Chapter 4 – Projecting Price Targets

One of the biggest advantages of Point and Figure charts is ‘the count’, as it is called, which is the ability to project price targets from the chart. As with 45° trend lines, these counts are objective. They are potential targets that give the analyst an idea as to how far the price can go. It is important to stress the word ‘potential’ . These targets are a rough estimate. It would be naIve to expect them to be anything else. Sometimes the price may reach the target with pinpoint accuracy; sometimes it will be some way off. However, the bonus is that not reaching a target, or indeed exceeding a target, does give the analyst additional information about the price action and this will be discussed.

Unfortunately, Point and Figure counts are also one ofthe most abused techniques, because it is alluring to ‘know’ a price target, and computers have made it all too easy to place counts on charts at points where perhaps there should be none. Whilst it is an excellent and satisfying technique, it is just as well to note that there are rules which should be followed.

Two methods are used to establish price targets: horizontal and vertical counts. 3 -box reversal charts allow both vertical and horizontal counts, whereas I -box charts allow horizontal only. The horizontal counting method is different for I -box and 3 -box reversal charts, although, as you will see, the I-box method can be used with 3-box charts, but has fallen out offavour. It is rare to use any other box reversal charts for counting, but there is nothing to prevent it. When counting was first developed by Thomas Sexsmith nearly 100 years ago, no mention was made of vertical counts. Counts on I -box and 3 -box charts were horizontal counts only.

Counts on 1-box reversal charts

It is not possible to cond�ct vertical counts on I -box reversal charts, because, as you have seen earlier, columns can contain both Xs and Os, so only horizontal counts are possible. These horizontal counts are not quite as effective as counts on 3-box charts, partly because the counting method is not as precise. The logic behind the count is that the width of the pattern determines the extent of the subsequent move, and the area where most of the action has taken place is the level at which the count is taken. This is the pivot or anchor point, about which the pattern balances.

It is important to remember that if a I -box chart is plotted incorrectly, as discussed on page 62 in chapter 2, the width of the pattern will be twice what it is meant to be; consequently, the count will be double and, therefore, wrong.

One advantage of horizontal counts on I-box charts, however, is that they can be used effectively for counting across any congestion pattern, including continuation patterns. Every pattern yields a valid upside as well as a downside count. Once obtained, you must decide at the time the breakout occurs whether it is the upside or the downside that is activated.

207

The Definitive Guide to Point and Figure

There are a number of rules associated with I-box counts so it is important that you understand them, because I -box counting is subjective even ifyou have a computer. You have to be sure you are counting the right part of the pattern. There are a number of ways to establish the width of any pattern, all of which are valid.

How to establish a horizontal count on 1-box reversal charts

Step 1 – Look for a congestion pattern

You must look for a congestion pattern, which could be a top, a bottom, or a continuation pattern. At the time of counting, you may not actually know what it is. It is only when the price breaks out ofthe pattern that you will. It is for this reason that every pattern yields valid upside, as well as downside, counts up until the breakout.

Step 2 – Measure the width of the congestion pattern using the following rules: There are four methods used to measure the width of the pattern in no particular order.

Method 1

Count the number of columns in the row that has the most filled boxes, that is to say the row with the most Xs and Os. This is where most of the price action has taken place and, therefore, where the strongest part of the pattern is. It is the level that has been crossed the most times and can be regarded as the anchor point for the pattern. The number of boxes in the row is counted from the far left to the far right of the pattern, including any empty boxes that do not have an X or o. You are, therefore, measuring the total width ofthe pattern based on the width of the row you have chosen. The logic is that the more times a price level is passed through within the congestion pattern, the more important that price becomes in defining the width of the pattern. Please note that the row with the most filled in boxes is not necessarily the longest row in the pattern. You are not counting the longest row; you are counting the length of the row that has the most activity within the pattern.

Method 2

Count the width of each row within the pattern and divide by the number of rows to give the average row width, rounded up to a full box size. The trigger row is taken to be the row in the middle of the pattern. This was the method introduced by Thomas Sexsmith and preferred by De Villiers and Taylor.

Method 3

If the pattern has ‘walls’, that is to say a clear column entering the pattern and a clear column exiting the pattern, then count the number of columns in the pattern at the level of the start of

208

Chapter 4 – Projecting Price Targets

,

the right-hand wall across to, and including ,the left hand wall. This is the method preferred by Alexander Wheelan. Very often, this row will coincide with the row chosen by method 1 .

Method 4

Count the width of the pattern (number of columns) at the breakout or catapult point. This will either be the width between the entry and exit walls at the catapult point, or, if none exist, it will be the width of the pattern one row below the catapult point.

This may seem unnecessarily complicated and ambiguous. If all four rules are employed on the same pattern, it is likely that you will obtain four different counts. This appears to be a bigger issue than it is, because the way the horizontal count is calculated means that the four counts will be within one or two boxes of each other, and you will recall that counts are an approximation anyway. You will also find that most I-box congestion patterns are much shallower than their 3-box counterparts. You will find that some of the methods coincide at the same row anyway; for example, the most XS and Os row may also be the middle row of the pattern, as well as the start ofthe breakout column. It is not suggested that you use all four methods on each pattern but rather choose one method, considering the pattern as follows:

From experience, counting the width of the row with the most filled in boxes (method I ) is the best method.

If the pattern has walls, then choose method 3 because this fixes the width of the pattern.

If you are prepared to wait until the pattern breaks out, then choose method 4.

If the pattern does not have walls but has a number of rows with the same number of XS and Os, choose method 2.

Ifthe pattern is shallow, as most I-box patterns are, it will not make much difference which method you use.

Step 3 – Project the count up and down by an equal number of boxes
Multiply the number of columns calculated in step 2 by the box size (the value of each

X and 0).

•

Add this number to the value of the box in the row from which the count was taken, to achieve an upside target.

Also subtract this number from the value of the box in the row from which the count was taken, to achieve a downside target.

If the pattern has already broken out then you will already know the direction, so only the target in the direction ofthe breakout is valid.

209

Upside target = (number of columns in the row at the count level) x (box size) + the level at which the count is taken

•

•

Method 3 – The count row is at the base ofthe exit column, so count the number of columns across the pattern until you reach the column entering the pattern. In this case it is from 1 3 back to column 3, which is 1 1 columns.

Method 1 – The count row is the row in the pattern that has the most Xs and Os. It does not have to be the longest row. This is also 1 1 columns.

•

• •

Count (ii) – using method 4

Count (ii) may only be taken after the catapult point and is done so by measuring the width ofthe pattern at the catapult point. Because by this stage the pattern has broken out, only the upside count can be taken.

•

•

The box size is 10.

The level at which the count is made is row (i) at 1130. Upside target for method 1 and 3 is ( 1 1 x 1 0) + 1 1 3 0 = 1 240 D o w n s i d e t a r g e t fo r m e t h o d 1 i s 1 1 3 0 – ( l 1 x l O ) = 1 0 2 0

Chapter 4 – Projecting Price Targets

If, however, you decide to obtain a target before the breakout, you will not know the direction, nor will you be able to see the right-hand wall. In this case, you would use Method 1 , which is the row marked (i), with the most filled in Xs and Os. From this, you will obtain

an upside as well as a downside count.

In this example, however, methods 1 and 3 give the same width of pattern.

•

Target = (the number of columns from the far left to far right of the pattern taken at the row at the catapult point) x (box size) + the level at the catapult point

The row for count (ii) goes from column 2 to column 13, which means there are 12 columns i n the row.

The box size is 10.

The level at which the count is taken is 1 160.

Target= (12×10)+ 1160= 1280

211

The Definitive Guide to Point and Figure

Count (iii) – using method 4

Notice that there are actually two catapult points in pattern A. The second one at count (iii) is for a wider pattern. This often happens as a smaller pattern breaks out and forms a second wider pattern. Count (iii) is taken at the level of the second breakout.

Target = (the number of columns from the far left to far right of the pattern taken at the catapult point) x (box size) + the level at the catapult point

The row for count (iii) goes from column 1 to column 14, which means there are 14 columns in the row.

The box size is 10.

The level at which the count is taken is 1 1 80.

Target = (14 x 10) + 1180 = 1320

As you have seen, pattern A has produced four counts, two of which are the same. A pattern often does this. If the targets are near to each other, then it gives you a better target area to aim at. If they are very different, it is possible that once the first target is reached, the price will go on to the next target. In the case ofFigure 4-1, the area is between 1240 and 1320 with a mid-count of 1290, which is the most likely target.

Although pattern A initially produced a downside count as well, this was negated once the pattern broke to the upside.

Figure 4-1 has another congestion area, pattern B, around the 1220 level. The pattern could be a fulcrum top or a continuation semi-catapult. When counting a target from this pattern, therefore, you must calculate an upside as well as a downside target, because at this stage it could tum into either.

Count (iv) – using Method 1 or Method 2

In this case, there is clearly a row that has the most filled in Xs and Os, and this is the row that should be used. It also happens to be in the middle ofthe pattern, which means the pattern is evenly balanced.

Upside target = (number ofcolumns in the row at the count level) x (box size) + the level at which the count is taken

Downside target = the level at which the count is taken – (number of columns in the row at the count level) x (box size)

The row for count (iv) runs from column 14 to column 19, which means there are 6 columns in the row.

The box size is 10.

212

The level at which the count is taken is 1220.

Upside target = (6 x 10) + 1220 = 1280

Downside target = 1220 – (6 x 10) = 1160

As you have seen, targets may be obtained from every congestion pattern in a I -box reversal chart. The essence of the method is identifying which row represents the pattern, then calculating and projecting the target from that row. There has been some discussion about the importance of each row within a pattern and four methods have been put forward. Two of the methods, 3 and 4, may only be used once the congestion pattern has broken out and indicated its direction. Methods 1 and 2, on the other hand, may be used before the pattern has indicated its direction and consequently they yield counts in both directions.

