Bullish Percent Index

Thomas Dorsey identifies the NYSE Bullish Percent Index in Point & Figure Charting.
Developed by Chartcraft for the NYSE, the index can be applied to any market.

The index is based on the percentage of stocks on the exchange showing buy signals:

• Each individual stock is charted on a Point & Figure chart (using a 2 percent logarithmic scale and a 3 box reversal), and
• Buy signals are taken as any column of X's that is higher than the previous column of X's.

The index itself is plotted on a Point & Figure chart with overbought level at 70 per cent and oversold level at 30 per cent. Boxes are fixed at 2 per cent and the reversal amount is 3 boxes (6 per cent).

Bull Confirmed Market

This signals the trader to go long when a column of X's on the index chart exceeds a previous column - the same as the buy signal on an individual chart.
The closer the signal is to the oversold level, the stronger the signal. Likewise, a signal close to the overbought level is weak.

Bear Confirmed

A column of O's falls below the previous column of O's. The closer to the overbought level the stronger the signal. Go short.

Bull Alert Market

When the index starts a new column of X's below the oversold level (remember that a new column is only started when price has reversed by 3 boxes). Take long positions but with caution.

Bear Alert

A new column of O's is started above 70 and drops to below the overbought level - without falling below the low of the previous column of O's. Take short positions but with caution.

Bull Correction

The index pulls back by 6 per cent after a Bull Confirmed signal. This may take place above or below the overbought level but must not cross the critical 70 per cent level. This indicates that the bull market is losing steam but does not indicate a reversal. Prepare to buy on the next rally that carries above 6 per cent.

Bear Correction

This indicates that the bear market is re-grouping but it is not a reversal signal. Prepare to go short on the next correction that starts a new column of O's

Coppock Indicator

Edwin Coppock developed the Coppock Indicator with one sole purpose: to identify the commencement of bull markets. The indicator was devised for use on the Dow Jones Industrial Average but is suitable for use on other market indices or averages

Although often late, the Coppock Indicator has produced very reliable signals in the past.

Trading Signals

A bull market is signaled when the Coppock Indicator turns up from below zero.

EXAMPLE

Mouse over chart captions to display trading signals.

Cycles

Many people seem to think that the change in prices in any one day is complete in itself and bears no relation to larger movements which may be under way. This is not so. (Nelson, 1903)

Dow identified 3 major cycles:

• Daily, or minor, movements last from a few hours up to three weeks;
• Secondary reactions normally retrace one third to two thirds of the primary trend (since the previous reaction) and last from 3 weeks to 3 months; and
• Primary movements of the market may last several years.

A number of other cycles have subsequently been identified within these categories:

• The business cycle (periods of expansion and recession) which varies roughly in line with the primary trend;
• The 4 year presidential cycle: the ruling party generally inflates the economy in the run up to elections, followed by a subsequent contraction at the start of a new term;
• Seasons: stocks generally perform better in Spring and Summer than in Fall or Winter;
• Calendar months: January, April and November are historically good months to hold stocks; February, May and October are poor months. Take a look at the performance of the Dow during October and February over the last 3 years;

• End-of-the-quarter window dressing: portfolio managers sell non-performing stocks that they don't want to show on their end-of-the-quarter books;
• Tax-loss selling in October: professional money managers sell non-performing stocks, creating losses to offset capital gains as they come to the close of their fiscal year;
• Triple witching hour: contracts for stock index futures, stock index options and stock options simultaneously expire on the third Friday of March, June, September and December: The last hour of trading is highly volatile as traders attempt to close their positions;
• The last day of the month and the first 3 days of the new month historically yield higher returns;
• Monday is often the worst, and Friday the best, day of the week for investors - possibly due to short traders closing their positions; and
• Similarly, the last half-hour of any day often experiences price surges as short traders attempt to close out their daily positions.

Sometimes cycles overlap to create larger than normal fluctuations and at other times they offset to create flat periods with little movement. The objective of chart analysis is to identify and predict those cycles which have a dominant effect in a particular time frame

Reading the Market

The direction of the overall market influences the behavior of individual securities. Study the overall market first before looking at any stocks in isolation

The Global Market

US stock markets represent almost half of the combined value of all global markets and significantly influence other markets. Always keep a weather eye on the direction of the 3 averages: the Dow Jones Industrial Average, Standard & Poor 500 and the Nasdaq.
Stay on the alert for events which could have a significant effect on the global (or US) economy. The usual suspects are:

• wars,
• interest rates,
• oil prices, and
• currency collapses.

The chart below shows 3 major indices over a 5 year period. Their growth rates may vary but all move in sympathy with each other.

Market Risk

The market is the major influence on the performance of an individual security. In Point & Figure ChartingThomas Dorsey compares the impact of the market to other influences on the price of an individual stock:

Market risk	66%
Sector risk	24%
Stock risk	10%

Actual figures may vary substantially but the largest influence for most stocks will be market risk.

