Chapter 7 – Optimisation of Point and Figure Charts

Before discussing the optimisation of Point and Figure charts, it is worth discussing optimisation in general, as it is a subject that requires clarification.

The case for and against optimisation

Optimisation of technical indicators and chart signals has its devotees and its detractors. Optimisation in technical analysis is the testing of various entry and exit conditions, as well as calculation parameters, in order to determine which combination has yielded the best results. The emphasis is on the past tense, but it is a wide remit. It can be helpful, but a hindrance as well. It is only helpful when the analyst undertaking optimisation understands the chart or indicator that is being optimised.

It is important, when considering optimisation, to look at it from a Technical Analyst’s, rather than a statistician’s, point ofview. In doing so, you need to stop and think for a moment what Technical Analysis is. Technical Analysis is the study of price through the use of charts. It is the study of the past in the belief that it can tell you something about the future. It is the understanding that patterns in price and indicator charts repeat. That is what chart reading is all about. Technical Analysts look for ‘things’ in the chart that have proved reliable in the past – such as shapes, patterns, indicator movement and so on – on the premise that they will occur in the future and will, therefore, assist the analyst in making decisions. If you look back and inspect a chart and notice that every time there is a particular pattern, the result is a significant price rise, or that every time there is another pattern type, there is a sharp fall, this is just mental optimisation – optimisation by inspection; looking at past history and applying what you have learnt to current data. Every time you look at a chart you are, in effect, mentally optimising it.

But optimisation goes further than that. Some charts can be altered by altering the parameters that are used to draw them. Point and Figure is an example of this. You have seen that the look of the chart can be completely altered by changing the box size and the reversal.

Optimisation takes a set of data and then works out what would have been the best way to have analysed it, with the benefit of hindsight. If those results are then applied to the same set of data you will have, by definition, fitted the curve. Some may argue that you should optimise one section of data, then test the results of the optimisation on a different section and finally apply it to a third section, but that would be said without an understanding of the market and market charts. Market data is not just a series of numbers. Technical Analysts understand that price data has ‘life’, and that there is a relationship between past data and future data because the data is created by human market participants. Whereas some will go to lengths to eliminate any sign of autocorrelation, and work on data that has had its ‘life’ removed, Technical Analysts yearn for autocorrelation.24 They want, and believe, that past prices have an influence on future prices because the same human beings who set the prices in the past will also be setting them in the future.

24 Autocorrelation describes a condition where data points in a time series are not independent of each other.
355

The Definitive Guide to Point and Figure

Technical Analysts understand the market and market charts. They understand that market charts are created by price data and that the price data is created by human beings who have human traits, such as following trends and behaviour pattern repetition. They understand that these human market participants are subject to human emotions, which affect the price and consequently the charts. If today’s price were independent of tomorrow’s, Technical Analysis would be worthless. Technicians realise that each instrument has different characteristics and so understanding the movement of one, does not mean that another is understood. They

understand that the characteristics of the price movement are different, depending on whether the price is in an uptrend, downtrend or sideways trend.

Whether you agree with optimisation or not, it is worth considering the side-benefits that may stem from an optimisation exercise. Optimisation is such a wide subject that it cannot be given full treatment here; only a dedicated text with thousands of tests and examples can do it any true justice. What follows are simply guidelines for you to follow, should you wish to embark on an optimisation exercise of Point and Figure charts.

356

Approaching Point and Figure optimisation

When you draw a Point and Figure chart, you have to decide what box size and reversal you wish to use. After doing so, you may look at the chart and decide you have chosen incorrectly and change the parameters again. This is, in effect, mental optimisation through the process of trial and error. It is the way a skilled Point and Figure analyst operates.

While deciding on the best box size and reversal, you also have to decide what constitutes a buy and what constitutes a sell signal. Is it better to wait for a triple-top buy and a triple­ bottom sell? What if these never occur? These are the sorts of questions that mathematical

optimisation can answer. To do this, you have to decide what you want to achieve. The assumption is that profit is the motive, so you could take a section of data and then test a range of parameters and conditions which generate buy and sell signals to mathematically decide which set of signals gives the greatest profit.

The input parameters of any Point and Figure optimisation are:

Reversal size – 1, 2, 3, or 5 etc.

Box size – arithmetic points size or log percentage size.

Construction using end of period close or end of period high/low.

Entry signals – normal Point and Figure signals such as double-top, triple-top, catapult, and simple column change.

Exit signals – normal Point and Figure signals such as double-bottom, triple-bottom, catapult, and simple column change or other exit signals which are not Point and Figure related, such as stop loss.

Waiting period or signal delay.25

Data under consideration.

Dealing costs.

Long or short.

A full optimisation would have to take every combination of these parameters: run through the data to find the best box and reversal which yielded the greatest profit; find which entry and exit signals achieved this, at the same time working out whether the Point and Figure chart should be constructed with close or high/low data. Before plunging headlong into a complex optimisation using all these parameters, you need to stop and consider a few things. Ground rules have to be laid.

25 Signal delay is the waiting period after a signal has been generated to see if the signal remains in place. If after the waiting period, the signal is still in place, it is acted on. The purpose of the signal delay is to avoid whipsaws. The disadvantage is that it introduces a lag to the signal.

Chapter 7 – Optimisation of Point and Figure Charts

357

The Definitive Guide to Point and Figure

Entry and exit signals must be unambiguous, which is seen as a great advantage of 3-box reversalPointandFigurecharts. I-boxcharts,youwillrecall,aremuchmoresubjectivein their interpretation and their signals are not quite so clear-cut. For example, a double-top buy on a 3-box chart could be a complex semi-catapult on a I-box chart. The first ground rule, therefore, is that all optimisations should be calculated on 3-box reversal charts. Although 5- box reversal charts can also be used, they are too long-term to be of interest. Although double-top and bottom signals work with 2-box charts, it is felt that they also rely on bigger patterns and so will produce too many signals.