In the case of methods 3 and 4, the row at which the count is taken is easy to reconcile. It is either the breakout row or it is the row at the start of the breakout column. In the case of methods 1 and 2, the row is the one which is the most important. Method 1 states that the row that has the most price action is the most important row. This is logical and easily reconciled. Method 2 states that an average row length should be taken and then the middle row should be used. This is less logical.

As to which method is best, it is impossible to say, other than to reiterate that the differences in the targets are so slight as not to be of importance. A checklist has already been given as to which method to use first.

I -box counts were designed initially for tick charts, that is to say charts constructed from the ticker where every price during the day is recorded. This is the most common application, but they may be used on any other time-frame including daily charts.

Chapter 4 – Projecting Price Targets

213

The Definitive Guide to Point and Figure

Chart 4- 1 is a 1 x 1 tick Point and Figure of the NASDAQ Composite Index. The price is in a downtrend and breaks out sideways, an indication that some congestion is taking place. Whenever there is congestion, there is an opportunity to obtain a count. At this stage, the direction from the congestion is uncertain. It could be a bottom or a continuation pattern, which seems the most likely. For this reason, the best count to take is along the row with the most filled in Xs and Os. The row marked on the chart has 1 4 Xs and Os, which is the greatest number in any row in the pattern. It should be noted that although this is also the longest row, it does not have to be.

Since the direction is uncertain, a downside as well as an upside count should be established from the pattern, giving two targets, 1 892 or 1 930. The pattern is quite evenly balanced, with support at the bottom and resistance at the top, so the break could occur in either direction.

1960

1900

1892 1800

Chart 4-1 : 1 x 1 tick chart of the NASDAQ Composite Index showing 1 -box horizontal counts

214

Chapter 4 – Projecting Price Targets

Chart 4-2 shows the congestion pattern becoming wider. Notice that the action appears to be towards the base of the pattern, indicating strength at the base and suggesting now that the price will break to the upside, but the direction is still unknown. What is interesting, however, is that although the pattern width is increasing, the count is exactly the same, because there is no other row which has more filled in Xs and Os.

NASDAQ COMPOSITE INDEX (IXIC) Up”9.!ili! Technical Analyst . 1 960

1 900

1892 1890

Chart 4-2: 1 x 1 tick chart of the NASDAQ Composite I ndex showing 1 -box horizontal counts

IxtemidTickPoint&F ure(cl)1 x 1

215

The Definitive Guide to Point and Figure

Chart 4-3 shows the congestion pattern eventually breaking out to the upside, at the same time as breaking the downtrend line, indicating that it is a bottom pattern rather than a continuation pattern in the overall downtrend. Because of this, the downside 1 892 count should be removed from the chart. Furthermore, because there is now a pattern with a wall of Os entering it and a wall of XS exiting, it is possible to establish another upside count at the base of the exit wall. The count gives a target of 1 928, which is only 2 points away from the already established 1930 count. Both counts are valid.

NASOAQ COMPOSITE INOEX (lXIC) UR.9.!!!� Technical Analyst D(lCmidTlCkPokd&F ure cl)1×1 1970

1960

1900

l B90

Chart 4-3: 1 x 1 tick chart of the NASDAQ Composite Index showing 1-box horizontal counts

216

Chart 4-4 shows the completed pattern. Notice that after the mid-pattern rally, which looked like a breakout, the price returned to support at the base, shown by the lower blue line. This is typical fulcrum action. The pattern continues to build around the centre, allowing another count to be established. Remember, to find the correct row to count, you must find the row with the most filled in XS and Os and once you have, you must count all the squares in the row including those that are blank. There are 22 XS and Os in the row marked A. There are actually 42 squares in the row, giving a target of 1 95 5 ( 1 9 1 3+42). There is nothing to stop you establishing a downside count at the same time if you think there is still a possibility that this congestion is a continuation pattern. This gives a target of 1 87 1 (not shown).

Eventually a column of XS leads the breakout of the pattern above the upper blue resistance line. Another count may be established at the base of the breakout column, marked B, giving a target of 1981.

Chapter 4 – Projecting Price Targets

NASDAQ COMPOSITE INDEX (IXIC)

IXIC mill Tick Point & Figure (el) 1 x 1

up<;J�t’!TechnicalAnalyst . 1990

Chart 4-4: 1 x 1 tick chart of the NASDAQ Composite Index showing 1-box horizontal counts

Another count could be established at the catapult point but this would yield a count only 1 point different from the 1981 count.

All the counts on the chart are valid. The 1928 and 1930 counts are achieved when the price consolidates and forms a semi-catapult marked C. The 1955 count is achieved at the semi­ catapult marked D, and the 1981 count is the top ofthe move at point E.

1900

217

The Definitive Guide to Point and Figure

Tick charts are not restricted to I -point box sizes. Chart 4-5 is a 2 x 1 Point and Figure chart of the NASDAQ Composite. A number of horizontal counts have been placed on the chart. In each case, the count has been taken at the row with the most filled in Xs and Os and, in each case, upside and downside counts have been established because, at the time of counting, the direction was uncertain. Notice how often the counts were achieved within a few points. The main thing about the counts is that they give you some idea of the potential upside or downside, should the pattern break out. Once the breakout does occur, the opposing count is removed.

NASDAQ COMPOSITE INDEX (IXIC) Up”9.!!1! Technical Analyst IKle mid Tick Point & Figure (el) 2 )( 1

Chart 4-5: 2 x 1 tick chart of the NASDAQ Composite Index showing 1-box horizontal counts

Horizontal counts are not restricted to tick charts. They may be established and applied to any time-frame including daily charts, as Chart 4-6 of eBay Inc. shows. The large base pattern breaks out sideways from the downtrend, indicating that accumulation is starting to take place. A small fulcrum bottom, marked A, is formed, yielding a target of 39 counted across the row with the most Xs and Os. The target was not achieved, indicating that bearish sentiment is prevailing. The price continued to trade sideways, creating a far bigger congestion pattern, B. Notice the small fulcrum top within the larger pattern yielding a downside target of 22, which was achieved. Pattern B eventually becomes so wide that a downside count cannot be taken. (See page 241 for more details on impossible counts.) The pattern does yield an upside count of 113, across the row with the most filled in XS and Os. This target coincides with the price at the top of the strong uptrend in eBay.

218

e� Inc ((BAy) Up”<;!�� Techncai l Analyst EBAYO ‘ Point&F e CI)1×1

B

Chart 4-6: 1 x 1 daily chart of e8ay Inc. showing the establishment and achievement of 1 -box horizontal counts

Large congestion tops on daily charts also provide an opportunity for downside I -box counts as the 50 x I Point and Figure chart ofthe FTSE 100 Index in Chart 4-7 shows. Within the large top marked B, there is a smaller top marked A, which has a wall of XS entering it and a wall of Os leaving it. The count for the smaller top A is taken at the start of the exit column and yields an accurate target of 4750. The larger top B also has a column of XS entering it and the same columns of Os exiting it. The count is taken along the same row and yields another accurate count of 3200.

Pattern C is an excellent example of distribution taking place after a rally during the downtrend. There is a clear wall entering and exiting, making the count easy to establish. Once again, it yields an accurate target of 3 800.

Chapter 4 – Projecting Price Targets

219

The Definitive Guide to Point and Figure

FTSE 100IUKXj

Uf’g’!.! Technical Analyst

�� �; � ====

———————–17000

UKKDai Point&

Figure(el) 0 1 5

4150

6500

6000

5500

5000

4500

3000

3800

3200

Chart 4-7: 50 x 1 daily chart of the FTSE 1 00 Index showing the establishment and achievement of 1 -box counts

Summary of 1 -box counts

I -box counts may be used to count continuation, as well as reversal patterns. The basis of the count is the width of the pattern determined by a number of methods, all of which yield a similar target. Although I -box charts, and hence I -box counts, tend to be used mainly for short-term charts using tick data, they may also be used with daily data and can count longer­ term targets, although this is where 3-box counts take over.

220

Counts on 3-box reversal charts

Not only are counts on 3-box reversal charts more popular, but also they are less ambiguous and easier to use than those on I-box charts. 3-box charts lend themselves to two counting methods, vertical and horizontal, because ofthe way the chart is constructed. The asymmetric filter against the trend means that 3-box reversal charts tend to have longer columns of XS and Os than I-box charts, thereby allowing vertical counts for which I-box charts are not suitable.

It is important to note at this stage, that 3-box counts cannot give an upside and a downside target from the same pattern. Only one direction is possible and is triggered on a breakout of the pattern, which is explained shortly.

Vertical counts on 3-box reversal charts

The vertical count did not exist until A.W. Cohen first introduced it in 1948. It has become more popular because there are more opportunities on a chart to use it. The vertical count measures the length of a column of XS or Os and projects it by 3 times that length. That is the easy part; however, choosing the correct column to count is important, so careful note must be made ofthe following guidelines. Vertical counts may only be established from the following columns:

The first move up off a bottom; in other words, the first column o f Xs after a bottom has been made.

The first move down off a top; in other words, the first column of Os after a top has been made.

The second move off a bottom if the second column is part of the bottom pattern, namely that the bottom is made up of either two Os at the same level, or two Os where the second 0 is only one box higher than the first and if the first column of Xs is a short column.

The second move off a top if the second column is part of the top pattern, namely that the top is made up of either two Xs at the same level, or two Xs where the second X is only one box lower than the first, and if the first column of Os is a short column.

Any other significant X or 0 column. This is not a licence to count every column you see. ‘Other significant column’ means either an intermediate mini-top or mini-bottom during an uptrend or downtrend, or a breakout column from a congestion area.