Market Direction

Chart patterns are often used to identify major market changes:

• Support and Resistance
• Head and Shoulders
• Double Tops
• Triple Tops
• Trading Ranges

Several market direction systems also deserve mention:

• Stan Weinstein's 30-week Moving Average Model from Secrets for Profiting in Bull and Bear Markets.
• William J. O'Neill's Market Top and Market Bottom signals from 24 Essential Lessons for Investment Success;
• The NYSE Bullish Percent Index explained by Thomas Dorsey in Point & Figure Charting;
• MA Systems use two Moving Averages to identify trend changes.
• MACD Histogram is more suited to traders than investors: signals are generated a lot earlier in the trend but are less reliable as a consequence.
• The Coppock Indicator - an oscillator with one sole purpose: to identify the commencement of a bull market.

Momentum Trading [Part 3 of 3]

Momentum Trading [Part 3 of 3]

Background

Bruce Vanstone is Assistant Professor at Bond University in Australia. He completed his PhD in Computational Finance in 2006 and is a regular presenter and publisher of academic work on stockmarket trading systems. He teaches stockmarket trading courses at university, and consults to Porter Capital Management on the design of mechanical, rules-based trading systems. More information on Bruce's research and methods can be found athttp://trading.it.bond.edu.au.

I have asked Bruce to write a series of articles based on his experiences with momentum trading.

~ Colin Twiggs

Introduction

This article is part 3 of a 3-part series. In this final article, I will summarize the key characteristics of investing using momentum based approaches. I will also discuss some approaches to managing risk in momentum models, and the benefits investors can expect when investing with rules-based funds.

Key Characteristics of Momentum Approaches

Perhaps one of the main benefits of the momentum approach is that it actually makes sense! It is not overly complicated to understand, it doesn't rely on split second timing, and it doesn't need huge sums of money to implement.

From a quantitative point of view, momentum trading also carries a number of clear benefits. It is simple to quantify, it is non-subjective, and it is robust to parametric modelling changes. The momentum approach also has academic credibility, and is the subject of ongoing research by some of the worlds best finance academics. The simulations in Part 2 showed that although there is slightly more "raw" risk than pure index investment, there is the potential for substantially higher returns.

It is precisely this possible mismatch between risk and reward which qualifies momentum as an approach worth further study. It's no wonder academics call momentum, "the premier anomaly"!

Managing Risk

Fortunately, there are a number of ways to manage risk that fit with the momentum approach. Distinguished academics like Andrew Lo have published useful research on the potential benefits of stop loss structures on momentum investment, demonstrating, at least in theory, that stop-losses can be of benefit to momentum based systems [1]. Other academic studies have investigated approaches like long-short investing, and hedging, which are techniques that have traditionally been used to reduce trading risk.

Another way of managing risk is to focus on investing when momentum models work the best. The time that momentum models outperform is when there is a clearly defined direction for the overall stockmarket. When the market direction tends to be sideways, it may be better to convert a momentum portfolio back to cash. In the shorter term, this may increase the number of months when small losses occur. However, in the longer term, it will help preserve capital during periods when there is no real expectation from the momentum approach. This ensures that when the market clearly establishes its overall direction, the investor is ready and able to take advantage of it. This is the approach employed by Porter Capital Management. I have included 2 graphs of quarterly returns for the simulation performed in article 2 of this series. The first graph shows the effect of keeping a momentum portfolio fully invested. The second shows the benefits derived by converting the portfolio back to cash when the market shows no clear direction. It is clear in the second graph that several of the quarterly returns have been shifted in a positive direction.

Figure 1: Portfolio fully invested

Figure 2: Portfolio converted to cash when market 'directionless'

Perhaps the most important consideration when using a momentum based approach is the idea of using models and rules to frame the way risk and returns are managed. Research tells us that to be successful in trading and investing requires a well defined plan, which encompasses not only entry and exit decisions, but also covers risk and money management. Having such a plan helps investors cope with the inevitable volatility that markets bring, especially during times when financial markets are under stress.

Importance of rules-based investment

For the average investor, trading and investing are difficult propositions. The average investor needs to invest a significant amount of time to develop the level of market expertise required to be successful.

As I have suggested in Part 2, many investors receive substandard investment returns, and in many cases, it may well be that the investors own behaviour is to blame. Research tends to show that investors have a number of well documented behavioural problems! For example, it has been shown that investors tend to overtrade their accounts to their detriment [2], and, that although investors may pick good stocks they tend to sell winners too early and sell losers too late [3].

Rules based investment allows the investor to specify the conditions under which he will buy or sell stock, and how much stock will be bought or sold. Using rules to quantify your trading decisions leads to clear entry and exit points, and reduces subjectivity, and worry. From a funds management point of view, well-defined rules reduce key man risk, and allow for using quantitative techniques to improve model performance.