Having fixed the reversal at 3-box, the first optimisation, therefore, could be to test a range of box sizes, testing various entry and exit signals, but this also presents a problem. You will recall that ifyou use a points box size, there will be times when it is not suitable because the price has either risen or fallen significantly from the area where the box size was determined. This means that long-term optimisations cannot be conducted using a fixed points box size. This should be reserved for instruments with a history of a narrow trading range. Log scale percentage box sizes, on the other hand, adjust the box size at every price level. It is, therefore, advisable to use a range of percentage box sizes rather than points. The number of permutations is significantly reduced if the test has a fixed reversal size and the need to test arithmetic points box sizes is eliminated.

There is no way to tell whether close only or high/low data is best for the Point and Figure construction, so it is best to include the choice in the optimisation. What about signals? You will have learnt that triple-tops and bottoms are more important than double-tops and bottoms, but you must understand that there is a big difference between entry signals and exit signals. There is no problem waiting for a rare entry signal to occur. If it doesn’t occur, you do not enter the trade and no loss is incurred. That is not the case with exit signals. A rare exit signal that may never occur will keep you in an unprofitable trade without any chance of exiting. Exit signals must be guaranteed, which means that triple-bottom exit signals on longs, and triple-top exit signals on shorts, cannot be considered in the optimisation. The same applies to the powerful catapult pattern, which cannot be relied upon to provide an exit for a trade. It is the reason why trailing stop loss is included as an exit condition, because it is a guaranteed exit. So, more input parameters may be eliminated by thinking about the optimisation before entering into it. Exit conditions should be limited to double-bottom (or double-top for shorts), as well as a range of percentage trailing stop losses.

The next parameters to consider are signal delay and data. Some analysts and traders dismiss the idea of a signal delay out of hand, because, administratively speaking, it is difficult to implement. It is difficult to have the procedures in place to receive a signal one day but act perhaps two days later. As this widens the field of research, no account has been taken of signal delay in the optimisations conducted.

Data, too, provides a point for some discussion and argument. The idea that you should divide your data set into three equal sets, to optimise on the first, test on the second and apply the

358

conditions to the third, is unrealistic. It may work with trend-less statistical data, but it will not work with data from the markets. As discussed in the introduction to this chapter, market data has trend. What worked best in an uptrend may well not work best in a downtrend. It means that to optimise effectively you must decide whether the current trend is up or down and how long it will continue. Of course, if you knew the answer to that, you would not need to optimise. The section on consistency and adaptability below discusses the issue of data in greater depth.

Dealing costs must also be considered. If you operate your optimisation with no dealing costs, then there is no cost in taking a trade that results in 0. 1% profit. This means that your most profitable situation may be dozens of small profit trades which add up to a high overall profit, when in fact, with dealing costs, those same trades would have all been unprofitable. The dealing costs you use in your optimisation are there to avoid small trades. Throughout the optimisations conducted, a 1.5 % commission on the price is applied on entry and exit. Prices dealt are always the middle price of the period. You could make your optimisations more onerous, and perhaps more realistic, by buying at the high and selling at the low.

Test parameters

Throughout all the optimisations undertaken, the following test parameters were used:

3-box reversal charts only.

Log scale charts only.

Box size range from 0.5% to 5%, stepping up 0.1% at a time. This means that 0.5% will be tested, then 0.6%, then 0.7% and so on.

Test using close only and the high/low construction methods.

Entry and exit signals. Buy on first occurrence of a double-top buy signal. Ignore any subsequent double-top buys. Sell on first occurrence of a double-bottom sell. Ignore all subsequent double-bottom sells, until the next double-top buy starts the process again.

Data. This is discussed in the data consistency and adaptability section below.

Data consistency and adaptability

The data under consideration is seen to be the most difficult issue when optimisation is considered, not only how the data is arranged, but also the number of data points. The number of data points relates to your time horizon. The more data you take into account with your optimisation, the longer your time horizon is. Conversely, the less data, the shorter the time horizon. The reason is that if the optimisation exercise has less data to work with, it needs to adjust the parameters to extract the greatest profit from a small amount of data. The longer

Chapter 7 – Optimisation of Point and Figure Charts

359

The Definitive Guide to Point and Figure

the data series, the more compromises are made to satisfy all the different characteristics within the data series.

For the results of the optimisation to be useful, they must be consistent as time passes, otherwise what worked best in the past can be of no use in the future. At the same time, they must be adaptable, so that, as new data becomes available, the parameters adjust by small amounts to account for the new information. Consistency and adaptability means that the results adjust slowly as time passes. It is, therefore, unlikely that you wiJI achieve this by selecting a period of 250 days, 5 years ago, and then another 250 days, 5 years later. The results of the two optimisations will most likely be entirely different because the two sections of data are independent of one another. The lack of data overlap means that the parameters have not been able to adapt to changes in the data’s characteristics as time passes.

Amazon.com Inc is a share that has risen as well as fallen over the last seven years, making it an ideal candidate to conduct an optimisation. Table 7-1 below shows the results of optimisations undertaken on Amazon.com Inc, a year at a time, without any overlapping of the data series. In each case, January-to-January data was used. Notice the consistency in the results for the first two years. This means that if the parameters found by optimising year 1 were used in year 2, they would have achieved the best result in year 2 as well.

Table 7-1 : Optimised box size for 3-box reversal charts for Amazon.com Inc- one year sections of data

The problem is that if you had used those same parameters in year 3, you would have suffered, because the box size has halved. It then remained consistent for years 3, 4 and 5, then in year 6 the box size doubled again. It halved again in year 7. There is no consistency because there is no adaptability. To achieve this, therefore, there must be an overlap of data so that the parameters can change as new data becomes available. Overlapping the data will explain how the ‘jump’ from 2.4% to l .0% occurred.