At this point it is important to understand that only these column types may be used to establish vertical counts. No other column should be counted, as this devalues the method and results in too many counts, giving a false impression as to the count’s effectiveness.

Chapter 4 – Projecting Price Targets

221

The Definitive Guide to Point and Figure

How to establish upside targets using the vertical count method on 3-box charts

Step 1 – Choose a column of Xs considering the rules above

•

• •

•

The most important count is one from an important bottom. Look for an important bottom at the end of a downtrend where a column of Os has reversed into a column of Xs.

The column of XS should be the first rising column from the bottom.

The column of XS must be a completed column. This means that the length of the column of XS must be fixed by the creation of a new column of Os to the right of it.

If, however, you have already counted from the bottom, then there are three other counts you may consider:

•

or

•

Look for a second column of XS off the bottom, but only if the first column of XS is a short column.

Look for an intermediate mini-bottom during an uptrend. The mini-bottom must have a ‘tail’ of Os. A simple 3-box pullback during an uptrend is not enough to trigger a count and is where most students make a mistake. The reason is that there must be some bearish action, which has been overcome by bullish action to increase the validity of the count. For example, a simple double-bottom sell would be sufficient to do this. As above, the column of Xs which you are counting must be complete and its length fixed by a new column of Os to the right of it.

or

•

Step 2 – Count the number of Xs in the column and calculate the count
Once the column has been chosen, and its length fixed by the emergence of a new

column in the opposite direction, count the number of Xs in the column.

Multiply the number of Xs by the box size (the value of each X and 0).

Multiply this product by the reversal, which is 3 .

Add this total to the value o f the lowest 0 in the column of O s immediately to the left of the counting column.

You now have the upside target, which you may mark on the chart.

Look for any significant X column breakout. This could be a column ofXs that breaks out of a sideways congestion, or any other X column that changes the look of the chart. As above, the column length must be terminated so that the count is fixed.

222

Chapter 4 – Projecting Price Targets

Figure 4-2 is a 5 x 3 Point and Figure chart showing three upside vertical counts. Each box is 5 points and the reversal is 3 boxes.

Figure 4-2: Vertical upside counts on a 3-box reversal chart showing columns that may be counted

Count 1

The price falls to a low in column 3. Column 4 is the first column ofXs offthe bottom. This allows count 1 to be established, once a reversal of 3 Os has been plotted in column 5 as follows:

Target = (number of Xs in column 4) x (box size) x (reversal) + lowest low in column 3.

There are 4 Xs in column 4, the box size is 5 and the reversal is 3.

223

The Definitive Guide to Point and Figure

The value of the lowest 0 in column 3 is 440.

Target= (4x5x3)+440= 500 Count 2

You may count the second column off a bottom if the first column is short. Count 2, therefore, utilises column 6 as follows:

Target = (number of Xs in the column 6) x (box size) x (reversal) + Lowest low in column 5

•

Count 3

To establish any additional count on the chart, you must find a significant column according to the rules detailed on page 221. The price spends a number of columns consolidating sideways after the move off the bottom but finally breaks out of the congestion area in column 12. Column 12 changes the look ofthe chart and is, therefore, a significant column. It is a column that has changed the look ofthe chart, so count 3 can be established as follows:

•

Do not assume that there will always be a third or fourth count. If they are not obvious, they don’t exist.

Note regarding vertical breakout counts

Some practitioners state that all vertical upside counts should use the bottom ofthe pattern as the anchor and that this is the value that should be added to. This is incorrect and illogical. Using the low of the pattern for all counts out of the pattern is self-defeating because there may be no direct relationship between the breakout column and the pattern low. Patterns where vertical counts are used are usually deep, making the connection between the low column and any column other than the adjacent columns spurious. For example, there is no

There are 10 Xs in the column 6, the box size is 5 and the reversal is 3. The value ofthe lowest 0 in previous column 5 is 445.
Target= (10x5x3)+445= 595

Target = (number ofXs in the column 12) x (box size) x (reversal) + lowest low in column 1 1

There are 1 2 Xs in the column 12, the box size is 5 and the reversal is 3. The value ofthe lowest 0 in column 11 is 475.
Target= (12x5x3)+475= 655

224

direct link between column 1 2 and the low in column 3 . If the low of the pattern had been used for count 3, the target would have been 620 instead of 655. This may seem trivial, but consider what would be the case ifthe breakout in column 12 had only been five Xs. This would have yielded a target almost identical to count 2, inferring that the sideways congestion after the bottom had no effect on the target from the breakout in column 1 2 . This contradicts the thinking on horizontal counts, which suggests that the width ofthe pattern does influence the count. So, in all cases, the low that is added to the count is the low of the preceding 0 column.

Count 4

Having counted the first move off the bottom, and the second move off the bottom, as well as the breakout column, the only other possible count is a mini-bottom during the uptrend – not simply a pause, but an actual bottom. This occurs in column 1 5, so the first move off the mini-bottom is therefore column 1 6, and must be counted as follows:

Target = (number of Xs in the column 1 6) x (box size) x (reversal) + lowest low in column 15

There are 8 Xs in the column 16, the box size is 5 and the reversal is 3. The value ofthe lowest 0 in column 15 is 500.
Target= (8x5x3)+500= 620

These four counts are the only valid counts in the chart in Figure 4-2. Do not be tempted to c o u n t a n y o t h e r c o l u m n s . y� s , y o u c a n e s t a b l i s h ‘ t a r g e t s ‘ o ff a n y o t h e r c o l u m n , b u t t h e y w o n ‘ t add anything to the counts obtained in the proper way and they will confuse the overall picture.

How to establish downside targets using the vertical count method on 3-box charts

The downside count is the inverse of the upside count, but it is just as well to go through it here so there is no doubt.

Step 1 – Choose a column of Os considering the rules above
The most important count is one from an important top. Look for an important top at the

end of an uptrend, where a column of Xs has reversed into a column of Os.

The column of Os should be the first falling column from the top.

•

Chapter 4 – Projecting Price Targets

The column of Os must be a completed column. This means that the length of the column of Os must be fixed by the creation of a new column of Xs.

225

The Definitive Guide to Point and Figure

If you have already counted from the top, then there are three other counts you may carry out:

•

or

•

or

•

Step 2 – Count the number of Os in the column and calculate the count

•

•

Although downside counts are really just the inverse of upside counts, it is important to have a reference example so that no misunderstanding exists. Figure 4-3 shows three downside vertical counts. Once again, the chart is a 5 x 3 Point and Figure chart. Each box is 5 points and the reversal is three boxes.

Look for a second column of Os off the top, but only if the first column of Os is a short column.

Look for a mini-top during a downtrend. The mini-top must have a peak ofXs, after a double-top buy signal. A simple 3-box rally during a downtrend is not enough to trigger a count. The reason is that there must have been some bullish action that has been overcome to increase the validity of the count. As above, the column of Os must be complete and its length fixed by a new column of Xs.

Look for any significant 0 column breakout. This could be a column of Os that breaks out of a sideways congestion, or any other 0 column that changes the look of the chart. As above, the column length must be terminated so that the count is fixed.

Once the column has been chosen and its length fixed, count the number of Os in the column.

Multiply the number of Os by the box size (the value of each X and 0).

Multiply this product by the reversal, which is 3 .

Subtract this total from the value of the highest X in the column of Xs immediately to the left of the counting column.

You now have the downside target, which you may mark on the chart.

226

Chapter 4 – Projecting Price Targets

T

550

540

2

510 X0X0X0X0 X0 0X0X0 500 xX o o o

–

530 X 0

520 X X0 X0X0X X

490X

480 X

oXX oX0 X0 oX0X X0
o oX0X0X

470

460

450

X X X

oX0X0X0 o 0X0X0

0 0 o0

Io 345678910111213141516

Figure 4-3: Vertical downside counts on a 3-box reversal chart showing columns that may be counted

Count 1
The price rises to a high in column 3 so column 4 is the first column of Os off the top. This

allows count 1 to be established as follows:

•

Target = Highest high in column 3 – (number of Os in the column 4) x (box size) x (reversal)

There are 6 Os in the column 4, the box size is 5 and the reversal is 3. The value ofthe highest X in column 3 is 530.
Target= 530- (6x 5×3)= 440

227

The Definitive Guide to Point and Figure

Count 2

Once again, to establish any additional counts you must find a significant column according to the rules detailed on page 22 1 . The price consolidates sideways but finally breaks out of the congestion area in column 8, which is a significant column because it has changed the look of the chart. Count 2 may, therefore, be established as follows:

•

It is important to remind you of the note regarding vertical breakout counts made on page 224. The high that the count is subtracted from is the high of the preceding X column, which is 5 1 5 in this case. It is not the high of the overall pattern.

Count 3

The third type of column that can be counted is a mini-top during the downtrend. This occurred in column 13, so the first move offthe mini-top in column 14 must be counted as follows:

Target = Highest high in column 1 3 – (number of Os in the column 1 4) x (box size) x (reversal)

There are 6 Os in column 14, the box size is 5 and the reversal is 3. The value ofthe highest X in column 13 is 485.
Target= 485-(6x5x3)= 395

Vertical count establishment and activation

It is important to note that there are two stages to the vertical count. There is the establishment stage and the activation stage. The establishment stage occurs when the length of the column being counted is fixed by the addition of a new column in the opposite direction. The count can then be performed and a target established.

The count and the target cannot, however, be considered active until there has been a break above the highest X in the counting column (in the case of an upside count), or a break below the lowest 0 in the counting column (in the case of a downside count).