Conclusions

During this 3-part series of articles, I have tried to describe the key benefits of momentum investing, and why it may offer potential advantages to investors. I have also tried to describe what I think are the primary benefits of a rules-based, or quantitative style of investment, and how the quantitative style works in with the momentum approach

Momentum Trading [Part 2 of 3]

Momentum Trading [Part 2 of 3]

Background

Bruce Vanstone is Assistant Professor at Bond University in Australia. He completed his PhD in Computational Finance in 2006 and is a regular presenter and publisher of academic work on stockmarket trading systems. He teaches stockmarket trading courses at university, and consults to Porter Capital Management on the design of mechanical, rules-based trading systems. More information on Bruce's research and methods can be found athttp://trading.it.bond.edu.au.

I have asked Bruce to write a series of articles based on his experiences with momentum trading.

~ Colin Twiggs

Introduction

This article is part 2 of a 3-part series. In this article, I will focus on using simulations to demonstrate the potential risks and rewards of the momentum approach. In the final part of the series, I will discuss the way in which investors can benefit from rule-based approaches to investment.

Creating Momentum Simulations

The results presented in this article are a quick demonstration of the potential of the momentum effect for Australian investors. I like to use simulations as they provide an excellent opportunity to see how well a strategy could have performed in the past. Simulations are also useful because they can give some clues as to how a strategy may perform in the future. However, we must always remember that past performance is no guarantee of future performance.

The simulation results in Table 1 have been created by calculating momentum on a historical rolling monthly basis for each member of the ASX200, and holding the top group of stocks each month. The data used contains delisted stocks, and is adjusted for survivorship bias as and where possible. Both simulations assume the same starting capital and account for transaction costs and slippage. The simulations cover the 10 year period from 2000 to 2009.

Applying Simulations to the ASX200

Table 1: Simulations of the Momentum Effect for Australian Investors

Historical Momentum Period	Risk (Max DD%)	Reward (APR%)	ASX200 benchmark Risk (Max DD%)	ASX200 benchmark Reward (APR%)
12	-60.07%	18.54%	-53.13%	4.76%

The columns contained in the table are explained below:

Historical Momentum Period	The number of months over which historical momentum was measured
Risk (Max DD%)	Risk as measured by the maximum drawdown
Reward (APR%)	Reward as measured by APR (annual percentage rate)
ASX200 benchmark Risk (Max DD%)	Risk as measured by the maximum drawdown in the equivalent benchmark (XJO)
ASX200 benchmark Reward (APR%)	Reward as measured by APR in the equivalent benchmark (XJO)

Figure 1 shows an equity graph plotting a simulated portfolio versus the ASX200 (XJO) index portfolio.

Conclusions we can draw from this

In the first article of this series, I pointed out that the momentum effect appears to be one of the most beneficial effects available to investors. The results in Table 1 and Figure 1 clearly confirm this observation.

The results above also confirm that to capture the benefits of a momentum approach, investors do not need to trade frequently, or with huge sums of capital, or in high-frequency timeframes. Instead, what is required is a disciplined, rules-based approach to investment, and a strong focus on risk management.

Although the momentum approach has slightly higher maximum drawdowns than the ASX200, the "potential" returns are substantially higher. Clearly though, investing using momentum alone is not a holy grail for investors! However, these results confirm that the momentum effect could be used to form the basis of an actively managed investment strategy, one that focused on trying to capture some of the outperformance, while still keeping an eye on risk.

In the third part of this series, I will discuss the benefits to investors of mechanical, rules-based trading approaches. Many investors receive sub-standard investment returns, particularly when managing their own capital. In some cases, it is not the strategy itself, but the way it is being implemented which is at fault. If you find yourself attempting to second-guess the way you trade, or you are unsure how to react during periods of market turmoil, then it is likely that you could benefit from the increased discipline and accountability that mechanical approaches can deliver.

Momentum Trading [Part 1 of 3]

Background

Bruce Vanstone is Assistant Professor at Bond University in Australia. He completed his PhD in Computational Finance in 2006 and is a regular presenter and publisher of academic work on stockmarket trading systems. He teaches stockmarket trading courses at university, and consults to Porter Capital Management on the design of mechanical, rules-based trading systems. More information on Bruce's research and methods can be found athttp://trading.it.bond.edu.au.

Financial market volatility over the next few years is likely to cause more actively managed rules-based approaches to out-perform long-term trend-following systems. I have asked Bruce to write a series of articles based on his experiences with momentum trading.

~ Colin Twiggs

Introduction

The purpose of this 3-part series of articles is to provide information about the potential benefits of momentum investing. In this series, I will try and explain what momentum is, the potential returns available to momentum investors, and the way that Porter Capital combine mechanical, rules-based strategies with the momentum effect to deliver benefits to investors.

