

Constructing & Trading Flags: Part 1

In this series of articles we take a look at the flag pattern and will detail the methods we use of detecting, analysing and constructing the pattern. We will also discuss a few ways in which to find these patterns and how these opportunities may differ from the text book to real life examples.

When we are trading shorter-term precision patterns, we prefer the use of bar charts. We find these patterns are easier to see on a bar chart and this is about the only time we prefer them over candles. We use candlesticks in our management of the trade once opened.

Special Note on Flag Patterns The flag pattern is a short-term technique, which we believe is generally not suitable to novice traders. These short-term techniques require a higher level of discipline than many new traders possess in their first few years. Decisions on these often need to be made instantly. If you have found yourself procrastinating on a trend trade or slower investment then shorter term trades such as flag patterns may not be for you.

When the pattern works it looks like the easiest trade in the world, and maybe it is. But when they do not work - this is where people can get in to big trouble as they become much more difficult to manage. Many new traders do not enter a contingent plan for this occurrence and it would seem very few people discuss how they manage the situation of a pattern that broke out but did not hit its target. We will do this in this series.

What is a flag pattern? A flag pattern is specifically a shorter term trading pattern based on momentum and an excited crowd. The way in which it appears gives it its name and the following diagram shows very basically how the flag pattern looks. It does have fairly specific criteria regarding its construction and these are covered in the next section.

ASXEQ Advances (Found in the !0 folder by default)

For now we basically want to show that the pattern is characterised by a very strong sudden burst of interest followed by a small retracement. These two occurrences are the setup. The excitement is characterised by the long flag pole and the small retracement creates the flag. The long flag pole is often on much higher volume than the surrounding periods and the retracement is on much lower volume relative to the flag pole.

The accepted psychology behind this is generally a large number of new buyers have entered the market and the excitement takes out the waiting sellers and they push prices up. The move has taken some shareholders by surprise and others may be happy to move in and take the price being offered while it is available. As these profit takers or shareholders move in to sell and buyers no longer have to chase prices, the price pulls back slightly. This creates the flag itself.

As the profit takers dry up however and the buyers remain the price generally resumes a strong upwards move. It is amazing how many times these patterns occur JUST PRIOR to announcements and in fact I jokingly refer to this pattern as using insider trading. (Our real time trade example at the time of writing this article was an example - MCR).

Now that we know a little about the psychology behind the move we will have a better understanding of what it is we are trading. On this note - we are not trading the stock but the brief amount of excitement in the crowd.

Construction Rules The construction rules that I cover here are those accepted by the majority of traders. These rules are fairly strict in this pattern and these have been created by the traders who go before us based on their mistakes. If you do not follow these rules then expect to get burnt with this pattern. In this series of articles we show an example of a flag trade that goes against expectations as well as a poorly drawn flag pattern and how quickly you had to react to save capital.

The Flag Pole While the pattern itself is not completed, it is detectable as the flagpole is forming. Of course we need to wait for the flag itself to be formed to confirm the pattern and there are some important notes regarding the construction of both the pole and the flag.

The first rule regarding the construction of the pole is that these generally occur over a short period of time from one day out to about 4 or 5 days at the outside. I had noticed in Leon Wilson's latest book "The Next Step to Share Trading Success" he mentions that he feels this is not set in stone and that he has seen flag poles that have taken up to ten days to form. He acknowledges that this is generally not the case however and most traders agree with our observation that they generally take one to five days to form. Daryl Guppy in his book Snapshot Trading writes in regards to the flagpole "it is created by one to five days of extreme and continuous price action".

The flagpole may have gaps in it however it will generally not have any serious retracements. If there are any down days at all they will generally be very small in range and have little or no effect on the direction. The following diagrams illustrate our point for those who like to visualise what we are talking about. Notice in the first chart there are gaps in the flag pole however price continues in one direction. In the second chart we have shown a very poorly constructed flag pattern (for many reasons which we will cover).

Lihir Gold

The chart on the left shows a well constructed flag pole. The price moves upwards for four days in a row on heavy volume without retracing at all. There are gaps showing the excitement of the crowd and while the daily trading range is not huge on its own (i.e. bars not overly long) the rate of change over these four days is significant. Note: the long line to the left of the flag is a line I have drawn measuring the height.

Lion Selection Group

This second chart shows a poorly developed flagpole. It did have a huge day and then a second up day as it broke away from its base. It then had a down (red) day (on reasonable range and volume) that closed lower and seemed to slow momentum. The price did gap up again after this day continuing the momentum and this move defined the top of the flag pole. Price then retraced in a flag like manner although there are issues with the development of the flag as well, these are covered later. NOTE: This is an example of poor flag pole construction.

Real-time flag pole

The diagram to the left shows one of the flag patterns we monitored in real time in our newsletter.

We believe this is a text book example of a flag pattern. The price moves up strongly for four days (the Doji shape on day 3 is of small concern as day 4 gaps up on the open and closes at its high).

There is no retracement in the flag pole itself - in fact there are no red days at all and the range had increased considerably. The flag was also well defined and we'll cover this in Part 2 of this article.

Before we move on from flag poles we feel we should again mention that flag poles are impossible to pick in real time. It is not until the flag itself begins to develop that we will know it is a flag pattern. I tend to be very strict regarding the rules of flag pole construction. These rules have been suggested by other traders as they have found that by sticking to the rules we have a better chance of probability.

In Part 2 of this article we will cover the construction of the Flag itself and how to set targets based on these.