This point was illustrated in Figure 4-2 on page 223. Column 4 is the first column ofXs offthe bottom. Its length is fixed by the addition of a column of Os in column 5 . A target of 500 may be calculated, but it cannot be considered active, in other words, it should not be relied on until

Target = Highest high in column 7 – (number of Os in column 8) x (box size) x (reversal) There are 9 Os in column 8, the box size is 5 and the reversal is 3.
The value ofthe highest X in column 7 is 5 1 5.
Target= 515-(9x5x3)= 380

•

228

The Definitive Guide to Point and Figure

The logic of the vertical count

As you have seen, the vertical count is taken on an upthrust or a downthrust column. The logic is that the greater the buying demand off a bottom or selling supply off a top, the stronger the participants creating the column. The analogy is throwing a ball: the harder or faster the arm is moved, the further the ball will go, overcoming any resistance. The greater the enthusiasm of the bulls in reversing the downtrend, the stronger they will remain during the uptrend as latecomers get on the trend expecting higher and higher prices. The converse is true for downtrends. The greater the move off a top, the more confident the bears become and the more shocked the bulls are at the change in trend, causing them to keep away.

It is no good, however, having just one thrust which could catch the other side off-guard; it is important that there is a reaction against the first thrust column, followed by a reassertion of the trend by at least a double-toplbottom breakout, and that is why the activation of the count is separately important.

Horizontal counts on 3-box reversal charts

Although only one method of3-box horizontal counts is really used today, there are, in fact, two methods. The original method was devised by Thomas Sexsmith, a colleague of De Villers and Taylor. It is along the same lines as that used for I -box horizontal counts described earlier, which was also devised by Sexsmith. In the interests of thoroughness, and in order not to disrupt the flow of the text, this method is discussed briefly at the end of this chapter.

The other method that is in common usage today is the method introduced by Cohen. 3-box horizontal counts can only be used when there is a wide congestion area, and when that congestion area is a top or a bottom. The prerequisite is that there must be a move into the top or bottom pattern, then a consolidation phase and a move out in the opposite direction. Some Point and Figure analysts, including this author, do not believe that continuation patterns can be counted on 3-box charts using this method. Ifyou wish to do this, you should use the original De Villiers & Taylor method described on page 252.

In the Cohen method, the horizontal count measures the width of any congestion top or bottom pattern and projects it by 3 times that width. Although not as popular as vertical counts, horizontal counts do yield worthwhile targets and should be used whenever possible, especially where they confirm vertical counts. Conditions for establishing a horizontal count are:

There must be a top or bottom pattern. There must be a move into the pattern. There must be some sideways consolidation.

230

There must be a move out of the pattern in the opposite direction, so the column type leaving the pattern is opposite to that entering the pattern.

The width of the congestion determines the extent of the move.

Chapter 4 – Projecting Price Targets

•

How to establish upside targets using the horizontal count method on 3-box charts

Step 1 – Look for a congestion pattern bottom
Look for any bottom patterns at the end of a downtrend that have the shape of a U, V or

W.

The shape of the pattern is not so important, but it must have walls.
There must be a column of Os entering it and a column of Xs leaving it.
Between the entry and exit columns, there must be some consolidation or congestion.

Step 2 – Count the number of columns in the pattern
Once the entry and exit columns have been chosen, count the number of columns across

the pattern, including the entry column of Os and the exit column of Xs.

Multiply the number of columns by the box size (the value of each X and 0).

Multiply this value by the reversal, which is 3.

•

Figure 4-4 overleaf shows four upside horizontal counts. For ease of calculation, the chart is once again a 5 x 3 Point and Figure chart. Each box is 5 points and the reversal is 3 boxes.

Add this total to the value ofthe lowest 0 in the pattern.

You now have the upside target, which you may mark on the chart.

231

The Definitive Guide to Point and Figure

Figure 4-4: Horizontal upside counts on a 3-box reversal chart showing rows that may be counted Count 1

Column 3 is a wall of Os entering the bottom pattern, and column 8 is the wall of XS leaving it. Between the two walls, there is some congestion. This allows count 1 to be established as follows:

Target = (number of columns in the pattern) x (box size) x (reversal) + lowest low between columns 3 and 8

There are 6 columns between the walls (including the walls), the box size is 5 and the reversal is 3.

The value ofthe lowest 0 in the pattern is in column 5, which is 455.

Target= (6x5x3)+455= 545

Very often, a bottom pattern has two or more parts to it. A smaller bottom pattern is often contained within a larger bottom pattern that develops at a later stage. This allows more than one count to be established from a single bottom. In the example in Figure 4-4, there are actually two additional counts.

232

Chapter 4 – Projecting Price Targets

Count 2

Notice that a smaller bottom pattern (count 1) bounded by columns 3 and 8 is contained in a larger bottom pattern bounded by columns 3 and 12. This allows count 2 to be established. Before doing the calculation, you can see that the target will be greater because the pattern is wider.

Target = (number of columns in the larger pattern) x (box size) x (reversal) + lowest low between columns 3 and 12

There are 1 0 columns between the walls (including the walls), the box size i s 5 and the reversal is 3.

The value ofthe lowest 0 in the pattern is still in column 5, which is 455. Target= (10x5x3)+455= 605

Count 3

In Figure 4-4 there is actually an even larger bottom pattern bounded by columns 1 and 12. This allows a third count to be established. The target will be greater than counts 1 or 2.

Target = (number o f columns i n the larger pattern) x (box size) x (reversal) + lowest low between columns 1 and 1 2

There are 12 columns between the walls (including the walls), the box size is 5 and the reversal is 3.

The value ofthe lowest 0 in the pattern is still in column 5, which is 455.

Target= (12x5x3)+455= 635

It is important to keep a look out for multiple counts from within the same bottom. As the bottom builds and the congestion area increases so additional, higher targets may be established. Do not, however, assume that there will always be a seco�d or third count. Ifthey are not obvious, then none exists.

Count 4

As stated before, using the Cohen horizontal method for counting continuation patterns is not recommended, so you may well ask why count 4 is being considered. It is a consolidation pattern contained within a continuation pattern, so in this instance it can be counted. In the same way that vertical counts may be established from mini-bottoms during an uptrend, so horizontal counts may be established from them as well. There is a move into the mini­ bottom in column 1 3 . There is some congestion during which a bottom is made, and there is amoveoutofthepatternincolumn 18.

233

The Definitive Guide to Point and Figure

The target is calculated as follows:

Target = (number of columns in the mini-bottom pattern) x (box size) x (reversal) + lowest low between columns 1 3 and 1 8

There are 6 columns between the walls (including the walls), the box size is 5 and the reversal is 3.

The value ofthe lowest 0 in the pattern is in columns 15 and 17, which is 500.

Target= (6x5x3)+500= 590
These four horizontal targets are the only ones available from the chart in Figure 4-4.

How to establish downside targets using the horizontal count method on 3-box charts

Step 1 – The first step is to look for a congestion pattern top
Look for any inverted U, V or W-shaped top at the end of an uptrend.
There must be a column of Xs entering it and a column of Os leaving it.
Between the entry and exit columns, there must be some consolidation or congestion.

Step 2 – Count the number of columns in the pattern
Once the entry and exit columns have been chosen, count the number of columns across

the pattern including the entry column of XS and the exit column of Os.

Multiply the number of columns by the box size (the value of each X and 0). Multiply this product by the reversal, which is 3 .
Subtract this total from the value o f the highest X i n the pattern.

• You now have the downside target.
Figure 4-5 shows three downside horizontal counts. The chart is also a 5 x 3 Point and Figure

chart opposite.

Count 1

Column 3 is a wall ofXs entering the top pattern. There is some congestion in columns 4, 5, 6and7,thenthereisawallofOsleavingthepatternincolumn8.Count1 maybeestablished for this pattern.

Target = Highest high between columns 3 and 8 – (number of columns in the top pattern) x (box size) x (reversal)

234

The Definitive Guide to Point and Figure

There are 1 2 columns between the walls (including the walls), the box size is 5 and the reversal is 3.

The value ofthe highest X in the pattern is 545. Target= 545-(12x5x3)= 365

Count 3

You will see that count 3 may be established from a mini-top during the downtrend. The mini­ top pattern is created by the column of Xs in column 1 3 entering the pattern and the column of Os in column 1 8 leaving it. The count is established as follows:

Target = Highest high between columns 1 3 and 1 8 – (number of columns in the mini-top pattern) x (box size) x (reversal)

There are 6 columns between the walls (including the walls), the box size is 5 and the reversal is 3 .

The value ofthe highest X in the mini-top pattern is 505. Target= 505-(6x5x3)= 415

Horizontal count stage

Horizontal counts don’t have two separate stages like vertical counts. The establishment and activation stage occur at the same time, when the entry column is matched by an exit column after some congestion. In the case of a bottom pattern, the exit column must rise above the highest X in the pattern so that there is a clear row running between the entry and exit columns. The width of the pattern is then fixed, the count can then be calculated, the target established and activated. The arrows marking the width of each count in the previous figures are placed at the point in the column where the count is established and activated.

In the case of a top pattern, the exit column must fall below the lowest 0 in the pattern so that there is a clear row running between the entry and exit columns.

The logic of the horizontal count

To understand the logic of the horizontal count, you must consider why congestion areas develop. They are a summary of the battle for supremacy by the bulls and the bears, neither wishing to progress too far. It is like a pressure cooker. When the lid blows, the contents are expelled with force. The distance it travels will depend on the pressure built up inside. It is similar with congestion patterns in Point and Figure charts. The longer the battle for supremacy, in other words the more columns built up by opposing sides taking contrary positions, the stronger the resultant move when one side is overcome.

236

33�

Chart 4-9: 5 x 3 of 8arclays pic showing vertical and horizontal counts

Targets have no time-scale

Perhaps the hardest thing for non-Point and Figure chartists to come to terms with, is that having been given a target, there is no time-scale for its achievement. Even an experienced Point and Figure analyst struggles with this. The fact is that Point and Figure charts have no time-scale, so it is impossible to calculate when a target – vertical or horizontal – will be reached. It is futile to attempt to do so. Simply accept that it is the situation.