In fact, it is easy to see why it changed by observing Chart 7-1 opposite. The section outlined in black is the year January 2000 to January 200 l . It is obvious that the characteristic of the chart has changed completely and those parameters determined prior to the outlined section are unlikely to work during it. This illustrates the problem of optimising market data.

360

The Definitive Guide to Point and Figure

Table 7-2: Optimised box size for reversal for Amazon.com Inc – stepping forward one month at a time

The table also shows the top four results in each case because often the difference in profit between these is so slight that any one ofthem could be used. Notice in the first optimisation, 2.4% was best and 1.7% was 3rd best. In the second optimisation, 1.7% moved from 3rd to 1st, and 2.4% moved from 1st to 3rd. This is adaptability. The parameters are adjusting to the new data, allowing the parameters to change slowly as time passes. The only one that seems out ofplace is 2. 1% in the 6th optimisation. In this case, 2.1% was the 9th best in the 5th optimisation, making it still acceptable. This will occur occasionally at the chart’s change over point when the chart’s trend has finally changed. In the Amazon.com Inc chart, it occurs in the year that ends with the column breaking down from the large top pattern.

Notice also that 1% proved to be the best box size from the seventh optimisation onwards, indicating consistency. The purpose of this exercise was to demonstrate that taking independent sections of data and expecting the parameters to remain constant is not the case. Optimisations should be conducted regularly, perhaps every week, every month or even every day, including the latest data. Many reading this will throw up their arms in horror and dismiss the technique as curve fitting. Indeed it is, but is it not important to adjust your parameters by using the latest data? There is one proviso, however: you must exclude any unrealised profit or loss resulting from an open position. This means that only data to the left of the last open position signal is used to determine the profit or loss that in tum determines the best set of parameters to use. Those parameters will, by definition, be the best possible ‘guess’ for the near future.

Finally, it is significant to note that in all the optimisations, the close only method proved best although the high/low method started to appear in the rankings towards the end. From this you may reach the conclusion to use the close only rather than the high/low construction method.

362

Chapter 7 – Optimisation of Point and Figure Charts

Entry: Double-top
Exit: Double-bottom or Stop loss

Exit

14.0% Stop 6.0% Stop 8.5% Stop 14.0% Stop 9.5% Stop Double-bottom 13.5% Stop Double-bottom 9.5% Stop 9.0% Stop 10.0% Stop 10.0% Stop Double-bottom 15.0% Stop 15.0% Stop 13.0% Stop 14.5% Stop 14.5% Stop Double-bottom 15.0% Stop Double-bottom 11.0% Stop Dou ble-bottom 13.5% Stop 14.5% Stop 10.5% Stop 12.5% Stop 14.5%Stop

1 1 .5% Stop Double-bottom Double-bottom 14.0% Stop 9.0% Stop

13.5% Stop_ Double-bottom 5.5% Stop 14.0% Stop 14.0% Stop 14.0% Stop 14.0% Stop 1 2.5% Stop
1 2.5% Stop 14.5% Stop 9.0% Stop 4.5% Stop 14.5%Stop 14.0% Stop 6.5% St()p

Entry: Double-top

Exit: Double-bottom or Stop loss

Code

Box x Rev.

Exit

10.0% Stop
1 5.0% Stop
1 5.0% Stop 12.5% Stop
1 3.0% Stop Double-bottom 14.0% Stop 11.0% Stop 14.5% Stop 11.0% Stop 12.0% Stop Double-bottom 10.0% Stop 15.0% Stop Double-bottom 15.0% Stop 7.5% Stop 14.0% Stop 12.5% Stop 14.0% Stop Double-bottom Double-bottom 12.0% Stop 7.5% Stop 13.0% Stop 13.0% Stop 14.5% Stop 6.0% Stop 11.5% Stop 5.0% Stop 13.0% Stop 10.5% Stop 12.5% Stop 5.5% Stop 9.0% Stop Double-bottom Double-bottom 12.0% St� Double-bottom 8.0% Stop 14.0% Stop 13.5% Stop Double-bottom 9.5% Stop 13.5% StOl> 14.5% Stop 9.5% Stop Double-bottom

2.1% x 3 Table 7-4: Optimisation of S&P 1 00 constituents from January 1 998

XRX

365

The Definitive Guide to Point and Figure

Table 7-4 shows:

1 .

2.

3.

The instrument code.27

The best box size for a 3-box reversal chart based on an entry signal ofa double-top buy and an exit either on a double-bottom sell or % stop loss.

Which was the better exit: a double-bottom sell or a stop loss. If a stop loss, the

percentage is shown.

The average box size is 1 .4%, slightly smaller than the FTSE 1 00. In most cases, the results can be improved by exiting on a stop loss ratherthan a double-bottom sell. 36% ofthe results required construction using high/low data.

These tables show the best box size, and in some cases the best stop loss percentage, at the time the optimisation was conducted. Although it is likely that they will remain the best for a short time, they cannot be expected to remain so for any length of time and should, therefore, not be relied on.

Optimising for shorts

All the optimisations conducted so far have assumed that your position is opened by buying. Ofcourse, it is just as easy to open a position by selling, called a ‘short position’.

Optimisation of FTSE 1 00 constituents for shorts

Table 7-5 shows the results of an optimisation on the FTSE 1 00 constituents using a double­ bottom sell as the entry signal and either a double-top buy or a stop loss as the exit signal.

Notice that the stop loss exit, in other words a buy based on a stop loss, is a far more common exit than one based on a double-top buy. Notice also that the box sizes are different from those achieved for longs and that the high/low construction method is less common.

This shows that if you intend shorting, you should be looking at a completely different Point and Figure chart. Optimisation will guide you to the best box size to achieve this because optimisation of shorts maximises the profit from shorts rather than longs.