The 'premier' anomaly

Since its initial discovery by DeBondt & Thaler in 1985^[1], the momentum effect has been documented and researched in many markets worldwide.

Many traders and investors would know of the academic notion of the ‘efficient’ market, and the implication that this efficiency has on the ability of investors and traders to earn profits.

What you may not be aware of is that the father of the ‘efficient market hypothesis’, Eugene Fama, refers to momentum as “the premier unexplained anomaly”^[2]. In other words, the success of momentum based investing is regarded by many as an exception to the efficient market hypothesis.

What is it?

In its simplest terms, momentum refers to buying stocks which exhibit past over-performance. Research shows that stocks which have exhibited strong performance over some defined historical period, have a tendency to continue to exhibit strong performance for some number of future periods. It means that investors can potentially hitch a ride on strong momentum stocks. In part 2 of this series, I will use simulations to explore the potential risks and rewards of the momentum approach.

A Typical Momentum Trade

The typical momentum trade has a history of clearly defined direction and strength. Figure 1 shows a chart of price activity for ALL (Aristocrat Leisure), from August 2004 to April 2005. During late August 2004, there is a clear price breakout on very heavy volume. This marks the start of the momentum opportunity. Over the next few months, the price activity demonstrates clearly defined direction.

Is it credible?

The momentum effect has been widely researched and documented in both the international and Australian equity markets. For example, Rouwenhorst^[3] tested momentum strategies in 12 European markets using data from 1980 to 1995, and found that momentum returns were present in every country, and their effects lasted for approximately one year. Griffin et al.^[4] found support for the profitability of momentum investing in over 40 countries, and concluded ‘Globally, momentum profits are large and statistically reliable in periods of both negative and positive economic growth’.

Momentum has been thoroughly researched in virtually all of the worlds equity markets. Momentum effects have also been documented in other asset classes, such as foreign currencies^[5], commodities^[6] and real estate^[7].

It is fair to say that the momentum effect appears to be one of the most beneficial effects for investors. Thorough research appears to indicate that momentum based investment does not increase investment risk, and that momentum effects are present during both economically good and bad cycles.

When does it work best?

Like all investment approaches, momentum investing is subject to the vagaries of the investor. For many investors, poor returns are not so much a function of their investment strategy, but of their own implementation of that strategy.

All investment strategies benefit from the increased discipline and accountability that mechanical, rule-based trading brings, particularly during difficult investment cycles. I will discuss this topic in more detail in the third part of this momentum series.

References

1. DeBondt, W.F.M. and R.H. Thaler, Does the stock market overreact? Journal of Finance, 1985. 40: p. 793-805.
2. Fama, E. and K. French, Common Risk Factors in the Returns on Stocks and Bonds. Journal of Financial Economics, 1993. 33(1).
3. Rouwenhorst, K.G., International Momentum Strategies. Journal of Finance, 1998. 53(1): p. 267-284.
4. Griffin, J.M., S. Ji, and J.S. Martin, Global Momentum Strategies: A Portfolio Perspective. Available at: SSRN: http://ssrn.com/abstract=492804, 2004.
5. Okunev, J. and D. White, Do momentum-based strategies still work in foreign currency markets? Journal of Financial and Quantitative Analysis, 2003. 38.
6. Taylor, S.J., Forecasting Market Prices. International Journal of Forecasting, 1998. 4: p. 421-426.
7. Stevenson, S., Momentum effects and mean reversion in real estate securities. Journal of Real Estate, 2002. 23: p. 47-64.

Blind Freddy Trends

We are frequently asked: "What is a Blind Freddy trend". The term ( often abbreviated to simply a "BF trend") was coined by members of the Chart Forum to describe a strong (or runaway) trend; one that even "blind Freddy" could spot. Here is a brief summary

Blind Freddy

Stan: The way I see it, the term 'blind freddy trend' has nothing to do with technical analysis as such.

Trend identification isn't always straightforward and there are many definitions of what constitutes an uptrend or downtrend. Add multiple timeframes and things get complicated.

You've had a look at Rocky's excellent trend analysis. That analysis takes times and you can't do it on every chart if you're scanning the whole market. Sometimes the trend is just so plain to see that you don't even bother wasting time analyzing it to technical definitions.

Sometimes it helps just to look at a chart for 3 seconds — shut your eyes — and say whether it's up or down.

Any five year old, my pet kelpie [she's pretty smart], or Blind Freddy can do the same thing.

Basic Requirements

Rocky had set the foundations in an earlier post:

Rocky: ......if I am unsure of the current trend I ask my 6 year old daughter and she tells me which way to trade.Moving average crossovers are a good indication of the trend changing direction but instead of using a fixed number on all charts, scrunch up the shares you are watching and try different combinations say 8/16 10/21 or whatever. Look for the two that don't give too many false signals but give an entry signal regularly for each different stock. You will find that some will have quite long periods and other will have 3/9 periods. Then to be safe, wait for a higher bottom to develop after the crossover before you buy and place a stop loss a few cents below this higher bottom. Another key is to only trade stocks that have a history of trending for reasonable periods of time. Many stocks are so volatile that they aren't worth watching and no sooner will you get an entry signal you are out again.