Chapter 4 – Projecting Price Targets

Things you should know about Point and Figure

counts

Chart 4-9 below is a 5 x 3 Point and Figure chart of Barclays pIc, showing a number of counts. Don’t concern yourself by the large numberof them, some are invalid. The chart is referred to in the following paragraphs to illustrate a number of points about vertical and horizontal counts. Although the chart below is a 3-box reversal chart, the points made in the following paragraphs apply equally to counts on I -box charts.

eoo

237

The Definitive Guide to Point and Figure

Nearest counts must be achieved first

Those of a bullish disposition will be tempted to seek out the highest possible count just as those of a bearish disposition will seek the lowest target. Before doing so, consider the logic. On the upside, lower targets have to be achieved before a higher count can be considered. If a lower target is not achieved, then the achievement of the higher count is obviously impossible. Counting is a progressive tool. By all means, select counts that cluster around a target, but do not select one that is double the next target. Its presence on the chart will give you a distorted view of the future. In Chart 4-9, count 1 5 of 670 may be established to show the strength of the chart, but should not be relied on or even quoted until counts 1 1 of 605 and 1 6 of 625 have been achieved. It is pointless relying on count 1 4 until vertical counts 1 2 and 1 3 have been achieved. In fact, they never were.

Clustering of counts

You have already been advised to avoid trying to count every column; however, there will be times when a number of valid counts, both vertical and horizontal, may be obtained from the same area of the chart. Normally when there is a horizontal colint, there will be a vertical count from the breakout column. Any clustering of these multiple targets reinforces the likelihood of that particular target being achieved. Vertical count 9 of 540 and horizontal count 10 of555 in Chart 4-9 show clustering, as do vertical counts 1, 3 and 5 to the downside. Counts 1 1 and 1 6 of 605 and 625 do as well. Often, three separate targets come in within a few percent of one another.

Remember, however, that although counts may be clustered, this does not mean that they will definitely be achieved. Counts 6 and 7, and 12 and 13, are good examples. Clustering of counts means that if the price does look like it is heading in that direction, then it is likely to stop somewhere within the clustering area. Clustering does not assure the count, it simply increases the likelihood that a target within the cluster area will be achieved.

Negating a count

Not all counts will be achieved, so there has to be a process for removing them from your analysis. A vertical upside count is negated when the price falls below the low that anchored the count. Conversely, a downside count is negated when the price rises above the top that anchored it.

Count 2 of 575 in Chart 4-9 is negated when the price falls through point A, opposite the bottom, or anchor, ofthe count. Count 2 should, therefore, be removed from the chart. Count 7 is negated at point B . Count 6 is negated at point C . Counts 1 2 and 1 3 are negated at points D and E respectively. Once a count is negated, it should be removed to avoid confusion.

238

Chapter 4 – Projecting Price Targets

Horizontal counts are negated in the same way. A downside horizontal count is negated when the price rises above the top of the pattern encompassing the count. An upside horizontal count is negated when the price falls below the bottom of the pattern. Horizontal downside count 14 is negated when the price rises above the level at point E.

Negation of a count shows weakness in the direction of the count. It shows that the determined bulls or bears that caused the count to be established and activated in the first place are not strong enough to follow it through. Understanding the negation of counts is closely related to the ideas in the following section on opposing counts.

Opposing counts

It is often the case that you will have an upside and a downside count working against one another. Do not feel that your analysis is weak. Downside count 1 is established from the top. Thereafter, the price falls and consolidates around the 450 level. Eventually it makes a bottom from where upside count 2 can be established and activated. The chart is now under the influence of two counts. Although count 1 is the most important, count 2 must be taken note of as it may be achieved if the trend has changed. The discussion on trend lines in the next section will help you to assess the likelihood of count 2 being achieved.

Once counts 1 , 3 and 5 have been achieved, the chart comes under the influence of counts 6 and 7 . These cluster around the 200 to 2 1 0 area and look distinctly possible, especially after count 8 is negated. Then another bottom is made and count 9 is activated. The chart is now under the influence of bearish counts 6 and 7, but, at the same time, there is an upside count 9 pulling in the other direction. Knowing when one side is pulling stronger than the other then becomes important. Initially counts 6 and 7 are the strongest, but as the bottom develops and the move from count 9 is extended, so count 7 is negated, which is a weakening of the bear side. It is a downside count that has not been achieved and has had to be cancelled; count 6 remains dominant though. Further development at the bottom results in the establishment of count 1 0, which is activated at the same time as the bearish resistance line is broken. Counts 9 and 1 0 now become the more likely targets, with count 6 looking doubtful. Remember that count 6 remains valid, however, so there are still valid upside and downside counts working against one another. This remains the case until count 6 is negated at point C.

Combining counts with trend lines

The validity of counts is enhanced by the use of trend lines, especially 45° trend lines. An upside count is more likely to be achieved if it occurs when the count column is above a 45° bullish support line. A downside count is more likely to be achieved if it occurs below a 45° bearish resistance line. Counts established against the prevailing trend should always be viewed with suspicion. They can, however, be useful because their achievement or non­ achievement does explain more about the underlying nature of the trend in place. Count 2 of

239

The Definitive Guide to Point and Figure

575 is below the bearish resistance line and so should be considered a speculative target, unless the chart pattern changes sufficiently to indicate that a major bottom has been made. What does happen, in fact, is a weak rally that reverses well short of the count, thus reinforcing the bearish nature of the trend. Further reinforcement of the bearish trend occurs when count 2 is eventually negated.

A similar situation occurs with count 8. A possible bottom is made around the cluster target of 320/335/340, generated from counts 1, 3 and 5. Count 8 is established and then triggered, but what occurs is a weak rally, which fails to follow-through, and the count is eventually negated, reinforcing the downtrend. At this stage, counts 6 and 7 become the most likely downside targets. See the next section on unfulfilled counts for more information about this situation.

Some countertrend counts are achieved, but these also provide more information about the chart. A countertrend target which falls short of the prevailing downtrend line, but which is then achieved, reinforces the prevailing trend. It means that there is sufficient power for a countertrend rally, but not enough to challenge the downtrend. During any trend, there will be countertrend rallies, or countertrend corrections. Iftargets from these fall short ofthe trend line, and are achieved, it reinforces the strength of that trend line. Count 4 is a countertrend count, which has a target short of the bearish resistance line. The fact that the count was achieved indicates that the bear trend is more powerful than the bottom from which the count 4 rally occurred.

Unfulfilled counts

The previous sections hinted at this. By definition, there will most likely be at least one upside and one downside count that will not be achieved on most charts. The reason is that as the trend matures, so new counts are established. The likelihood of a count near the end of a trend not being achieved is increased. It is not always the case, but, when it does happen, it is a clear indication that the trend is coming to an end.

Downside counts 6 and 7 in Chart 4-9 are examples of counts that were never achieved. At the time they were established, the downtrend in place was strong and there was no indication that these counts were not going to be achieved. Then a possible bottom was made and count 9 established first, followed by count 10. At the same time, count 7 was negated. Count 10 became active at the same time that the bearish resistance line was broken. The conclusion is that it is now highly unlikely that downside count 6 will be achieved. This non-achievement is further evidence that the bear trend is over.

Non-achievement is also common in uptrends, as count 1 7 on the far left of the chart shows. It is a valid upside count, which was not achieved. You would not know this at the time, however, but the establishment and activation of count 1 , cancelling count 1 7, confirms that a new bear trend is being established.

240

Improbable and impossible counts

Always be on the look out for counts that are impossible or improbable. Impossible counts are easier to spot because they give a figure that is impossible. This can only happen with downside counts and occurs when the downside count yields a value less than O. Most software makes the target O. Ofcourse, a count ofzero could mean that the company is going out of business but it is more likely that the column you took the count from is invalid.

Chart 4-10 shows a 2 x 3 ofAlizyme pic. The count from the top produces a target ofzero. Of course, there is a possibility that the company could go out of business, but what is more likely is that your box size is too big or the column cannot be used for the count. Often, if this happens, the next column can be used, in this case yielding a target of 22, which was achieved almost exactly.

up.data Technical Analyst ‘ 280

260

240

220

200

180

160

1 40

120

100

80

60

40

20

AlIZVME ORO 2P (AZM)
AZMDol Point&FIIIr.cl2x3

Chapter 4 – Projecting Price Targets

Upd�J rtchnlClJI A’lalyn

22

Chart 4-1 0: 2 x 3 of Alizyme pic showing impossible and improbable counts

241

The Definitive Guide to Point and Figure

Improbable counts are another matter. The probability that a count will not be reached is a subjective interpretation and should, therefore, be treated with caution. Many improbable counts have been achieved. Xerox Corp. in Chart 4-11 is a good example, using a 0.5 x 3 chart. When the price was around 50, it broke down out of a large fulcrum top allowing a horizontal downside target to be calculated. The target came out as 3 . Anyone looking at the chart at the time, with the price at 50, may well have thought that this was an improbable count. In fact, within 1 8 months it had fallen to 4.5, only 1 .5 points from that target.

You will not know at the time whether the count will be achieved, but you must look at the chart for signs. The signs are that a large top has been made at the same time an important 45° trend line has been broken. The top pattern yields a horizontal target of3, the breakout column yields a vertical target of 15.5 (not shown). Both are so far from the current price, that it is difficult to justify remaining long. It doesn’t matter whether they are achieved with any accuracy. The point is that they are less than half the current price. Once you have no long positions, it is much easier to follow the price down and make decisions during the downtrend about where the price will stop. The bottom pattern between 20 and 25 looked like it would halt the downtrend but that quickly faded and, when it did, the target of 15.5 and then 3 became more likely.