27 A table of instrument names is shown in Appendix D. 366

Chapter 7 – Optimisation of Point and Figure Charts

Short Entry: Double-bottom

Short Exit: Double-top or Stop loss

Short Entry: Double-bottom

Short Exit: Double-top or Stop loss

Code

ABF

AL-

ALLD

ANTO

AUN AV- AVZ AZN BA- BAA BARC BATS BAY BG- BLND BLT BNZL BOC BOOT BP- BSY BT-A CBRY CCL

CNA

CPI

CS-

CW-

DGE

DMGT

DXNS

EMA

EMG

ETI

EXL

GLH

GSK

GUS

HAS

HBOS

HG-

HNS

HSBA ICI IHG

Box x Rev.

1.5% x 3 2.3%x3 2.2%x3 1.6% x 3 2.0%x3 2.2%x3 2.4%x3 1.2% x 3 H/L 2.1% x 3 2.4% x 3 2.3% x 3 2.5% x 3 1.0% x 3 2.5% x 3 1.4% x 3 2.5% x 3 1.7% x 3 2.4% x 3 1.3% x 3 2.3% x 3 H/L 1.2% x 3 2.5% x 3 2.3%x3 2.2%x3

1.6%x3 2.3%x3 0.9%x3 2.5%x3 2.5%x3 2.2%x3 2.4%x3 0.5%x3 2.3%x3 0.8%x3 1.4%x3 2.4%x3 1.2%x3 2.1%x3 2.5%x3 2.5%x3 2.5%x3 1.8%x3 2.4%x3 2.2%x3 2.5% x 3 H/L

Exit

15.0% Stop 7.5% Stop 14.5% Stop 3.5% Stop 1.0% Stop Double-top 11.5% Stop 15.0% Stop Double-top 8.0% Stop 1.5% Stop Double-top 12.0% Stop 1.0% Stop 13.5% Stop 10.0% Stop 2.5% Stop Double-top

Double-top

1 3.0% Stop 11.5% Stop Double-top 13.0% Stop 6.5% Stop 13.0% Stop 7.5% Stop 7.5% Stop Double-top 13.5% Stop 14.0% Stop 14.0% Stop 14.5% Stop 1.0% Stop 4.5% Stop 2.0% Stop 2.0% Stop 1 1 .5% Stop Double-top Double-top 5.0% Stop 11.0% Stop 4.0% Stop 13.5%Stop Double-top 5.5% Stop

Code Box x Rev. III 1.2% x 3

IMT 2.4% x 3 IPR 0.5% x 3 lTV 2.5% x 3

JMAT 2.0% x 3 KGF 2.5% x 3

LAND 2.2% x 3 H/L LGEN 2.4% x 3 LLOY 0.8% x 3

MKS 1.9% x 3 MRW 2.5% x 3 NGT 2.0% x 3 NRK 2.3% x 3 NXT 1.1% x 3 PRU 2.2% x 3 PSON 2.1% x 3

RB- 2.2% x 3 H/L RBS 2.5% x 3 REL 2.5%x3 REX 2.5%x3 RIO 1.7%x3

RR- 1.7%x3 RSA 2.0%x3 RTO 2.4%x3 RTR 2.4%x3

SBRY 1.5%x3 SCTN 0.7%x3 SDR 1.0%x3 SGE 1.0%x3 SHEL 2.3%x3 SHP 1.1%x3 SMIN 1.4%x3 SN- 2.5%x3 SPW 2.3%x3 SSE 1.8%x3 STAN 1.2%x3 SVT 2.2%x3 TATE 2.4%x3

TSCO 1.6% x 3 H/L ULVR 2.3% x 3 H/L UU- 2.4% x 3 VOD 1.7% x 3 H/L

was 2.3% x 3 WPP 2.1% x 3 WTB 1.1% x 3

Exit

Double-top

5.0% Stop 15.0% Stop 12.5% Stop 3.0% Stop 8.0% Stop 14.5% Stop 13.5% Stop 15.0% Stop 14.0% Stop 8.5% Stop 5.0% Stop 0.5% Stop Double-top 13.0% Stop Double-top 1 3.5% Stop 0.5% Stop Double-top 7.0% Stop 12.5% Stop 13.0% Stop 11.0% Stop Double-top 12.0% Stop 13.5% Stop 14.5% Stop 15.0% Stop 15.0% Stop Double-top 6.5% Stop 6.0% Stop 4.0% Stop 4.5% Stop 11.5% Stop 7.5% Stop 6.0% Stop 4.5% Stop

14.5% Stop 6.5% Stop Double-top 8.5% Stop 3.0% Stop 12.5% Stop 9.0% Stop

Table 7-5: Optimisation of FTSE 1 00 constituents based on short trades

367

The Definitive Guide to Point and Figure

Optimisation of S&P 1 00 constituents for shorts

Table 7-6 shows the results of an optimisation conducted on the S&P 1 00 constituents using a double-bottom sell as the entry signal and either a double-top buy or a stop loss as the exit.

Once again, the trailing stop loss exit has proved best in most cases and the high/low construction method is only used in a few cases.

Short Entry: Double-bottom

Short Entry: Double-bottom

Short Exit: Double-top or Stop loss

Short Exit:

Double-top or Stop loss

Exit

3.5% Stop

8.0% Stop 8.5% Stop 10.5%Stop 8.0% Stop 6.5% Stop 2.5% Stop 9.0% Stop 6.0% Stop 1.0% Stop 14.0% Stop 9.5% Stop 2.5% Stop Double-top 10.0%Stop 5.5% Stop 8.5% Stop 6.5% Stop Double-top 14.5% Stop 10.5%Stop 9.5% Stop 15.0% Stop 13.0%Stop 3.0% Stop 9.0% Stop Double-top 3.5% Stop 3.5% Stop

Double-top_

1 1 .0% Stop

3.5% Stop

6.5% Stop

13.5% Stop

7.5% Stop

Double-top

14.5% Stop

11.5% Stop

4.5% Stop

7.5%Stop 3.5% Stop 6.0% Stop 14.5% Stop 6.0% Stop 3.5% Stop 4.0% Stop Double-top 7.0% Stop

Code

INTC

IP

JNJ

Box x Rev.