He then gave a number of examples.

Trendline Break

Rocky: A two-day reaction ending December 12th that sits on top of a previous swing high like this (leaving a space in price is better, of course) is the first indication that an up—trend is starting, possibly a fast one.

The double bottom in the up—trend on Jan 8th and 12th (note the lows are 30 days apart — a typical Gann time cycle) also indicates a trend with great strength.

These subtle signals the Stock has been giving offer a far earlier entry into this trend than waiting for a breakout that may or may not confirm.

Buying breakouts is often a risky proposition — being on board when they happen is sweet.

MA Crossover

Most users employ three simple moving averages: 7—day, 18—day and 30—day; while Rocky used 6, 16 and 30 on some earlier charts. Some members later evolved a system using just two MAs: 10—day and 30—dayexponential moving average crossovers (and 10—week and 30—week EMAs for long—term trends). The length, or type, of MAs used is not critical to the system: use your own if you prefer.
The MAs are not used in a conventional Three Moving Averages fashion (where the fast MA must rise above the medium MA, and the medium MA rise above the slow MA, for a valid up—trend). The slow MA (or sometimes middle MA —— see CBA chart below) is discarded if the other MAs offer a better fit on available price history. Signals are taken from the remaining Two Moving Averages.

Rocky: I just had to post this chart — this is a powerful example of why I like this setup and the US market so much...

This Stock gave a BF buy signal at around $45, powered away, had a 2 day reaction and then leapt 16% on Thursday night to close above $56.

Here is an earlier example of a 6—day and 16—day crossover:

Higher Bottom

In a strong trend, the swing low will often not overlap the previous swing high, leaving a space (or gap) that confirms trend strength. I believe that this concept was first introduced by Bill McLaren, who also counts the number of bars in a reaction to determine trend strength —— another concept discussed below. For a full exposition, see Bill's DVD/video: Foundations for Successful Trading.

Rocky: The space left from the low of December 11 back to the previous swing high made on November 25 was the first indication that a 'Blind Freddy' trend (discussed elsewhere in the forum recently) had the potential to unfold. See chart below —

Since that time, the largest reaction to the rally has been 2 days (up until Friday), and there has been a space in price between each subsequent swing high and the swing low that follows, indicating a very fast move underway. The sellers have been overwhelmed on each occasion and the buyers have won the day — the very essence of an up trend.

Standing in front of this type of trend, without some indication it may be coming to an end (lower top, blow—off gap and reversal) is not the way to make a lot of money trading.

Yes this could be a top, but the highest probability trades have been buying during the pullbacks, not trying to pick the end of the move.

This is danger of relying on indicators — they often get overbought or oversold very quickly and stay there for several days (or weeks?).

Count the days of the reactions — if the Stock is going up 6 days and down two, the trend is strongly up — that's the way to trade until it gives some tangible evidence that the trend is changing

Summary

Indo: Rocky ..... Even though you have spent considerable effort in setting out your BF system they are scattered around a number of posts and I thought it might be useful to try and distill them into a single statement...........

Below is the system for going long — going short is the exact opposite.

SET UP

• A swing low must either be higher or equal to the swing high prior to the higher high currently retracing
• 10 Day EMA has crossed above 30 Day EMA
  (Rocky later confirmed that he still uses 7—day, 18—day and 30—day SMAs)
• Real breakout of downtrend
• 30 Day EMA trending upwards
• If reaction more than 4 days best not to enter position.
• Watch CURRENT resistance levels — not old swing highs and lows

BUY TRIGGER

• Trade a fixed number of cents above the lowest high of the swing low depending on the previous price characteristics of the stock. Can be varied where this is near the previous swing high.

TRADE MONITORING

• Best if reactions only 3 days and then takes out swing high within 3 days
• Reactions should not last more than 4 days and price bars remain above 10 and 30 day EMAs and trendline.
• If stock makes new high and retraces to rest on old high an explosive move may occur
• Initial stop below last swing high then when new swing high established move stop to below the last swing low.
• Number of cents below each level depends on previous price character of the stock.

Indo later went on to explain that the breakout from the down—trend (trendline break) may occur some time before the BF setup

How To Trade Trends

If you are going to trade trends, no matter what the time frame, you are likely to encounter three major problems: false starts, early shakeouts and late exits.

False Starts

Also known as whipsaws, false starts occur when your setup gives a positive signal, immediately followed by a reversal. You are no sooner stopped out of your position when the setup gives another buy signal. This is frustrating; and expensive. Many traders lose heart and fail to take the second entry, only to find that the trend spikes sharply upward, leaving them ready to throw their PC (or themselves) through the window.