 

Constructing & Trading Flags: Part 2

Construction of the Flag

Length of Development
Just as we have some rules regarding the development of a flag pole, there are also strict rules regarding the construction of the flag itself. As part of a short term movement the flag itself is generally formed over about 4 to 5 days. It can be less although it is hard to confirm a pattern with less than three days. The flag can take a little more time to develop than the pole but generally 5-6 days is about as long as they last. We again noticed in Leon Wilson's new book "The Next Step to Share Trading Success" that he recognises that flags can take up to 10 days to form, we would tend to agree with this figure more than we did in the development of the flag pole. It should be noted however flags can form very quickly, and the breakouts happen even quicker. Daryl Guppy suggests that flags that take longer than 10 days to develop lose their "vitality" and are not as likely to hit or exceed their target.

Rules of Development
The flag itself is made up of two parallel lines. These lines must be parallel for the flag to be valid. If the lines come together this is a pennant pattern, if they move apart then it is definitely not a flag pattern.

Flag Pattern - Parallel
Flag Pattern - Parallel

 

Pennant Pattern - come together
Pennant Pattern - come together

The flag should also be on a downward sloping angle and not to the side or upwards. Surprisingly flags with upward angles actually have less likelihood of achieving their targets in fact some say it is a bearish pattern. The angle of a flag is hard to determine although Leon Wilson suggests a figure of 70% is an optimum angle and anything less than 45% is usually less reliable. We do not calculate the angle in our own trading.

One of the most important features of a flag pattern which determines its strength is the height at which the flag is on the flagpole. The flag should be at the very top of the pole and not towards the middle. The lower the flag the less probability of the pattern reaching its upside target.

Flag: Top of Pole/Half Mast

If the flag is flying half mast (therefore the flag begins halfway up the pole) then the pattern is unlikely to reach its target. We do not trade the pattern if it is not towards the very top of the flag pole.

LSG

Looking back to the earlier chart of LSG, which we felt showed a weak flag pattern (shown left) we would point out a few issues with the flag itself, (we already discussed the pole). While the parallel lines defined almost the entire flag one of the first day's highs actually went above the line. If we decide to call this the top of the pole as some trader's may then the pattern started just below the top of the pole, which is also a weaker sign. Either way the pattern is not as strongly defined as the others that have been shown.

The fact the pattern did seem to follow the psychology of the flag shows that it was a flag - just a very poorly constructed one. These generally have less probability of hitting their target and the fact that momentum did not increase immediately on the breakout confirms this pattern was not as strong as others. It did exceed its target in the end - partly due to a bullish market.

Should prices break below the bottom line of the flag pattern the whole pattern is invalidated. Price must stay between these lines for the pattern to be valid. This is with the exception of course of the upside breakout. This is a very important rule.

The 50% Line
This is a rule that some traders use and many other traders don't and we almost did not include it for this reason. We tend to use it ONLY if the flag pole is made of one large day and preferably is a dominant candle. This is similar to the 50% mark on a dominant candle. We would suggest however that we only use it in certain situations and we mention it only for completeness in the article. We do note many of Australia's authors and traders (such as Daryl Guppy, Leon Wilson and others) do not tend to speak about this rule. The 50% line is the halfway point between the top and bottom of the flag pole. If the flag goes below this point then the premise is that the chance of it hitting its target on the breakout is reduced considerably. We do not subscribe to this as a general rule and some traders believe using such a rule will take you out of many good opportunities.

The 50% mark is simply measured by halving the height of the pole. For example if the pole went from $0.22 to $0.29 then the halfway point would be $0.245. This is found by finding the height of the pole in a dollar amount.
Eg. $0.29 - $0.22 = $0.07

Half of $0.07 is $0.025. Therefore we add this amount to the bottom of the flag pattern and find the half way mark (note we could also take the amount from the top of the flag pole - we just felt it was easier for people to add).
$0.22 + $0.035 = $0.255

Setting Price Objectives
One of the best things about patterns such as the flag and the ascending triangle is the pre defined price target that the pattern gives us to sell at. This allows us to ascertain a risk reward calculation as well as have a defined management plan should the trade go right. (We cover what to do when they go wrong in a subsequent article).

Again there are some strict rules regarding the setting of targets with a flag pattern. The flag pole is what we use to measure the target and the upper trendline of the flag itself is used as a projection point. What I mean by this is the height of the flag pole is measured in terms of dollars (as we did for the 50% line). This time however we take the height of this line and add it to the value of the point where the flag pattern is penetrated. It should be noted that the target set for flag patterns is a minimum target and is often exceeded.

For example let's say we have a flag pattern again with the flag pole going from $1.20 to $1.38.
The height of this flag is: $1.38 - $1.22 = $0.16

We then use this figure and add it to the value of the upper line in the flag pattern at the point of breakout. If the stock broke out of the flag pattern when the trendline was at the value of $1.32 then the minimum target is: $1.32 + $0.16 = $1.48

If the value of the line had been at $1.31 then the target would have been $1.47 ($1.31 + $0.16). This figure must be re-adjusted daily as the value of the trend line drops in the flag. This is of course until the pattern experiences a break above the line. The following diagram illustrates visually the concept of adding the projected height from the upper line of the flag itself.

Target - Height of Pole

If in the above chart the price did not break out on the day it has we would need to re-adjust the target based on the value of the upper line of the flag. If the line drops by 1c per day so does the target.