2.0% x 3 2.3% x 3 2.5% x 3

Exit

Double-top

4.5% Stop

8.5% Stop

10.0% Stop

7.0% Stop

13.0% Stop

6.0% Stop

DOUble-top

15.0% Stop DOUble-top 0.5% Stop 7.0% Stoj:)_ 10.0% Stop 4.0% Stop Double-top_ Double-top 11.0%Stop 14.0% Stop 8.0% Stop 14.0% Stop 2.5% Sto� 13.5% Stop 8.0% Stop 14.5% Stop 11.5%Stop 5.0% Stop 14.0% Stop 5.5% Stop Double-top 9.0% Stop 4.0% Stop 10.0% Stop 2.0% Stop 15.0% Stop 6.5% Stop 9.5% Stop 0.5% Stop 6.0% Stop Double-top 6.5% Stop 4.5% Stop 2.0% Stop 0 . 5 % Stop 11.0% Stop 3.5% Stop 7.5% Stop

7.0% Stop

10.5% Stop

JPM 1.6%x3 KO 1 .3% x 3 LEH 0.7% x 3 LTD 1.0% x 3 LU 0.5% x 3 MAY 1.7%x3

2.5%x3

1.9% x 3 H/L

2.3% x 3

2.4% x 3

MCD

MDT

MEDI

MER MMM MO MRK MSFT MWD NSC NSM NXTL ORCL PEP PFE PG ROK RSH RTN S SBC SLB SLE SO T TOY TWX TXN TYC UIS

2.4% x 3 H/L

2.3% x 3

2.3% x 3 2.4% x 3 2.1% x 3 1.1% x 3 2.2% x 3 0.5% x 3 2.0% x 3 1.6% x 3 2.1% x 3 1.3% x 3 1.8% x 3 2.3% x 3 1.1% x 3 2.2% x 3 2.3% x 3 H/L 1 .8% x 3 1.7% x 3 0.8% x 3

1 .4% x 3 H/L

2.3% x 3

1.4% x 3

2.1% x 3

0.7% x 3

0.8% x 3

2.4% x 3 H/L

2.4% x 3

2.0% x 3

GE

1.8% x 3 H/L GM 1.3%x3 HAL 2.5% x 3 HCA 2.1%x3 HD 1.2%x3 HET 2.2% x 3 HIG 1.8%x3 HNZ 2.2% x 3 HON 2.5% x 3 HPQ 1.9%x3 IBM 1.5%x3

1.2% x 3 USB 1.4%x3H/L

UTX 2.3% x 3 VZ 2.5%x3 WFC 2.4%x3 WMB 0.7%x3 WMT 2.1%x3 WY 1.2% x 3 XOM 2.4%x3

XRX 1 .4% x 3 H/L Table 7-6: Optimisation of S&P 100 constituents based on short trades

368

Optimising for specific patterns

As discussed earlier, the only pattern guaranteed to occur on every chart is the double-top buy and double-bottom sell. This does not mean that you have to limit your optimisations to these. You may prefer to enter your long trade on a less common pattern such as a triple-top or a catapult, provided you don’t try to exit on one; they may never occur. Therefore, ifyou optimise for entry on either a catapult or a triple-top, the results will tell you the best box size and construction method to achieve the greatest profit from these patterns.

Catapult entry signals

Optimisation of FTSE 1 00 constituents for catapult entry for longs

Table 7-7 shows the results ofan optimisation conducted on the FTSE 100 constituents again, but this time using a bullish catapult as the entry signal and either a double-bottom sell or a stop loss as the exit.

Chapter 7 – Optimisation of Point and Figure Charts

Code

ABF

AL-

ALLD

ANTO

AUN AV- AVZ AZN BA- BAA BARC BATS BAY BG- BLND BLT BNZL BOC BOOT BP- BSY BT-A CBRY CCL CNA CPI CS- CW- DGE DMGT DXNS EMA EMG ETI EXL GLH GSK GUS HAS HBOS HG- HNS HSBA ICI IHG

Box x Rev.

0.5% x 3 H/L 1.4% x 3 H/L 1.9% x 3 H/L 0.8% x 3 0.8%x3H/L 0.5%x3H/L 1.5%x3H/L 0.9% x 3

1 .8% x 3

0.6% x 3

0.7%x3H/L

1.7%x3H/L

2.5%x3H/L

1.6%x3H/L

0.7%x3H/L

2.3%x3H/L 1.0% x 3 H/L 1 .7% x 3 H/L 0.8%x3H/L 1.9% x 3 1.9% x 3 1.8%x3H/L 1.8%x3H/L 1.3% x 3 H/L 1.7% x 3 1.1% x 3 2.2% x 3 H/L 0.6%x3H/L 0.7%x3H/L 0.7% x 3 0.6%x3H/L 1.3%x3H/L 0.5%x3H/L 0.9%x3H/L 1.0%x3H/L 1.6%x3H/L 1.8%x3H/L 1.9%x3H/L 1.3% x 3 1.1%x3 0.6%x3H/L 0.5% x 3 1.6%x3H/L 0.5%x3H/L 0.5%x3H/L

Entry

Catapult Buy

Catapult Buy

Catapult Buy

Catapu l t Buy

Catapult Buy

Catapu lt Buy

Catapult Buy

Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy CataultBuy Catapult Buy Cata ult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy CataDult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy CataDult Buy Catapult Buy Cata ult Buy Catapult Buy Catapult Buy CataDult Buy Catapult Buy Cat”Pult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapu lt Buy Catapult Buy