242

The Definitive Guide to Point and Figure

S&P 500 INO£ K (S500)

S5OO 0

cl) 10 x 3

P_&f •

10

1 500

1400

1300

700

Chart 4-1 2: 1 0 x 3 of the S&P 500 Index showing vertical and horizontal counts

Chart 4- 1 2 shows a l O x 3 of the S&P 500 Index. The first bottom in October 2002 yields a vertical count of 1 200, which, at the time of writing, had just been achieved. A further count off the March 2003 bottom yields 1 050, which was also achieved. Only three other counts are possible: the large horizontal count of 1320 from the bottom; a vertical count of 1330 from the breakout column; and the minor vertical count of 1280 from the mini-bottom, three columns from the end. Note, however, that no count targeted the important resistance at 1 1 50.

Chart 4-13 is a 5 x 3 ofthe S&P 500 Index, halfthe size ofthe box in Chart 4-12. Reducing the box size to 5 has exposed more detail and many more counts can be established. The two bottoms yielded almost identical counts of 1080 and 1075, and then two mini-bottoms during the uptrend yielded two identical counts of 1150, which is where major resistance was encountered.

244

Chart 4-1 3: 5 x 3 of the S&P 500 Index showing vertical and horizontal counts

10

Chapter 4 – Projecting Price Targets

e cl)5)(3

up.!:!1′-ta Technical Analyst 15

850

800

It is always worth looking at counts on various box size charts, as you may see something not apparent on the box size chart you normally use.

A short-term trader will use intra-day charts of I-box and 3-box reversal to give intra-day counts. Chart 4-14 shows a 1 x 3 tick chart of the FTSE 100 Index covering 28th January 2005 to 15th February 2005. Notice how, as the price rises in a strong uptrend, short-term vertical (marked with V) and horizontal (marked with H) counts are established, activated and achieved. Notice too that any downside counts (not shown) are not achieved.

245

matching of column lengths that is the first thing to notice. If this happens, it means the price is not trending in a true Point and Figure fashion. It means that the bulls take hold and the bears take it back and vice versa. Point and Figure relies on a move from the one camp, a weakish response from the latter and a reassertion by the former. Without this, some of the value of the Point and Figure charts is lost and part of that loss is the accuracy of targets from the chart. Instruments that have traded sideways over a long time tend to be bad counters as Chart 4-15 of Bradford and Bingley pIc shows. A selection of past counts indicates that targets are unlikely to be met and should therefore be treated with caution.

Chapter 4 – Projecting Price Targets

1I0l&_012SP(11

116 110
Chart 4-1 5: 3 x 3 of Bradford and Bingley pic showing a example of a bad counting chart

Counts on close or high/low charts

38

390

360

330

300

210

240

There is no evidence to suggest that counts are more accurate on close or high/low charts. Chart 4-9 of Barclays pIc discussed on 237 is a close only chart. Compare this with Chart 4-16 overleaf, which is the same 5 x 3 box and reversal, but is constructed with high/low data. All relevant counts have been placed on the chart. You can see some are the same, some are very different. Because high/low charts have wider congestion patterns there are more horizontal counts. Using high/low data increases the volatility ofthe chart, so there will be less longer-term vertical counts because the columns will be shorter. Some readers will want to know whether it is best to use close or high/low charts. Unfortunately, it is not possible to say. Some instruments will perform better with close, some better with high/low. It is the same problem with deciding on log or arithmetic. It is best to apply the rule in the previous paragraph. Draw both the close and the high/low charts, perform the counts and see which type of chart has worked best in the past. Ifyou feel uncomfortable about either ofthem, do not use the chart.

247

The Definitive Guide to Point and Figure

Chart 4-16: 5 x 3 daily (hi) of Barclays pic showing counts on a high/low constructed chart

Counts on other box reversal charts

Counting tends to be conducted on I-box and 3-box charts only. It is rare to see them used on 5-box charts because the counts tend to be too long-term and often transpire to be completely wrong. The columns on 5-box charts are very long because the price has to reverse by 5 boxes to change columns. Vertical counts, therefore, yield improbable counts and are best ignored. You will recall that near-term counts must be achieved before the far­ term counts can be considered. With 5-box charts, you will not obtain any near-term counts. Horizontal counts tend to be better because the 5-box reversal makes congestion patterns narrower, and this makes up for the fact that the width must be multiplied by 5.

The vertical count formula does not work very well on 2-box charts either. Vertical counts always fall short of the correct target, because the column is only projected by twice its length instead of 3 times. As with 5-box charts, however, horizontal counts can be used on 2-box charts and do produce reasonable results. This is because there are more columns and so the width of the congestion patterns is greater. As explained already, it is best to try a number of earlier counts to see whether the counts have been accurately achieved in the past.

248

Chapter 4 – Projecting Price Targets

Counts on log scale charts

As if it weren’t enough to contend with vertical and horizontal counts, and then two methods of conducting horizontal counts, mention also needs to be made of counts on log scale Point and Figure charts. A log scale Point and Figure chart, as you have seen, is a chart where the box size is set to a percentage of the price and, therefore, varies exponentially as the price rises. This means that you cannot use an arithmetic calculation method. Whilst it is inadvisable to draw log scale Point and Figure charts by hand, and therefore calculate the counts by hand, there will be those who wish to do this.

Consider what is important when calculating a count. The three parameters you use to calculate the target are the box size, the reversal and the number of boxes in the column or across the pattern, which are multiplied together and added to, or subtracted from, a base price to give the target. In the case of log scale Point and Figure charts, however, the box is a percentage change in price. It means that every X in a column of Xs is slightly larger, by the percentage, than the previous. Ifthe box size is 1 % then each box will be 1 .0 1 times larger than the previous one as the price rises. So, a 3-box reversal would not simply be formed by a 3% move, but by a 3.03% up or a 2.97% down.

To obtain a target on a log scale chart, therefore, you must proceed as follows:

Count the number of boxes in the vertical column in the case of a vertical count, or across a pattern in the case of a horizontal count.

Multiply the number of boxes by the reversal and by the natural logl5 (In) of the box size. This gives the extension, which is either added to or subtracted from the anchor point.

Take the In of the anchor point (the high or low to which the count is anchored).

For upside counts, add the extension to the In of the anchor point. For downside counts subtract the extension from the log of the anchor point.

Finally, take the exponential (e ) of this figure to give the target.

Chart 4- 1 7 overleaf is a 1 % x 3 Point and Figure of the S&P 500 Index. A number of counts have been shown.

I S The natural log (In) is the preferred method of logging data.

249

Horizontal count C

There are 20 columns in the pattern.

The extension is In(l.OI) x 20 x 3 = 0.597020

The lowest 0 in the pattern is 778.03.

The target is the exponential of In(778.03) + 0.597467 which is e(6.656765+0.597467) = 1413.44

Accuracy of counts on log scale charts

Log scale counts tend to overstate both the downside and upside targets as you can see by comparing log scale Chart 4- 1 7 and arithmetic Chart 4- 1 8 .

Chapter 4 – Projecting Price Targets

950

1500

1400

1300

Chart 4-1 8: 1 0 x 3 of the S&P 500 Index showing counts on an arithmetic chart

The difference is normally so small as not to make much difference. The reason is that the extensions are compounded by 1% per box. You should understand that you will never estimate the same targets or, in fact, retracements if you change from arithmetic to log and vice versa. Every aspect ofthe chart is different, including trend lines. This applies not only to Point and Figure, but all other chart types as well.

251

The Definitive Guide to Point and Figure

This should not prevent you from using one method over the other. In any case, it’s not really a true test, because it is impossible to draw an arithmetic and a log scaled chart which are the same. In this example, one is 10 points and the other is 1%. There will only be one place on the chart, at the 1000 level, where the box size is identical. So, accept that counts on log scale Point and Figure charts will be different. The suggestion made earlier applies: try a few counts on past data and see how accurate they have been.

De Villiers and Taylor 3-box horizontal counts

The De Villiers and Taylor method of 3-box horizontal counts was devised by Thomas Sexsmith. It is not really mentioned any more, but that does not mean that you can’t use it. Unlike the Cohen method, it is very effective for counting targets from continuation patterns.

It uses the same logic as the I-box horizontal count discussed on page 207, the only difference being that the total is multiplied by 3 before adding it to the row at which the count is taken. So in a 10 x 3 chart, if the counting row had 12 squares, the total would be the number of squares times the box size times the reversal, 1 2 x l O x 3 . It seems that this method of 3-box horizontal counting fell into disuse once Cohen’s book was published. Its use was discontinued because 3-box patterns are deeper and so it was quite time consuming to find the row with the most filled in XS and Os.

Risk and reward

There are two questions that every trader must ask and then answer before placing a trade: ‘What do I do if! am wrong?’ and ‘How do I know I am wrong?’ The first is easy to answer: you must close your position. Many people are unable to answer the second question, and if it remains unanswered, the first becomes redundant. At the time you make the decision to trade, you must look to see what would make you close that trade. Some traders use a simple trailing stop loss, but Point and Figure charts allow you to be more specific, based on signals you can see in the charts – normally a double-bottom (top) or a breakout from the other side of a compound pattern. Once you have decided what would ‘cancel’ your position, resulting in a loss, you know the risk of placing the trade. That is one side of the equation, because there is a potential reward as well and estimating this is an important part of the process of making the decision to trade. This is where Point and Figure counts, discussed earlier in this chapter, come in.

252

In fact, combining the risk and the reward creates the misnamed1 6 risk-reward ratio. Point and Figure charts are ideally suited to computing these ratios because there is always a clear-cut entry and clear-cut exit point. For this reason, the risk-reward ratio is more suited to 3-box reversal charts rather than I -box, because in I -box charts the signals, although effective, are not quite so clear-cut.

The risk-reward ratio is the ratio ofthe potential gain from any trade, derived from the Point and Figure count, to the possible loss if the trade goes wrong and the price goes in the opposite direction, derived from Point and Figure double-top and double-bottom signals.