Shake-Outs

If you move your stops up to below the low of each subsequent correction, there are going to be times when you are shaken out; no matter how strong the trend. It is just in the nature of the beast. Even if you are more cautious in ratcheting up your stops, applying some kind of filter, there are still going to be times when you are stopped out. And they are going to be expensive -- because of the filter. Shakeouts are covered in more detail elsewhere in the Trading Guide.

Even the strongest trends attempt to shake you out. I am reminded of riding a rodeo bull. It will head off at breakneck speed while you hang on for all you are worth. Just when you feel that you have adjusted to the speed, it will reverse direction. If you are not prepared for this you are going to eat dirt! Another trick is to turn and turn and turn, until again your body senses the rhythm, then, just as suddenly, it will reverse direction. Or the bull will fake to go one way and then take off in the opposite direction.

EXAMPLE

Here is a classic fast-trending stock from the ASX 200. Arc Energy has increased by more than 1000% in the last 4 years. The stock completed a broad double bottom in 2001, with a breakout at [C], and has never looked back. I have used point and figure charts because they are less subjective than the normal closing price or weekly candlestick charts.

Can you see how difficult it is to stay with the trend? Either insiders or professionals had recognized the stock's value and tried every trick in the book to shake out existing positions and claim a bigger stake for themselves. In all there are 16 potential false breaks or shake-outs

Late Exits

Late exits were covered under Selecting A Long-Term Moving Average but a quick recap may help.
Trend-following systems either suffer from a large number of shake-outs or are slow to exit when the trend reverses; and often both. You can't have your cake and eat it. We will deal more with exit strategies in a later article.

Taking Profits Too Early

Apart from systemic problems we also need to consider the human aspect. A trend trading system builds psychological pressure as the trader witnesses repeated gains followed by significant retracements and frequent late exits at trend reversals. Pressure can build to such an extent that the trader overrides his system, attempting to take profits at a perceived high point in the trend. At this point he/she is following their emotions rather than a system -- a recipe for disaster.

Managing psychological pressure will be dealt with in a later article.

Back Fitting Systems

It makes sense to back-test your system on historical data, to ensure that it works, but don't waste time fine-tuning your settings to fit the data. You may design the perfect system to maximize profits on historical data but it will not work as well when you trade live: the future is never an exact replica of the past.

Design Considerations

Type of Stocks and Trends

Blue chip stocks are more reliable but don't expect to find many fast-moving trends. They tend to fall into theslow but steady category; though occasionally they may still spring a surprise.
Fast-moving, highly volatile stocks are more exciting, but also more difficult, to trade. Expect larger numbers of false starts, shakeouts and trend reversals.

Retracements

What type of retracements are you likely to tolerate? This is difficult to answer in abstract, because of the lack of emotional pressure, but give it a try. Imagine that you have a million dollars riding on a single stock** that you have held for several months. How much of a retracement would you comfortably tolerate?

• 10 per cent?
• 20 per cent?
• 30 per cent?

I believe that not many traders could honestly answer more than 20%; and many would be more comfortable with 10%. If that is your comfort zone, do not try to trade a system that tolerates higher retracements.
** Not that I would ever advise you to invest a million dollars in a single stock. It is just that a large sum of money brings home the significance of the percentage losses.

Time Frame

What time frame do you plan to trade? Obviously the tolerated level of retracement will vary depending on the time frame. Also, a stock may be trending at different speeds (and directions) in separate time frames. Swing traders may be very happy with a stock that oscillates rapidly in a broad trend channel, while long-term traders may find that the swings overlap, resulting in a slow-moving primary trend.

Transaction Costs and Slippage

What are your transaction costs per trade? Easy to calculate if your broker works on a straight percentage; a bit more difficult if you are charged a flat fee; even more difficult if you are charged a fixed fee per share.

Slippage is often underestimated. Your stops will often be filled at below the trigger price (or above when selling short) and entries often occur at obvious trigger points -- when everyone is trying to get in at the same price.

You need a rough and ready estimate of what your total transaction cost is likely to be for any given trade; even if it is just an average like 0.5% for each leg of the trade.

Consolidations

Do you want your money tied up in a stock that moves sideways for 6 months to a year? It may look great on a 10 year chart to be able to stay with a trend from start to finish, riding out the mid-point consolidation without being unseated. Bear in mind that you may end up with a negative return for a year or more, when you could have been making money elsewhere in the market. It is hard to maintain your enthusiasm for a system after several months of watching your stock becalmed. And mid-point consolidations look remarkably like tops, until they complete an upward breakout.

Realistic Expectations

Newcomers often expect to catch trends from start to finish. It is not that easy.