Exit

Double·bottom

8.0 % Stop

Double-bottom

Double-bottom

7.5 % Stop

7.5 % Stop

1.5 % Stop

Double-bottom

Double-bottom

1 1 . 5 % Stop

8.0 % StoD

7.5 % Stop

6.5 % StQQ

8.0 % Stop

9.5 % Stop

13.0 % StoD

9.0 % Stop

10.5 % StoD

15.0 % Stop

7.5 % StoD

Double-bottom

6.0 % Stop

9.5 % Stop 8.5 % Stop 14.5 % Stop Double-bottom 2.5 % Stop 12.0 % StoD 8.5 % Stop 12.0 % StoD 12.5 % Stop Double-bottom 10.5%Stop 6.0 % Stop 13.5%Stop 11.5%Stop 8.0 % StoD 6.0 % Stop 14.5%StO]) 13.5%Stop 11.0%Stop 6.0 % Stop 6.0 % Stop 7.5 % StoD 10.5%Stop

Table 7-7: Optimisation of FTSE 1 00 based on 3-box catapult buy signals & double-bottom or stop loss exits

369

The Definitive Guide to P oint and Figure

Compare Table 7-7, which uses a catapult entry with Table 7-3, which uses a double-top entry. Notice that the high/low (h/l) construction method is found to be the best in most cases when the catapult is used, indicating that if you wish to trade on catapult entries, you are better off drawing high/low Point and Figure charts. This is one of the great benefits of optimisation; it can tell you how best to construct your charts.

Notice also that an exit based on a stop loss is preferred in most cases to an exit based on a double-bottom sell. The reason is that, having entered on a catapult buy, you will want to run the trade for as long as it is profitable. Ifyou get closed out on a double-bottom sell signal, you have to wait for another catapult to occur before you can buy again. As you have seen, that may never happen again and you will have to forego a profit opportunity. Consequently, the less sensitive trailing stop loss ensures you remain with the trend for as long as possible.

Optimisation must, therefore, be tailored to suit your trading style. If you wish to know what box size and construction method is most likely to produce the most profitable catapult entry trades, then optimising based on these conditions is the only way to find out.

Although the Table 7-7 shows the results for long trades using catapult buys, a similar optimisation (not shown) may be conducted for shorts, where the entry signal is a catapult sell and the exit is a double-top buy or stop loss.

370

Optimisation of S&P 100 constituents for catapult entry for longs

Table 7-8 shows the results of a similar optimisation conducted on the S&P 1 00 constituents. The entry signal is a bullish catapult and the exit signal is either a double-bottom sell or a trailing stop loss.

Once again, the highllow (hll) construction method is found to be the best in most cases. The trailing stop loss exit is also preferred to the double-bottom sell in most cases.

Chapter 7 – Optimisation of Point and Figure Charts

Code Box x Rev. AA 1.3x3H/L AEP 1.1x3H/L

Entry

Catapult Buy

Catapult Buy

Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy

Exit Code 1 0.5% Stop INTC

2.5% Stop IP 1 0.5% Stop JNJ 12.5% Stop JPM 1 1 .0% Stop KO

9.5% Stop LEH 1 1 .5% Stop LTD 7.5% Stop LU

Double·bottom MAY 5.5% Stop MCD 14.0% Stop MDT 10.0% Stop MEDI 14.5% Stop MER 14.5% StoP. MMM

10.5% Stop MO

15.0% Stop MRK Double·bottom MSFT 13.5% Stop MWD

13.5% Stop NSC 6.5% StolL NSM Double-bottom NXTL

4.5% Stop ORCL 10.0% Stop PEP 5.0% Stop PFE 9.0% Stop PG Double-bottom ROK 12.5% Stop RSH 8.5% Stop RTN

3.5% Stop S 12.5% Stop SBC 14.5% Stop SLB 14.0% Stop SLE 15.0% Stop SO 13.0% Stop T

Double-bottom TOY 12.0% Stop TWX 14.0% Stop TXN 14.0% Stop TYC

Dou ble-bottom UIS 5.0% Stop USB 1 1 .0% Stop UTX

1 3.0% Stop VZ 11.5% Stop WFC 8.0% Stop WMB

6.0% Stop WMT

7.0% Stop WY 14.0% Stop XOM 10.5% Stop XRX

Box x Rev.

0.6x 3

0.9x 3 H/L

1.7x 3

1.4x3H/L 2.0x 3 H/L 1.5x3H/L 0.5x 3 H/L 1.1x 3 H/L 2.0x 3 H/L 0.5x 3 H/L 0.8x 3 H/L 1.3x 3 LOx 3 H/L 2.4x 3 1.8x3H/L 1.3x3H/L 0.6x 3 H/L LOx 3 H/L 1 .0x 3 1.6x3H/L 1.6x 3 2.5x 3 H/L 0.5x 3 H/L LOx 3 H/L 0.6x 3 1.0x3H/L 1.1x 3 H/L 1.2x3H/L 0.7x 3 LOx 3 H/L 1.1x 3 H/L 0.7x 3 H/L 0.7x 3 H/L 0.6x 3 H/L 1.5x3H/L 1.7x3H/L 1.3x 3 1.5x3H/L 1 .4x 3 H/L 1.8x3H/L 1.1x 3 H/L 0.7x 3 H/L 0.7x 3 H/L 1.2x3H/L 1.9x 3

2.3x 3 H/L

0.6x 3

2.1x3H/L

Entry

Catapult Buy

Catapult Buy

Catapult Buy

Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buv Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult B u v Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buy Catapult Buv Catapult Buy Catapult Buy Catapult Buv Catapult Buy Catapult Buy Catapult Buy Catapult Buy