How Do We Measure the Flag Pole?
There are reasonably precise rules regarding how to measure the height of a flag pole as this is the figure used for the target at which we will sell at. Prices can fall back from their target reasonably quickly in pattern trades and so we need to be very careful to set the right target so that we know when to act. Even being out on the target by 1 cent in certain shares can mean the difference between a good result and an ordinary result. If the target is hit and price pulls back and you are not out then you may need to chase price downwards to get out.

The flag pole is deemed to have started on the first bar that experienced stronger than usual upward movement. As Daryl puts it in Snapshot Trading (p.119) "The flagpole starts with the day signalling the beginning of the fast rise".

This can be a little bit subjective however it most cases it is reasonably clear. The change in pace in a share price is often visible quite easily on a chart in hindsight and this is all we need to do with this measurement.

The fast rise can include days that gap up. In the calculation of the flag pole, if the initial day of the pole gapped up then we include the gap as part of our calculation.

The diagram above illustrates taking the gap of the move in to consideration for the height.

So the bottom most point of our flag pole is the point at which it starts to rise quickly, or to put it another way when momentum kicks in.

The top of the flag pole is calculated at the highest point prior to the retracement in to the flag pattern. The above diagram shows how the flag pattern starts just off the top of the pole however the height of the pole is measured to the point of the highest high.

    Basic Guidelines for Setting Targets on a Flag Pattern:
  • The bottom of the pole starts where price begins to accelerate or rise quickly
  • The top of the pole is the highest high before prices retrace in to the flag pattern
  • If there is a gap on the initial day of the flag pole this is included in the calculation
  • The height of the flag pole is projected from the upper line in the flag at the point of breakout
  • This target is generally a minimum - flags often exceed their targets.

In further articles Jason covers some of the more important points on trading flag patterns including how to find flag opportunities, how to enter in to a flag trade (looking at both an aggressive and conservative entry methods) and most importantly how to manage both successful and unsuccessful trades. He also shows a selection of actual examples of flag trades and patterns to help people recognise the differing ways in which these may appear on a chart in real time.

 

Constructing & Trading Flags: Part 3

Over the last few weeks we covered what a flag pattern was, the rules regarding its construction and how I measure targets. In the original articles from my newsletter I showed a real time example of a flag trade so that readers could see our management methods. I have included the notes of this trade directly from the newsletter as an example of a successful pattern.

When they work they are easy to trade. This is why I have also included an example of a flag trade that did not perform and the subsequent management of this. It should be noted that these notes were prepared between the 12th and 26th February 2005.

Looking at the XAO review we have decided not to add any position trades to our case studies or web site portfolio again this week. While this may surprise the newer readers, long time readers have probably already guessed this would be the case and have realised that this is partly why there have been no new charts on the web site lately. (We are not suggesting we are bearish yet).

This week we have decided to look at a shorter term trade using a pattern as our setup and trigger as this is in line with our own trading strategy.

The way in which we found this was very simple. We simply chose a letter (completely at random) and quickly scanned through each stock in the database. We were specifically scanning for patterns and in the first pass we looked for ascending triangles. We did not see any of these and so went through again looking for flag patterns. We found a stock that was developing a possible flag pattern and have decided to show how we may analyse this type of opportunity.

The above chart shows the stock price as it was at the close on Friday going back to April last year. The stock recently broke away from its base in a sharp upward move on heavy volume and has pulled back slightly on lower volume. The tops of each bar in the pull back have been in a straight line allowing us to plot a short-term downward trendline. Using this we place a parallel line and note that it is touching the bottom of price action. This is a short-term channel of a few days to maybe a week and creates the flag. The strong rise prior to this creates the flag pole and combined we have a flag pattern forming.

The JICD (Jason's Indicator Convergence Divergence) is a basic trend/rally following indicator originally designed by Jason Mitchell to help improve and locate entry points into new intermediate term trends.
This indicator as well as Berg, Guppy and Hull indicators are available in EzyChart 6.

Looking at the JICD (in the previous chart) we can see that it's peaks have been getting higher, it crossed zero on the breakout and has now returned to the trendline and has turned up and may be about to rebound away again. This is the type of setup we like to see with the pattern. We would also accept a rebound from the zero level however the trendline is more bullish.

I have taken the height of the flag pole and projected the height of this above the upper line of the flag. This was projected based on a break on Monday and will be re-adjusted if the break does not come at this time. The value of the line on Monday will be $0.71. I have drawn a line across the value of the lower trendline on Monday and a move below this invalidates the pattern.

We will buy as soon as possible on a break above the upper trend line. Using a contingent order at $0.72 on Monday could get us in - we will buy at $0.72 or $0.73 and will adjust these values each day there is no breakout. I can not chase prices any higher based on my desired risk reward strategies. As I trade full time I place my orders manually.

The above spreadsheet shows the position sizing for the trade. We expect prices will move quickly once they break above the pattern and as such we have used a very tight stop to get the biggest position size and the best risk reward ratio. We understand that this may work against us and that using a tight stop can actually reduce the success rate of a system. We are happy to take this risk however due to the high probability of the pattern.

If we get in at our preferred level of $0.72 then the trade offers a risk reward ratio of 4.34:1, which is acceptable. If the trade does not work we will have $336.00 at risk based on the stop and this could be larger quite easily. The possible profit that is available $1458.65 based on the target.