Risk-reward ratio from vertical counts on 3-box charts

You will recall that there are two stages to the vertical count – the establishment and the activation. The risk-reward ratio can only be computed once the count has been established and the activation column is in the process of being built.

Figure 4-6 shows a typical bottom. After the first column of XS in column 2 off the low, the establishment stage takes place when the length of the column of XS is fixed by the reversal of a column of Os in column 3 . At this stage the vertical target can be established and the reward calculated. The reward is the difference between the vertical target and the price at the breakout above the highest X in the counting column of Xs.

The length ofthe correction column of Os in column 3 is fixed by a reversal of a column of XS in column 4. At this stage, the risk may be calculated. The risk is the difference between the price at the breakout above the highest X in the counting column – above the blue line – and the value of the 0 below the correction column of Os – below the red line. The calculation is as follows:

The vertical target is the number of Xs in column 2 multiplied by the box size, multiplied by the reversal, added to the low in column 1 .

Verticaltarget= (7x 1 x3)+20= 41
The reward is the vertical target minus the price at the double-top breakout above the blue

line, marked A.

Reward= 41-28= 13

The risk is the price at the double-top breakout minus the price at which the first double­ bottom sell would appear on the chart if the price went against the trade – below the red line in Figure 4-6, marked B .

Risk1= 28-22= 6
Risk-reward ratio 1 = Reward/Risk = 1 3/6 = 2 . 1 7

1 6 The ratio is called risk-reward, but in fact it is actually reward divided by risk.

Chapter 4 – Projecting Price Targets

253

The Definitive Guide to Point and Figure

This means that the potential reward for every actual point of risk is 2. 1 7 points. Although it is a matter of personal preference, governed by your time horizon, a good risk-reward ratio is around 3 or greater. Short-term traders will accept lower risk-reward ratios that are sufficiently greater than, say, 1.0 to 1.5

Figure 4-6: Calculating risk-reward ratios from 3-box vertical counts

You can see that the risk is governed by the length ofthe corrective column of Os (in column 3) after the initial X column off the low (in column 2). The longer the corrective column of Os, the greater the risk of taking the trade at the breakout.

Although the reward is fixed, the risk may be adjusted by looking for support levels lower down in the pattern. Often two or three risk-reward ratios may be calculated from the same count. For example, risk could be established at the box below the lowest 0 in the pattern, at 1 9, marked C in the example in Figure 4-6.

254

The calculation is as follows:

Reward remains the same at 1 3 .

Risk 2 = 28 – 19 = 9

Risk-reward ratio 2 = 1 3/9 = 1 .44

Risk-reward ratio from horizontal counts on 3-box charts

Not as common, and not as easy to calculate, are risk-reward ratios from horizontal counts. The problem with horizontal counts is not the reward but the risk, because the exit from the pattern can be quite complex. Although it is customary to place the stop below the lowest low in the pattern, this results in increased risk levels. It is, therefore, best to study the pattern and determine subjectively where your stop should be placed.

Figure 4-7 is a typical bottom pattern from which a horizontal count may be established.

Risk level 1 is the price at which the first double-bottom sell would appear on the chart if the price went against the trade (below the red line in Figure 4-7, marked B), whereas risk level 2 is the row below the low ofthe pattern, marked C.

Chapter 4 – Projecting Price Targets

H C = · 43
… t

0 0 0

“0… co

Q)� n::

25 0 X…t

A

+i …

N

� II)

0XXX 0X0X X0X �II) 0X0X0X0X
0 0X0X0

+�
20 0X0X … B

�
… C

23456789
Figure 4-7: Calculating risk-reward ratios from 3-box horizontal counts

00

255

The Definitive Guide to Point and Figure

The horizontal target= (8 x 1 x 3) + 19 = 43 The breakout row is at 25, marked A.
The reward = 43 – 25 = 18
Risk1= 25-20= 5

Risk2= 25-18= 7

Risk-reward ratio 1 = 18/5 = 3.60

Risk-reward ratio 2 = 1 817 = 2.57

You therefore have two risk-reward ratios to work with and may allocate your trade accordingly.

Chart 4- 1 9 is a l O x 3 Point and Figure ofWhitbread pic. The horizontal count across the base pattern yields a target of 1 320. The level at the breakout is 690. The reward is, therefore, 630 points (1320-690).

There are two possible risk levels. The level for establishing risk 1 is the price at which the first double-bottom sell would appear on the chart if the price went against the trade; at 520 on the chart risk level 2 is the row below the low ofthe right-hand side of the pattern, at 450. In Chart 4- 1 9 you will notice that there is a lower risk level below the low of the whole pattern at 410; however, the use ofrisk 2 negates the need for using the risk 3 level, which is why it is not shown. Normally, however, the low of the whole pattern would be the greatest risk level.

256

The Definitive Guide to Point and Figure

WHlTIIR£ADORO50P(WTB)
WTB DolPoint! & Ftauro cl IO . 3

80

900

BOO

700

H

Risk 1

��
600

� �8 � Activation Point

� >�

-_ Rlsk2

Chart 4-20: 1 0 x 3 of Whitbread pic showing risk-reward ratios from 3-box vertical counts

Often there is a vertical count from the same pattern. The horizontal target shown in Chart 4- 19 was 1320, whereas the vertical target in Chart 4-20 is only 820. The activation point is at 590, making the reward 230 points.

There are two risk levels at 520 and 450 making risk 1 , 70 and risk 2, 140. Risk-reward ratio 1 = 230/70 = 3 .29
Risk-reward ratio 2 = 2301140 = 1.64

Risk-reward ratio 1 is confirming the ratios from the horizontal count, but risk-reward ratio 2 is on the low side. This means that a stop below the low of the pattern is unacceptable for a target based on the vertical count only, but, since the horizontal count yields a greater target and hence a more favourable risk-reward ratio, the trade can be considered on the basis of a stop below the low of the pattern.

258

500

Chapter 4 – Projecting Price Targets

Risk-reward ratio from horizontal counts on 1 -box charts

As you will have realised by now, I -box charts are more subjective than 3-box charts, so the placement of counts and stops is not as clear-cut and, consequently, risk-reward ratios are more difficult to obtain. However, if you follow some basic rules they are possible.

You will recall that there are a number of ways to conduct a horizontal count. When calculating risk-reward ratios, it is the count across the row which acts as the balancing point of the pattern that is important. This is the row with the most filled in boxes. A break below this level after the pattern has broken out and given a buy signal, is your level of risk. The second level ofrisk is the row below the low ofthe pattern. The reward is the horizontal count target minus the value at the breakout of the pattern.

Figure 4-8 is a typical fulcrum pattern from which a horizontal count may be established.

Figure 4-8: Calculating risk-reward ratios from 1 -box horizontal counts from fulcrums

Thehorizontaltargetacrosstherowwithmostfilledinboxes= (11xIxI)+52= 63

The catapult point is at 5 5 .

Thereward= 63-55= 8

There are two levels of risk. Risk I is the row below the pivot row: the row with most filled in boxes. Risk 2 is the row one box below the low ofthe whole pattern.

RiskI= 55-51= 4 Risk2= 55-48= 7

259

The Definitive Guide to Point and Figure

Risk-reward ratio 1 = 8/4 = 2.00

Risk-reward ratio 2 = 8/7 = 1.14

Risk-reward ratio 1 i s acceptable, but i fyou intend taking the trade based on a stop below the low of the pattern it is not, based on risk-reward ratio 2. Stops below the low of the pattern in I -box charts are usually too far away.

Figure 4-9 is a typical I -box continuation semi-catapult.

1 200

Figure 4-9: Calculating risk-reward ratios from 1-box horizontal counts from semi-catapults

The horizontal target across the row with most filled in boxes = (6 x 10 x 1) + 1220 = 1280

The catapult point is at 1 240.

Thereward= 1280-1240= 40

Again, there are two levels of risk: one below the pivot row and one below the low of the semi-catapult.

Risk1= 1240-1210= 30 Risk2= 1240-1200= 40 Risk-reward ratio 1 = 40/30 = 1 .33 Risk-reward ratio 2 = 40/40 = 1 .00

The risk-reward ratios from semi-catapults are characteristically low, but remember that these are continuation patterns where, unlike reversal fulcrums, the likelihood of a reversal is low. Consequently, a low risk-reward ratio might be considered acceptable. It would not be if the pattern were a reversal pattern.

1250 I� X�

II
Catapult point

I

XX XX� X00XX0 I
X 00 …
X Risk2 X

X

Count Row R i s kI 1

260

Chart 4-2 1 is a 20 x 1 Point and Figure of Whitbread pic. It shows a horizontal count of 1 1 60 out of fulcrum pattern A and 960 from continuation semi-catapult B.

The catapult point offulcrum A is 700. The reward, is therefore, 460 points (1160-700). Risk 1 is the row below the counting row at 540 and risk 2 is the row below the low of the

pattern at 400, making the two risks 1 60 and 300, respectively.

Risk-reward ratio 1 = 4601160 = 2.88 Risk-reward ratio 2 = 460/300 = 1 .53

WHITBREAD ORD sOP (WTB) upc:la!a Technical Analyst =� . � � �; � . � � i �nt . ,. � � f i u�r . c 1 2 � ‘ . � _ – – – – – – – – – – –l 1 0 0 0

�9 0

Chapter 4 – Projecting Price Targets

:CP R1 BOO

_R2

�—-� – — ‘- Catapult point _. _ _ _ _ _ _ _ R i s k 2

Chart 4-2 1 : 20 x 1 of Whitbread pic showing risk-reward ratios from 1 -box horizontal counts

The catapult point of semi-catapult B is 860. The reward is, therefore, 1 00 points (960-860). Risk 1 is the row below the counting row at 800 and risk 2 is the row below the low of the

pattern at 760, making the two risks 60 and 100, respectively.