One of the most helpful things that anybody can learn is to give up trying to catch the last eighth -- or the first. These two are the most expensive eighths in the world. They have cost stock traders, in the aggregate, enough millions of dollars to build a concrete highway across the continent. ~ Edwin Lefevre: Reminiscences of a Stock Operator (1923)

If you try to catch the absolute start of a trend, the number of false starts is likely to be so high that it will wreck the viability of your whole system. Similarly, if you try to catch the absolute peak, in most cases you are either going to exit too early or too late; again destroying the viability of the system. Most good systems are content to leave the early and late (and most uncertain) parts of the trend alone, only trading the strongest segments.

Think of it like riding a surfboard. If you want to be the first man on the wave, you will have to paddle out further than anyone else. The problem there is that the waves aren't yet properly formed and you will spend your time fruitlessly paddling after swells without picking up a single wave. If you are content to stay closer to the shore, however, the waves are more powerful, having built up momentum, and you will enjoy plenty of good rides.

System Objectives

The challenge is to bring the above considerations together in a cohesive set of realistic objectives:

• If you are only prepared to tolerate a 10 per cent retracement, do not try to ride out major corrections.
• If you are prepared to tolerate a 10 or 20 per cent retracement, consider trading lower-risk blue chip stocks in the hope of limiting the number of times that you are stopped out in a trend. Or are you prepared to endure more stop-outs in the hope of greater returns from faster-moving stocks?
• If your transaction costs are high, you will need to limit the number of times that you are stopped out in the trend. Are you prepared to tolerate deep retracements or will you trade more reliable, slower-moving stocks?
• If you intend to trade a shorter time frame, you will need to minimize your transaction costs.
• How will you identify suitable trends, separating faster-moving and slower-moving stocks?
• How much of the early part of the trend do you need to sacrifice in order to minimize false starts? This is especially important if you are going to trade more volatile stocks.
• How will you identify and cope with shake-outs: where there are two (or even three) successive down-swings before the up-trend resumes?
• How will you cope emotionally if you are repeatedly stopped out? Will you have the resolve to stick to your system?
• How will you identify and cope with blow-offs, where price goes into an almost vertical spike? These tend to reverse sharply, leaving trend-following exits far behind.
• If you trade breakouts, how will you cope with false breaks or fake-outs?

EXAMPLES

Using the same Arc Energy as earlier, here is an example of a long-term system that may be used by a trader with high transaction costs and who is prepared to tolerate fairly deep (20%) retracements. Again, I have usedpoint and figure charts because they are less subjective. Box size is 10 with reversal amount of 2.

1. Entry is made at 20 cents on the start of a new column [1] after the pull-back respected support at [0].
2. Exit would have been made in column [2] if the retracement had continued to 40 cents (20%).
3. Stopped out at 42 cents when the retracement [3] falls below the previous low [2].
4. Re-enter at 46 cents on the start of a new column after the higher low [4].

An alternative approach, that could be employed by a trader with lower transaction costs and possibly a lower tolerance for retracements. Box size is 5 with reversal amount of 2.

1. Entry at 19 cents on start of a new column after the pull-back respects support.
2. Stopped out at 17.5 cents.
   Re-enter at 18.5 cents on completion of false break.
3. Exit at 49 cents, immediately after "long pole" (or "flagpole") with more than 10 boxes above the previous up-swing.
   Flagpoles are one method of identifying market spikes or "blow-offs".
4. Re-enter at 45 cents after double bottom.
5. Stopped out on lower low at 49 cents.
6. Re-enter at 44 cents on new high (above previous up-swing).
7. Stopped out on lower low at 47 cents.
8. Re-enter at 45 cents precedes another possible flagpole (ignored here).
9. Stopped out at 55 cents. Another shake-out.
10. Re-enter at 65 cents on higher high.
11. Another shake-out at 87.5
12. Re-enter at 90 after higher low.
13. Stopped at 175 cents after higher low.
14. Re-enter at 175 cents.

Gross trading profits are almost identical at 178.5 cents compared to 176 cents for the earlier example. But when we factor in transaction costs and slippage (at a combined 0.75% per trade) the last example falls to 169.5, because of the higher number of transactions, while the earlier example falls only slightly to 173.5 cents. As you can see, there is not much between the two approaches. And the example chosen, with a fast trend, few strong corrections and many attempted shake-outs, may flatter the first approach over the latter.
What the above examples do illustrate, though, is that, provided we are able to identify fast-moving trends, transaction costs are secondary, relative to overall system performance.

Summary

Trend traders face three major obstacles: false starts (or whipsaws), frequent shake-outs and late exits. Frequent shake-outs can present a problem even with the strongest trends. Your system should consider the time frame that you are trading; your transaction costs; your ability to endure frequent stops; and your tolerance for strong corrections. Finally, you need to decide whether you are going to trade fast-moving, highly volatile stocks (with greater risk) or slower, steadier blue chips.
In the weeks ahead I hope to answer more of the questions posed under Design Objectives (above) and compare the performance of several different trend filters and exit strategies.