Now, this is where low price shares kill traders - if we need to enter at $0.73 (only 1 cent difference) then this puts $468.38 at risk (an extra 39.4% more than the purchase at $0.72). Based on the target the reward will be $1,301.70, this puts the risk reward ratio back to 2.78:1. We prefer trades that give us a 3:1 ratio although due to the probability of the pattern we will still add the position to the case study at $0.73 if need be. The plan is:

Raw Price Signal: Flag Pattern - Looking For Break
Confirmation: JICD trending up and possible rebound off trendline (5 periods)
Maximum position size available and risk reward of 4.3:1 on preferred entry
Sector Trending Up and outperforming the general index.
Initial Stop: $0.70
Position Size: 13,805 @ $0.72 for a cost of $9970.05 or
13,616 @ $0.73 for a cost of $9969.63
Risk: $336.00 (plus) on entry at $0.72
$468.39 (plus) on entry at $0.73
Management Tools: JICD (5 Period) to measure momentum if prices fail to reach target.
Candlestick Patterns.
Plan: Buy on the breakout from the flag pattern using a tight stop for risk reward.
Sell as prices hit the target.
If pattern fails to hit target within 5 days re-assess the stock and position.
Sell on weakness of the JICD (eg. Divergence)

The following notes are the update on this trade for the following Week.

Last week we decided to look at a shorter term trade using a pattern as our setup and trigger as this was in line with our own strategy. We elected to use a pattern trade and came across a flag pattern that was well constructed and offered an adequate risk reward ratio.

The chart above shows the successful completion of this pattern as it exceeded its target on Friday. This article shows the management of the trade.

On the Monday we added the trade at our preferred price of $0.72 just after the open, we believe better trading could have gotten an entry at $0.71 as the market settled down, however for the sake of the article we take an entry at $0.72. The price did not shoot up on this day until just after lunch time so there was plenty of time to enter with this particular trade. As soon as the buyers stepped in however the price shot to $0.76 and rose from here. The last chance to enter was on Monday morning before the big buyers moved in.

The following two days failed to go anywhere however seasoned traders would have realised that the fall on the first day failed to make a significant dent to the previous day and the following day was a very short inside day. This created a very small but significant triangle pattern, which suggested that the price was in all probability pausing. Admittedly the first down day was not ideal however price failed to penetrate the previous days range by more than 50%. Penetration of the 50% mark of a large candle is an occurrence a number of momentum traders pay attention to.

Price continued to move up on the Wednesday to hit the target of $0.83. It was not possible to exit at this level on the day and the highest available exit was at $0.82. The huge number of buyers that had moved in however on market depth showed that there was likely to be short term (intraday) support at the $0.82 level. Using this knowledge (and knowledge of an increase in the dividend) we held off exiting until the Friday and took advantage of the price spike to exit just above the target at $0.84. We feel this was easily achievable on the day and prices in fact went to $0.85.

There was no need to confirm this exit strategy with any indicators as it was a set target. The trade returns $1,596.70 in one week which is equivalent to 16.66% and the trade ended up showing a risk reward ratio of 4.75:1. In a less bullish market we would have taken an exit at $0.82 on the previous day.

Interestingly there was an announcement made on the Monday and on the Thursday in this stock. The first was an announcement regarding a high grade drill intersection and the other was an increase to the dividends that were being paid after the release of the Half Yearly report. Without any knowledge of these announcements we have taken a good chunk of profit from their effect on the market.

As can be seen the management of this trade was relatively easy as it did what it was supposed to. This is an example of a conservative entry as we waited for the break above the pattern. The time that we can get in to trouble on these patterns with a conservative entry is when they fail to move ahead as expected or fall below our stops and we do not act.

Any one can manage a flag pattern that hit's the target but all successful traders will tell you it's about how you manage losses or things that go against the plan that makes a trader successful. As such you will rarely see much on failed trades as people generally don't want to discuss these. The rest of this article is devoted to showing management of a pattern that has not worked.

Managing a Flag Pattern that hasn't worked The following chart shows a flag pattern and its subsequent failure to reach its target. Going back to last week's article we can note a few points of weakness in this flag as it was forming which hopefully would have kept our readers out. The first is the fact that the flag has started just below the top of the pole. The higher the flag the stronger and more likely the pattern is to reach and exceed its target.

This flag started about 85% of the way up the pole, but still this was not up the top. The second issue is the angle of the flag which we estimated to be about 25 degrees. As we noted last week we do not generally measure the angle of the flag however we do take in to account if it is overly steep. Leon Wilson suggests that any flag with an angle of less than 45 degrees is less likely to hit its target. Combine this with the fact the flag started at 85% of the pole height and we should be aware of the reduced likelihood of this pattern reaching its target. This turned out to be accurate in this case.

Before we discuss possible management methods we should remind readers that the flag pattern is a short term trading technique. We are not in this for the long haul - we are expecting momentum to drive prices quickly to their target as they did in the recent trade example of MCR. A flag trade usually takes between 2 - 6 days to hit it's target after it has broken out although they can take up to 10 and can reach it or exceed it in as little as one day. So how could we have managed it had we entered? For the sake of the exercise let's say we took a more conservative entry on the breakout. It doesn't really matter where we take the entry as in this exercise we aim to show the management if it fails to reach its target.

The first thing we should understand about this pattern as I said in the first article is - we are not trading the price action but the excitement of a crowd. We expect the excitement of the crowd to push prices up in a reasonably quick timeframe. We call this type of move a rally and it is produced by an increasing momentum in the share price. We do have tools that are designed to help us measure the strength of this momentum and there are a number of momentum indicators out there to choose from. We have a preference for the JICD, the Williams %R or the RSI. There is no need to use all of these however and one should really do the job.