Exit

14.0% Stop

9.0% Stop 14.0% Stop 12.5% Stop 7.0% Stop Double-bottom 14.0% Stop 6.5% Stop 12.0% Stop 8.5% Stop 12.0% Stop 11.5% Stop Double-bottom 6.0% Stop 6.5% Stop Double-bottom Double-bottom 10.0% Stop 9.0% Stop 8.5% Stop Double-bottom 12.5% Stop 9.0% Stop 6.0% Stop 6.0% Stop Double-bottom 1.0% Stop 9.0% Stop 11.0% Stop 3.0% Stop 13.0% Stop 7.5% Stop 12.5% Stop 7.0% Stop 7.5% Stop 14.5% Stop Double-bottom 11.0% Stop 12.0% Stop Double-bottom 4.5% Stop 12.0% Stop 6.5% Stop 14.0% Stop 9.0% Stop 14.0% Stop

1 1 .0% Stop

5.5% Stop

AES

AIG

ALL

AMGN AVP AXP BA BAC BAX BDK BHI BMY BNI BUD C CCU CI

CL

CPB

CSC CSCO DAL DD DELL DIS DOW EK EMC EP ETR EXC F FDX G GD

2.0x 3 H/L

0.6x 3

1.3x3H/L

1.8x 3 H/L 0.7x 3 0.7x 3 H/L 2.1x 3

1.2x3H/L 0.7x 3 H/L 0.8x 3 H/L 1.9x3H/L 1.4x3H/L 0.6x 3 H/L 0.5x 3 H/L 2.2x 3 H/L 2.5x 3 H/L 2.1x 3 H/L

2.3x 3 0.9x 3 H/L 2.2x 3 H/L 1.1x 3 H/L 1.1x 3 H/L 0.5x 3 H/L 0.7x 3 H/L 1.4x3H/L 0.7x 3 H/L 1.1x 3 H/L 1.1x 3 H/L 0.9x3H/L 1.0x3H/L 1.0x3H/L 0.9x 3 H/L 2.3x 3 H/L 0.9x 3 H/L 0.8x3H/L

GE 1.1x3H/L

GM

HAL

HCA

1.2x 3 0.7x 3 H/L 0.7x 3 H/L

HD 1.8x3H/L HET 1.7x3H/L HIG 1.3x3H/L HNZ 1.5x 3 HON 2.3x 3 H/L HPQ 1.4x3H/L IBM 1.4x3H/L

Table 7-8: Optimisation of S&P 1 00 based on 3-box catapult buy signals and double-bottom or stop loss exits

371

The Definitive Guide to Point and Figure

Triple-top entry signals

It is possible to run an optimisation based on any entry pattern, and the next logical one is a triple-top buy signal. As with catapults, they don’t occur that often, but when they do, they tend to produce good gains.

Optimisation of FTSE 1 00 constituents for triple-top entry for longs

Table 7-9 shows the results ofan optimisation conducted on the FTSE 100 constituents using a triple-top as the entry signal and either a double-bottom sell or a stop loss as the exit.

Code

III

IMT

IPR

lTV

JMAT

KGF

LAND

LGEN

LLOY

MKS

MRW

NGT

NRK

NXT

PRU

PSON

RB-

RBS

REL

REX

RIO

RR- RSA RTO RTR SBRY SCTN SDR SGE SHEL SHP SMIN SN- SPW SSE STAN SVT TATE TSCO ULVR UU- VOD WOS WPP WTB

Box x Rev.

1.7% x 3

2.1% x 3 0.9% x 3 H/L 0.9% x 3 H/L 1.7% x 3 1.5%x3H/L 0.6% x 3 2.2% x 3 1.1%x3H/L 1.1%x3H/L 1.0%x3H/L 1.0%x3H/L 2.4% x 3 H/L 1.9% x 3 2.2% x 3 2.5% x 3 1.6% x 3 H/L 0.6% x 3 0.6% x 3 H/L 0.8% x 3 H/L 2.1% x 3 H/L 1.7% x 3 2.0% x 3 2.0% x 3 H/L 1 .4% x 3 H/L 2.0% x 3 H/L 2.5% x 3 H/L 1 . 1 % x 3 H/L 2.5% x 3 1.9% x 3 H/L 2.0% x 3 0.7% x 3 1.3% x 3 H/L 1 .5% x 3 1.0% x 3 1.8% x 3 H/L 1 .4% x 3 1.9% x 3 H/L 1.5% x 3 0.5% x 3 H/L 2.0% x 3 1.8% x 3 H/L 1.7% x 3 H/L 2.2% x 3 1.5% x 3 H/L

Table 7-9: Optimisation of FTSE 1 00 based on triple-top buy signals and double-bottom or stop loss exits

372

Chapter 7 – Optimisation of Point and Figure Charts

Once again the stop loss is the favoured exit, although the choice ofconstruction methods is evenly balanced between close only and high/low. This is because the high/low spikes are more likely to produce catapults, which are, in effect, triple-tops which have failed on the first break.

It is worth noting that when a triple-top occurs, a double-top is also occurring. This means that some of the triple-top buys may be at the same point as a double-top buy. Indeed, some of the triple-tops in Table 7-9 will also have been double-tops.

373

The Definitive Guide to Point and Figure

Optimisation of S&P 1 00 constituents for triple-top entry for longs

Table 7- 1 0 shows the results of an optimisation conducted on the S&P 1 00 constituents using a triple-top as the entry signal and either a double-bottom sell or a stop loss as the exit.

Exit

1 1 .5% Stop

6.0% Stop

Double-bottom

12.5% Stop

7.5% Stop

Double-bottom

Double-bottom

Double-bottom

9.5% Stop 4.5% Stop 9.5% Stoo 1 0.0% Stop 14.5% Stop 1 5.0% Stop 11.0% Stoo 1 3.0% Stop 14.5% Stop 1 3.5% Stoo 1 3.5% Stoo 6.5% Stop_ 8.0% Stop 1 5.0% Stop 11.0% Stop 3.5% Stop 11.0% Stop 10.0% Stop 1 3.0% Stop 1 0.5% Stop 8.5% Stop 7.0% Stop

1 4.5% Stop

1 4.0% Stop

1 5.0% Stop Double-bottom Double-bottom 4.5% Stop

Code

INTC

IP

JNJ

JPM

Box x Rev.