Risk-reward ratio 1 = 100/60 = 1.67 Risk-reward ratio 2 = 100/100 = 1.00

600

400

261

The Definitive Guide to Point and Figure

Risk-reward ratios on shorts

All the risk-reward ratios shown have assumed long trades. You may establish them in exactly the same way for downside counts and short trades.

Risk-reward ratios in aiding the placement of stops

The placement ofstops is vital for good trading, making the calculation ofrisk-reward ratios useful in determining where to place your stops. Having calculated the various risk-reward ratios from a pattern based on a number of different stop levels, you may then look at the risk­ reward ratios achieved and place your stop according to the risk-reward ratio you feel comfortable with.

Finally

Risk-reward ratios are calculated to help you decide whether or not to take the trade. The reward is established from the count, the risk from where you will close the trade, should it go wrong. Once the trade has been entered, however, the ratio itself is no longer of interest. Of course, the count still applies and the risk remains the same until a new stop can be placed at a higher level, but knowing the risk-reward ratio does not affect the way you manage the trade. What is important is that if the decision to trade has been based on a particular risk­ reward ratio, then the level ofthat risk must be the level at which you close a bad trade. It is foolish to assess a trade based on a risk-reward ratio and then ignore the signal to close a trade that goes wrong.

From one target there is one level of reward, but there may be a number of levels of risk. In the examples above, risk-reward ratio 1 is calculated using the risk from the first double­ bottom sell signal. You may decide that this will not be your exit point and, ifthat is the case, you should not use risk-reward ratio 1 to determine your level of risk. Very often, the first double-bottom sell is too close to have your stop. Looking left in the pattern will tell you whether there is a better level lower down to use.

In practice, risk-reward ratios are just a guide, but it is good practice to always calculate them. Many traders will use trailing stops to protect an open position, but very few, it seems, have a reliable method for arriving at the reward, and hence the risk-reward ratio. This is yet another advantage of Point and Figure charts. Of course, risk-reward ratios may be arrived at in a number of ways. You may use the Point and Figure count and combine that with a money management stop rather than a Point and Figure sell signal.

262

Other ways of projecting targets

Although the Point and Figure counts discussed so far are unique to Point and Figure charting, they are not the only way to project targets. It may seem out of place to have a section on Fibonacci17 retracements in a chapter on counts but it is not: the two are inextricably linked. Prices often retrace previous moves by key Fibonacci levels and often these levels match the targets from vertical and horizontal counts. The most common Fibonacci retracement levels are 23.6%, 38.2%, 50%, 61.8% and 78.6%.

Fibonacci retracements

Fibonacci retracements are used to establish the possible ‘bounce’ levels of a correction against a previous trend. A bounce level is the amount of the previous trend – up or down – that is retraced by the price during the correction. They provide possible support or resistance levels, which are calculated on Fibonacci ratios rather than observation of previous congestion areas, although they often coincide with these.

Chart 4-22 is the same 1 0 x 3 Point and Figure chart of Whitbread plc used earlier. Some of the Fibonacci retracement levels of the trend from A to B are shown. Once the top at B has been made, the vertical height between A and B is divided into the Fibonacci ratios or percentage retracement levels. For example, a retracement back to the level ofA is a 100% retracement. The price will not always retrace to every Fibonacci level. The levels are potential target areas – in this case, potential support levels. Notice that the 23.6% retracement level provides the initial support at point C. The 38.2% level provides no support at all and is, therefore, not shown on the chart. Do not assume that every Fibonacci retracement level will provide support, and do not assume that the price will stop exactly at every level. Fibonacci levels are target areas rather that exact target levels. The 50% level provides the strongest support at point D. Notice that this coincides with the previous bottom on the chart to the left. The price tests the 50% retracement level a second time at point E, before rising to point F, level with point B.

After reaching point F, the price falls, finding support again on the 50% retracement level at point G. It eventually falls through this level and finds support on the 78.6% level at point H. Notice that the 61.8% level (not shown) provides no support at all, but the 78.6% level is tested a second time at point I. You will often find that if one Fibonacci level has held, the next will not.

17 Leonardo of Pisa, more commonly known by his nickname, Fibonacci, a 13th century mathematician, discovered a number series, 1 , 1 , 2, 3, 5, 8 etc., where the next number in the series is the sum of the previous two. What is important about these numbers is that eventually the ratio between any two consecutive numbers is the golden ratio, 0.618. This ratio, as well as many derivations of it, are widely used in Technical Analysis.

Chapter 4 – Projecting Price Targets

263

Chapter 4 – Projecting Price Targets

Chart 4-23 shows the retracements of the trend from B to D. The 23.6% level provides resistance to the uptrend at point a. The price then rises to encounter resistance from the 38.2% level at point b. The 50% level at point c provides the strongest resistance, but the 61.8% also provides enough resistance to cause the price to pause at point d. The price then rises to point F, which is almost 1 00% retracement of the B to D trend. Notice that, in this case, the 78.6% level (not shown) provided no resistance at all.

Wl Tll AO ORO SOP (WT B) wmo ‘ Point&F’ e(c010x3

100.0% 1100 1000

O- ———–236% 800 _- ___0.0% 700

Chart 4-23: 1 0 x 3 of Whitbread pic showing Fibonacci retracement levels

265

The Definitive Guide to Point and Figure

Chart 4-24 shows the retracements ofthe trend from E to F. The 23.6%, 38.2% and 50% levelsprovidenosupportatall.The61.8% leveldoeshoweverprovidesupportatpointe,as does the 100% retracement level at point f. The price then falls to point g, which is a retracement of 161.8%.

Chart 4-24: 1 0 x 3 of Whitbread pic showing Fibonacci retracement levels

266

Chapter 4 – Projecting Price Targets

Chart 4-25 shows the retracements of the trend from F to I. The 23.6% level (not shown) provides no resistance, but the 38.2% does at point a, as does the 50% at point b and 61.8% at point c. Notice that the target of870 from the vertical count offthe low at point I matches the 6 1 .8% level. 78.6% is the next Fibonacci level that has not been reached.

�.

Chart 4-25: 1 0 x 3 of Whitbread pic showing Fibonacci retracement levels

As with counts, retracement levels are target areas, which may or may not provide support or resistance to the price. It is customary to draw all the levels but then remove those which have had no effect, as has been done in the charts above.

267

The Definitive Guide to Point and Figure

Summary

You have seen the benefits and the pitfalls o f counting. You have also seen, how counts help with the general analysis of the chart. It is almost impossible to conduct Point and Figure analysis without establishing counts, just as it is impossible without trend lines. Counts and trend lines go hand in hand and either confirm or contradict one another. Confirmation gives confidence to your analysis. Contradiction casts doubt, which requires further investigation and a search for more evidence.

Counts assist in determining the strength, and the possible decay, of trends. This is because, in bull trends, upside counts are usually exceeded, and, in bear trends, downside counts are usually exceeded. In fact, the degree to which they are exceeded goes some way to explaining the underlying strength of the bull or bear trend. Conversely, in a bull trend downside counts are not usually achieved and in bear trends upside counts are not usually achieved. A downside count which is not achieved during a bull trend reinforces the trend, as does an upside count which is exceeded.

The corollary of this is perhaps more important. In any bull trend, there will always be one upside count that is not achieved. This is normally the last count undertaken in the trend, but does not have to be. A count not being achieved is evidence that a bull trend is coming to an end, but you will not know that this is the case until you see a top pattern taking place. The degree or size of the top pattern is dependent on the size of the trend that precedes it. When an upside count is not achieved, it indicates a decay in the bull trend and provides evidence of an end to the trend.

The same occurs in bear trends. There will always be one downside count that is not achieved, normally the last count conducted in the trend, which gives evidence for the end of the bear trend.

As you have seen, you will often have counts working against one another. You will have the upside count that has not been achieved working against a new downside count, or a downside count that has not been achieved working against a new upside count. Do not concern yourselfwith this; it is quite normal when a trend is changing. The count that ‘wins’ defines and reinforces the change in trend. At some stage, however, a count will be negated or cancelled as explained on page 238.

To summarise, therefore:

Counting yields potential price targets.

There are two counting methods, vertical and horizontal.
I -box reversal charts allow horizontal counts only.
3-box reversal charts allow vertical and horizontal counts.
The horizontal counting method is different on I -box and 3-box charts.

268

•

There can be upside and downside counts working against one another.

There is no time-scale for the achievement of a count.

There is no rule as to whether counts are better on close only or high/low charts.

Vertical upside counts are negated when the price breaks below the bottom of the counting column.

Vertical downside counts are negated when the price breaks above the top of the counting column.

Horizontal upside counts are negated when the price breaks down through the low ofthe pattern.

Horizontal downside counts are negated when the price breaks up through the high of the pattern.

In bull trends, upside targets tend to be exceeded and downside counts are not achieved.

•

•

•

•

In bear markets, downside targets tend to be exceeded and upside counts are not achieved.

The achievement or non-achievement of counts gives clues as to the state of the trend. Counts should be combined with trend line analysis.
Counts are approximate and a guideline. The targets given are potential targets.
The more counts that coincide or cluster at a particular target, the stronger the target is. Counts give unambiguous targets. They may not be right but they are unambiguous. Counts work well on some charts but not on others.

Counts reinforce support and resistance levels.

Don’t keep adding new counts to your chart just because you feel bullish or bearish. There is no point in finding a count that yields, say, 865 if the price has not yet achieved an earlier count of 620.

Counts on log scale Point and Figure charts are also possible. Their calculation leads to them overstating the targets achieved from arithmetic charts.

Counts, combined with standard Point and Figure sell signals, mean that risk-reward ratios may be determined each time a count is established.

Fibonacci retracements of a prior trend may also be used to establish targets.