The chart below shows the breakout point and our entry (circled).

It also shows our chosen initial stop at $1.03 indicated by the thick red line. This is chosen as price has not closed below this level and should the excitement of the crowd continue we would expect momentum to take over and for prices to stay above this. If they do not we can accept we were wrong in our analysis and get out. Being wrong is going to happen - this doesn't matter too much. It is being able to judge how we will know we are wrong and then acting on this that matters. Being wrong happens to every trader.

As we have mentioned waiting for the breakout sometimes gives us a better chance of getting a good position size. We show the position sizing spreadsheet (available from our members section) that we use for most of our trading and an entry on the open as prices break out (this is a short term trade - timing is everything). This entry allows us to take 9466 shares at a cost of $9969.25. Based on the stop loss level at $1.02 (just under the stop at $1.03) we have 343.90 at risk. Using the target this gives us a slightly better than 3:1 risk reward ratio. If we used the figure of $1.01 however the risk reward ratio drops to 2.45:1 (this is not shown).

So far we have the following parts to our plan:

    Raw Price Signal
  • Flag Pattern
  • Buy on the breakout from a flag pattern
    Confirmation
  • Risk Reward ratio > 3:1
    Initial Stop
  • Short term resistance at $1.03
     
    Exit Conditions / Management
  • Sell at target level of $1.17

Unfortunately some people leave it at this point and are happy to enter the trade. We however have only looked at 2 out of the 3 possible outcomes. The first outcome is price falls back and we exit using our initial stop. This is easy and MUST be followed. It's going to happen and that does not on its own mean you are a bad trader. A number of unsuccessful trades however would make you re-think your strategy on flags.

The second outcome is that price reaches the target; this has been covered and is easy. The third outcome is the price goes sideways or moves ahead but fails to reach the target. For every trade we need to develop a plan for the three possibilities of price action, which are:

  1. Price goes up
  2. Price goes down
  3. Price goes sideways

Earlier in this article (and in the first) we said that we feel the pattern shows excitement and that we are trading this excitement. If the excitement slows down our theory is that we would also expect to see momentum slow down. Combining this with our expectation that price will rise fairly quickly, the first thing we could do is to say we will re-assess the position if it has not reached its target within 3 days (short term trades call for close monitoring). At this point if we see a bearish divergence between our chosen momentum indicator and price action we could take this as a sign that momentum (and therefore the excitement) has slowed. We are trading the excitement so if this happens it is probably time to get out.

Seeing as we do not trade knowing the data on the right hand side of the chart, the easiest way to show this is by a day-by-day blow of how we may have managed it using our methods and ideas. This will show us the chart as it would have appeared at the point of a decision and not how it appeared in hindsight.

Day 2 (Tuesday)
The Day After the breakout prices did move ahead however we notice that the close is at about the midway point of the day. This would give us an upward shadow. The JICD is rising showing momentum continues to rise.

There is no need for any concern on this day other than to note the fact price closed at about the halfway mark.

Day 3 (Wednesday)
The next day is very short in its range and the open and close are at the same level showing a lack of dominance by either group. Combined with the fact that that the close was the same as the previous days close we can see there had been a lack of short term momentum. There is no need to be overly concerned on this day although we again note that the momentum is not very strong.

Day 4 (Thursday)
By day 4 we are beginning to become a little more worried as price has failed to get very far in the four days and the close has been at the same level for three days. There is no momentum here at the moment and we have placed a very tentative short-term trendline that points downwards on price action. On the JICD we have placed a horizontal line at the last peak. It is too early to plot a divergence as the 2nd peak has not been formed but the line shows that the indicator has failed to make a new high at this stage.

It is possible that some traders would be considering an exit at this point as price fails to move ahead. This is a valid strategy however I would be inclined to give it a little bit more time seeing as it has not retraced. An exit on the close of this day would have given the trader $318.74.

Day 5 (Friday)
Now the closing price has fallen and the short term downward trendline has had its third touch. Further to this a new low for the week had been created. The JICD indicator had shown a divergence as the peak had been formed (JICD now headed down).

This is the point that I would have considered taking an exit (if I hadn't already). There is enough evidence to show that the excitement had died down and five days is a fairly long time in a flag pattern. An exit on the close at $1.08 would mean a small profit of $224.08 after brokerage.

Remember we are trading the excitement and after 3 days of nothing and the close actually getting lower we were happy that the excitement had seemed to disappear.

Day 6
Had I stayed in the stock until this point I would have decided now that price and momentum were well and truly gone. The close had dropped back again and the range of the day was very short again showing that there was no dominant force at this point. Obviously the divergence in the indicator is still visible and is even more pronounced now as the JICD continued to fall.

An exit on this close would have meant the trade did not lose money and we would have taken $129.42.

Day 7
We have now had 7 days without one day of upward momentum in the last 4 days. This 5th day that lacked any major movement would have been a definite sign that this pattern has failed. It has been given every chance by this stage and has literally gone nowhere. Price still failed to close above the downward trendline.

Patient trading may have been rewarded here with a small profit of $224.08 on an exit on the close. Often in these patterns patience is not always the greatest thing. I would have exited on day 5 or 6 (and possibly earlier if better opportunities had presented themselves.

I have also included the following chart, which illustrates other momentum indicators showing a divergence by the time we suggested we would take an exit. We usually use the JICD however appreciate that readers may prefer to use other indicators. We have shown the Chande Momentum Indicator, the Williams %R and the RSI. These are all shown with their 14 period versions. We would not recommend using all of these and skill with one should be more than enough.