Entry Exit Triple-top 9 . 5 % Stop Triole-too Double-bottom

0.7x 3

2.3x 1 .2x 1 .8x

H/L

H/L H/L H/L H/L H/L 3 H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L H/L 3

3

H/L

H/L

H/L

H/L

H/L

H/L

3

H/L

H/L 3 H/L H/L H/L H/L H/L H/L H/L H/L

3 3 3 3 3

Triple-top

Triple-top

Triple-top

14.0% Stoo 12.5% Stop 4.5% Stop 10.0% Stop 14.0% Stoo 11.0% Stop 12.0% Stop 11.0% Stop 14.0% Stoo 14.0% Stop 9.0% Stop 15.0% Stop 10.0% Stoo 8.0% Stop 8 . 0 % Stop

KO 1.1x LEH 0.7x LTD 1.4x

Triple-top
Triole-too
Triple-top
Triple-top
Triple-top
Triole-too
Triple-top
Triple-top
Triple-top
Triole-too
Triple-top
Triple-top
Triple-top Double-bottom

LU 0.7x MAY 2.0x MCD 0.7x MDT 1.6x MEDI 1.7x MER 1.3x MMM 0.7x

MO 1.3x MRK 1.8x MSFT 1.0x MWD 2.4x NSC 1.0x NSM 1 .6x NXTL 1 .0x ORCL 1 .8x PEP 0.6x PFE 1.0x PG 1.1x ROK 0.8x RSH 0.7x RTN 0.7x

3

3

3

3

3

3

3

3

3

3

3

3

3 3 3 3 3 3 3 3 3

Triole-too

Triole-too

1 5.0% Stop Triple-top 13.5% Stop

Triple-top

1 2 . 5 % Stoo

1 3.5% Stoo

S 1.4x SBC 0.9x

Triple-top

Triple-top

Triple-top

Triple-top

Triole-too

Triple-top

Triple-top

10.0% Stop

1 0.5% Stop

1 3.0% Stop

1 5.0% Stop

1 4.5% Stop

11.0% Stop

1 0.0% Stop

Triple-top Double-bottom

SLB 0.6x SLE 1.2x SO 0.7x T 0.6x TOY 1.5x TWX 0.8x

3

3

3

3

3

3

TrioIe-too

14.0% Stop

6.5% Stop 7.5% Stop 12.5% Stop 15.0% Stoo 13.5% Stop 11.0% Stop 12.0% Stop

3.5% Stoo

14.0% Stop TXN 1 .7x

TrioIe-too
Triple-top
Triple-top
Triple-top
Triole-too
Triple-top
Triple-top
Triple-top
TrioIe-too Double-bottom

14.5% Stop

10.5% Stop

Double-bottom

11.5% Stop

9.0% Stop

14.5% Stop

8.0% Stoo 6.0% Stop 13.5% Stop 14.0% Stop Double-bottom

TYC 1.2×3 UIS 2.0x 3 USB 1.5x

UTX 1.1x VZ 2.0x WFC 1.7x WMB 1.4x WMT 2.2x WY 1.4x XOM 1 .2x XRX 1.1x

3

3

3

3 3 3 3 3

Triple-top

Triple-top

Triple-top

1 3.0% Stop 1 2.0% Stop 1 .5% Stop

Triole-too Double-bottom

Triple-top

Triple-top

Triple-top

Triole-too

11.0% Stop

1 4 . 0 % Stop

9 . 5 % Stop

1 4.0% Stop

Table 7-1 0: Optimisation of S&P 1 00 based on triple-top buy signals and double-bottom or stop loss exits

The trailing stop loss continues to be the favoured exit, but it is significant to note that the high/low construction method has been used 70% of the time. This is a significant increase over that when the FTSE 1 00 was optimised. This says more about the state of the high/low data on FTSE 100 stocks than anything else.

374

Unless you are looking very short-term, percentage box sizes should be used.

Including uncommon Point and Figure patterns is fine for entry signals, but they should never be included as possible exit signals.

Chapter 7 – Optimisation of Point and Figure Charts

Conclusion

I f you think that optimisation will replace chart reading, you should think again; i t will not. Optimisation will guide you to the best parameters to use, based on past performance, but it will not provide you with a formula for untold wealth. It can tell you whether an instrument’s characteristics require the high/low construction rather than close only, but you must be aware that it can change. It will tell you what box size to start with, based on taking every buy and sell signal, but that does not mean that is the box size you must use, because you won’t take every buy and sell generated from the chart.

Optimisation, therefore, gives you the parameters to get you started, from which point your subjective analysis and knowledge of Point and Figure charts should take over and adjustments be made. Remember when analysing Point and Figure charts there are many things to take into consideration, such as trend and pattern formation, but before doing that you have to make the decision about box and reversal. Optimisation can assist with that decision.

This chapter has only scratched the surface as far as optimisation is concerned. It is a vast subject and far more sophisticated optimisations need to be performed. Provided you understand that the ‘holy grail’ is not the prize, then further work on optimisation should be considered with an understanding of the following:

Optimisation is only of value with 3-box reversal charts, because of their unambiguous signals.

•

•

•

•

Optimisation should always be conducted using the latest data because this allows the parameters to adjust as new data is received.

The unrealised profit from any open position should be excluded from the optimisation, so box sizes are calculated on closed positions only.

Dealing costs must be taken into account to prevent lots of small trades from being included.

Optimisation can tell you what box size is working best, based on double-top and bottom signals. In doing so, it tells you about the characteristics ofthe instrument.

In most cases, profits are increased considerably when trailing stop loss is used as the exit rather than a double-column signal.

Optimisation can tell you whether the close only or high/low construction method is working best for the particular instrument.

375