Before we leave it for this week, I thought I'd show another failed flag pattern. One reason is to show some issues with the flag and the other is to make sure readers know these patterns do not lead to definite results. They do put the balance of probability on our side if well constructed, although this is not.

This chart shows a stock where there has been a flagpole form. There is a somewhat definable flag created as price retraces. The major issue is the speed of, and the gaps created, during the retracement. The flag starts below the top of the pole and the angle is too severe for my liking. This is not a high probability pattern.

The subsequent price action did experience a break out however price pulled back from this level and it is unlikely that an exit would have been achievable at a profit.

 

Constructing & Trading Flags: Part 4

In the final part of this article on Flag Patterns we look to methods I use to locate trade opportunities and briefly cover the difference between the two types of entry strategies - conservative and aggressive. Finally I provide a number of recent Flag Trade opportunities and how these patterns were presented by the charts.

Finding Flag Patterns There's some good news and some bad news about trading flag patterns and that is they are not the easiest things to find in the world, but they can be found. There are basically three methods that can be used.

  1. Eyeball search
  2. Price Volume Breakout Searches
  3. Specific Flag Detection Searches

Eyeball Search The first method we have mentioned is the eyeball search. We know that this does not suit some of our readers however the truth of the matter is that it really is one of the best ways. Further to this the eyeball search gets you looking at a large number of charts repeatedly. We feel this is one of the best methods of practice when learning about technical analysis.

{Filter}
C > 0.10
AND
C < 1.50
AND
(Mov(V,50,e) * Mov(C,50,e)) >100000

You can run a search that restricts the stocks that you will look at and then eyeball the remaining candidates. For example you may use a search that finds all the stocks between $0.10 and $1.50 with the average volume being higher than $100,000. You may adjust this as you see fit.

Alternatively you can do what I often do (and copied from Daryl Guppy) and choose an alphabetical folder at random and scan this. For example you may choose the letter M and then look through this folder. This was how we found MCR. It is possible to use any folder including a list of your preferred stocks, the ASX300 stocks, a list of stocks from fundamental reports or any criteria that you feel comfortable with. It really does not matter how you choose the folder as can be seen by the random letter approach.

The main thing is to make sure that you do not spend too long at any one chart. 3 - 4 seconds should be more than long enough. If you don't see the pattern it probably is not there - move on and don't over analyse. If you are in doubt leave it. There will be more opportunities around the corner.

When I scan through the stocks then I am looking for one of two things. The first is a possible flag pole. I save any charts I see that are possibly developing a flag pole in to a separate folder or write them on a list called Developing Flags. This allows me to monitor their progress and see if a flag pattern develops.

The second thing I am looking for is a flag that has started to develop. This is what we saw with MCR last week and we only just saw it in time. The trick is again to look for the flagpole and then assess the subsequent development using the rules covered in part 1 of this article. The flagpole due to its change in range is the easiest thing to spot.

Price Volume Breakout Searches The price volume breakout search can come in many forms and is basically searching for a flag pole developing. Once these have been found it is necessary to put these in to another folder and monitor there progress to see if they develop a flag pattern. In Daryl Guppy's book Chart Trading (I think) it has a search for a price volume breakout. This means price has increased by more than 5% and when the volume is greater than 50% higher than the average volume.

The formula as Daryl has it is already in MetaStock as: Equis - Price Volume Breakout

I however have altered this formula to get rid of a lot of the stocks that it brings up. The main change I have made is I do not want to know about stocks worth less than $0.10. I generally do not like trading these types of stocks due to the excessive price leverage. Everyone thinks price leverage is good but it goes both ways and can become difficult to set reasonable stops. As such I wipe them out.

The second change I have made is to get rid of the many illiquid stocks where the chart is full of days with gaps in them. If the stock is showing excessive gapping I find these become much harder to trade and further to this the majority do not give me any reasonable trading opportunities. Inserting a simple code helps us to get rid of any stocks that have had more than 15 gaps in a period of 50 days. The alteration to Daryl's formula is:

ColA	Close
ColB	Ref(C,-1)
ColC	ROC(Close,1,Percent)
ColD	Volume
ColE	Mov(Volume,50, Exponential)
ColF	((Volume-Mov(Volume,50,Exponential))/Mov(Volume,50,Exponential))*100
Filter	V1:=If((GapUp() OR  GapDown()),1,0);
	Gaps:=Sum(V1,50);
	colC >= 5 AND colD >= (colE*1.5)
	AND
	C > 0.10
	AND
	HHV(Gaps,50) < 15

We have found this search is adequate for looking for flag poles however it should be noted that it will miss those that do not rise more than 5% in a single day and it will also need to be run daily to ensure that you do not miss any developing opportunities. From this point the flag still needs to develop.

Essentially ANY search that looks for a sharp rise in price will help pick up flag poles. Many people make use of the Rate of Change function, which allows us to look for stocks that have moved a certain amount.

One of the other searches that we created and have used occasionally is what we term the flag finder search. It does bring up a lot more candidates than the above search, as it does not require price to have moved 5% in a single day. It looks for a spike in the price over the last five days. As such this search will sometimes bring up candidates that are just beginning to form a flag.

{insert in filter}
ATRP:=(ATR(2)/ C)*100;
V1:=ATRP/Mov(ATRP,50,E);
V2:=If((GapUp() OR  GapDown()),1,0);
Gaps:=Sum(V2,50);
Pole:=(V1 >=2.1) AND (PDI(3) > MDI(3)) AND V > (1.5*Mov(V,50,E));
(BarsSince(Pole) < 5) AND (PDI(14) > MDI(14))
AND ((HHV(Gaps,50) < 15)) AND (C > 0.10)
AND ((Mov(C,50,E) * Mov(V,50,E)) > 80000)

The best way to work out which search is best for you is to run both over a folder and scroll through the candidates. Choose the search that gives you the candidates you feel comfortable looking at. Of course many of the results between the two will be the same and I have found most people prefer to use the price volume breakout search and monitor the developing activity for a flag pattern. Most methods revolve around looking for a flag pole. Eg. Price volume breakout, or the flag pole search mentioned above.

If you do use the search for the flag pole and run a separate folder watching the stocks as they develop then we would suggest that you delete any stocks that do not conform to the rules on a flag pattern immediately. There is generally no need to watch the stock once it has been invalidated in any way and this will help keep this list very small. (Often within a day or two a flag pattern can be ruled out).

Specific Flag Detection Searches We are not going in to this in too much depth except to say that these are available. I haven't come across any that I could confidently say picks flag patterns with any great success, although there are some interesting products.

Daryl Guppy's newsletter ran an article on CPFinder, which is pattern recognition software that integrates with MetaStock format data to look for a number of chart patterns. If you want to see the review that Daryl gave of this we would refer people to his "Tutorials in Applied Technical Analysis" Newsletter, October 30th to November 20th editions.

I personally do not use the software but I acknowledge that it does find patterns and can be very useful. We would recommend checking it out to see if it suits your trading needs. Make sure however that you have stricter rules regarding your pattern construction than the program - it will pick up a lot of "flags" that are not flags. This is not a major issue as the program helps to save time - it does not make our decisions.

Entering In to a Flag Trade Generally it is accepted that there are two methods of entry in to a flag trade. One is considered a more aggressive approach and the other a more conservative approach. The aggressive approach aims to enter in to the trade as the flag is forming. This often allows for a better reward if the trade goes ahead however there are a few issues that make this type of trade more difficult.

The second or conservative method is to wait until price breaks above the downward channel formed by the flag. This is considerably easier in some respects however it can mean missing out on opportunities as the price can move exceptionally quickly once it breaks out. On the site and in our newsletter we tend to show the conservative method more as it is easier for people to use however we do often take a more aggressive approach and buy during formation occasionally.

The Aggressive Approach
As mentioned this approach involves taking an entry in to a flag trade as the flag has formed and in anticipation of a breakout. This method takes some skill and confidence and our experience has shown it is a more risky method and has a slightly lower probability of success. When you get it right the rewards can be great - get it wrong and things can go pear shaped very quickly.

In order to demonstrate the aggressive approach we will use the chart from our recent trade example MCR but approach it in a different way. Our actual example and personal trade took the most conservative approach and this was partly to do with the time at which we found it. However we had anticipated the breakout when we looked at it on Friday.

Taking an aggressive entry requires very good timing and nerves of steel. The reason for this is that we need to be sure that the flag has formed. We then need to ensure we do not enter too early and then need to try and time the breakout. It would not have been so difficult in the example that we use however as we look at the plan you will see the downfall's to this approach.

Looking at the chart we would start drawing the upper trend line on the day marked 2. This is the 2nd day of the flag formation. At the same point we draw a parallel line touching the lows. With this chart it was easy, however in other charts we would place this tentatively. We would not wait for confirmation of this line with this pattern (this brings up an interesting point which we will cover later).

On day 3 we would be reasonably confident that the pattern had formed - it could still fail - these patterns are no guarantee that they'll hit the target, but we are comfortable that the flag had formed. It would be a very aggressive entry on this day however as prices closed in the middle of the day and did not even get near the top trend line. Some traders would be happy to enter however so let's look at the issues with an entry at this point. We take an entry at $0.70 on the close.

While it is not hard to take an entry on the close we need to define our position size for risk management. This sounds easy at first, the pattern is invalidated by a close below the lower trendline so this is our exit point - this is fine except the trendline is going down - this means our stop is LOWERING. This is a dangerous way to trade and for almost all cases we would NEVER recommend that you lower a stop. So where do we place the stop. If we place it below the value of tomorrow's trendline and price does not break out tomorrow then we need to lower the stop to the value below the line on the next day and so on. Our stop could be lowered 3 or 4 times before it takes out or we are stopped out. Let's look at what happens in this case.

For the example we use the entry at $0.70 and a stop below the next day's trendline. The trendline itself is at $0.67 so if we say $0.66 then our position size available is only 11000 shares.(based on maximum risk and trade size of $500 and $10,000 respectively). This costs $7,729.95 and puts at risk $499.90. Based on the target the risk reward is 2.74:1. Readers know that we prefer 3:1.

On the following day price does not break out and the stop needs to be lowered. It needs to be lowered because the lower trendlines value is $0.66 and so a close at this level does not render the pattern invalid. To lower the stop to $0.65 then puts $627.90 at risk suddenly going over our maximum allowable limit. The fact more is now at risk means the risk reward ratio has dropped to 2.84:1. If this kept going on for a few days then we are just putting more and more money at risk.

This illustrates the importance of two topics when dealing with flags. The first is we need to try and anticipate the breakout and not just enter and wait - this takes a lot of skill and many hours of watching charts on an intraday and end-of-day basis. The second is that the steeper the flag the quicker this stop will lower and the more difficult it is to take an aggressive entry.

By the time we get to day 4 the prices gap up on the open and hit the value of the downward trendline. Despite the small range of this day the gap up on the open is a bullish signal and a sign for some aggressive traders to enter using market depth for timing. Entry on this day was available at $0.71 and $0.72. Again however we still face the dilemma of the lowering stop and this time the stop is further away. This again illustrates the point we make in that timing is the successful trade off between risk reward and probability. The chance of continuation would seem slightly more probable on this day however the risk reward is reduced.

Based on the exit below tomorrow's trendline value we can only take a $6,279 position, which still has $500 at risk. This also puts the risk reward ratio back to 2:1.

I have heard the solution that a flag pattern should only take up to 10 days to form and 3 or 4 have gone so why not base the stop on the value of the trendline in another 6 days, the floor is going to be the severely reduced risk reward. Even based on our most aggressive entry point at $0.70 if we had used the trendline value of 6 days away the position would have been reduced to $5,134.50 and the best risk reward down to 1.79:1. If it did take six days the risk reward ratio would be less, as the target would also be reduced. This is an unacceptable alternative in most cases. We believe this should be assessed based on the risk reward calculations.

Aggressive entries are best made in slower sloping triangles when we are as sure as can be of a breakout. At this point it is often better to use a tight stop. This does reduce the success rate of the pattern as a tight stop can mean being whipsawed out of the trade, however our experience has been that during a bullish market the small reduction in the success rate is well worth the increased risk reward ratio.

We could have just shown you a historical example and said "An aggressive entry here with this position size leads to this profit". The reason we show the position sizing and initial stop methods however is because these are the issues that we actually face with a trade as opposed to just looking at one. When you do these calculations you often realise the trade offered a terrible risk reward ratio and had we tried it a number of times we may have lost money.

The Conservative Entry Approach
We do not intend to go over this in detail as the trade example in section 3 based on a real time example shows exactly how we would have traded it given the point we found the pattern. What we will do however is recap the fact that the conservative approach is entering as price break's out above the pattern. This often relies on a contingent order or full time access to the market to get in. It is essential however to have a limit as to how far prices may be chased and this limit should be based on the risk reward ratio available.

As we are waiting for the breakout it is more acceptable to run a tighter stop as the pattern has a relatively high probability of not coming too far back in to the flag (or even falling back in at all) - but it can happen. We would refer readers to last weeks update and the previous week's trade example for an example of a conservative entry.

We do not have a preference for either, as long as we are happy with the risk reward ratio. We do believe that waiting for the breakout is the easiest and least stressful option.

Summary
Over the last few weeks we covered what a flag pattern was, the rules regarding its construction, how we measure targets and last but not least the management of these trades once opened. This week we will wrap up our recent articles on trading flag patterns with a quick summary of some of the more important points that we covered. These have been set out below and then we have included a few examples of flag patterns and how they appear "in the wild".

    Pole Construction Rules:
  • Created by 2 - 6 days of continuous, strong, upward price action (can take up to 10 days although this is less common).
  • Should have no major retracement or "kinks".
  • Flag pole may have upward gaps
  • Down days tend to be small in range and have little effect on the upward momentum

    Flag Construction Rules:
  • The flag is made of two parallel lines
  • The top line is placed first using two points. The bottom line is placed tentatively to this using only one point.
  • Flag should be at the top of the flag pole
  • Generally take less than 6 days to form. Can take longer but this is the exception. If they do, they generally tend to be weaker patterns.

    Height Measurement Rules:
  • The flag pole begins on the first strong up day.
  • If there is a gap on the first day then the gap is included in the height of the flag pole.
  • The height of the flag is projected from the upper line of the flag at the point of breakout to measure the target.

    Invalidation Rules:
  • Close below lower trendline of flag
  • Price falls below the low point of the flag after the breakout.
  • Some people use the 50% mark of a candle or the flag pole.

    Entry / Initial / Management Stop Rules:
  • Aggressive entry requires anticipating the breakout.
  • Conservative entry generally easier (on break)
  • Place initial stop one tick below the lowest point in the flag or chose an appropriate chart based stop.
  • Sell as target is hit. Note however: flag patterns in a bullish market often exceed their target.
  • Sell on lack of momentum and upward movement after breakout.

    Methods of Finding Flag Patters:
  • Most methods revolve around looking for a flag pole. Eg. Price Volume breakout
  • Eyeball search of folder(s) looking for the flag pole and studying subsequent price action
  • Specific Pattern Detection programs such as "CPFinder"

The picture shown above for the setting of targets is just a lovely little flag pattern isn't it? Unfortunately not all patterns are this easy to see. We have included a few more flag patterns from actual charts so that you may see how they differ from the above text book example. The charts on the left show the pattern as it developed and the charts on the right show the subsequent price action. These were all recent opportunities (at the time of writing) and two are taken from actual trades.

These articles have been taken from a series on Trading Flag Patterns by private trader Jason Mitchell and his newsletter the StarTrader Report. During this series of articles Jason showed two real time flag trade's (i.e. chart shown before the entry) being MCR and CEY. These returned 16% and 14.57% respectively - both within 7 days.

First Published: 20 May 2005 - Copyright © Jason Mitchell

This document is copyright. This document, in part or whole, may not be reproduced or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise without prior written permission. This article needs to be viewed as educational reference only. It is not intended, nor is it to be regarded, as investment/securities advice or any other advice.