Breakout Model

Stan Weinstein, in Secrets for Profiting in Bull and Bear Markets, provides one of the most complete models for trading long-term trends. The model employs a combination of proven techniques to identify breakouts from a trading range, to follow the progress of a trend and to identify appropriate exit points. It is important to read the book to understand the full model which is briefly summarized below.

Trading Ranges

The long-term cycle has four distinct phases:

• phases 1 and 3 being trading ranges or reversal patterns; and
• phases 2 and 4 being trends.

The phases are not always as easy to identify as in the above illustration: a trend may last more than a year and a reversal pattern may be over within a week. Up-trends (or down-trends) may also be interrupted by a trading range before continuing the trend.
The model uses trendlines and breakouts above resistance levels to identify the start of a new trend

Volume Confirmation

The breakout must be confirmed by higher than usual volume activity.

30 Week Moving Average

No trades may be entered if price is below the 30-week weighted moving average or if the moving average slopes downwards.

Trailing Sell-Stops

Stop-losses are moved up to below the Low of each successively higher trough in the up-trend or the 30-weekMA, whichever is the lower. See Adjusting Stop Levels for details.
If the 30-week MA starts to level out and it appears that the stock is entering a Phase 3 top, then the stops are moved up to below the bottom of each successive trough and the 30-week MA is ignored.

Exit

Trailing sell-stops account for most of the exits from the trend. Exit immediately, however, if price falls below the 30-week MA and the MA is no longer rising.

EXAMPLE

Yahoo is shown with  30-week weighted moving average.

1. Price breaks above the $3.00 resistance level [R] in June 1997. This is followed by a correction before a second breakout above the resistance which is confirmed by large volume. The entry point is marked by [E] and the 30-week MA is rising strongly.
2. Stops (depicted by  trendlines below the MA) are adjusted upwards as the trend progresses, but never above the 30-week MA as long as it is rising.
3. Price crosses below the MA at [?] but the position is not closed as the MA is still rising.
4. The position is stopped out at [X] when price falls below the previous stop level set just below $60.00

Standard Deviation Channels

Standard deviation channels are plotted at a set number of standard deviations around a linear regression line. They can be usefully applied to swing trading (as well as for detecting changes in momentum).

Visually identify a stable trend on the chart and fit standard deviation channels by dragging your mouse over the selected time period.

EXAMPLE

Sydney Gas Limited [SGL] is plotted with trend channels drawn at 2 standard deviations around a linear regression line. After equal lows in December 2005 and August 2006 SGL started to rally.

1. The signal to go long is identified when the secondary reaction (or consolidation) at [1] respects the 100-day Exponential Moving Average.
2. Accompanied by strong accumulation with Twiggs Money Flow holding above the zero line.

Mouse over chart captions to display trading signals.

Channel lines are extended using Auto-Extend (right-click on the trendline).

3. Opportunity to increase the position presents itself with another higher trough in the lower half of the trend channel
4. A further pyramid opportunity
5. Exit when price reverses below (or respects) the top trend channel
6. Another entry signal when price completes a double bottom at the lower trend channel
   (price is above the 100-day Exponential Moving Average and Twiggs Money Flow has respected zero)
7. An opportunity to increase (pyramid) your position when price respects the lower channel line.

Mouse over chart captions to display trading signals.

Trend Channel Stop Loss Orders

1. Place your first stop loss below the down-swing at [1] on which you entered.
2. On subsequent troughs, move the stop loss up to below the lesser of:
   • the lower channel line, and
   • the low of the trough.
   If the new stop level is lower than the previous stop, leave the (higher) old stop in place.
3. Even if the trough is in the upper half of the trend channel, set the stop loss at the lower channel line.
4. When price penetrates the top trend channel, move your stop loss to the upper channel line.
   If price breaks clear of the channel, keep on moving your stop up along the upper channel line each time that a new trough respects the channel line.

Mouse over chart captions to display trading signals.

5. On the second entry, place your initial stop loss below the trough
6. Move the stop loss up to below the subsequent trough, (which is also below the lower channel line)
7. The next trough is in the upper half of the trend channel, so set the stop loss at the lower channel line.

Standard Deviation Channel Setup

See Help: Trend Channels for directions on how to set up standard deviation channels.

I normally use 2 standard deviations, which enclose roughly 95% of the selected data. Using 3 standard deviations encloses about 99% of the selected data but the channel often appears too wide.

Trend Channels

Trading Forex Trend Channels

An article (Channels in the Forex Markets by Gareth Burgess) in the latest Stocks & Commodities magazine caught my eye today. It had trend channels drawn with the lower channel in a rising trend touched by only two bars and the third touch was shown as a high touching the upper channel.

I believe that the lower trendline needs to be respected (not necessarily touched) three times for it to be valid — and that the upper channel line is merely a parallel line drawn through the highs of the up-trend; it should not be used for confirmation.

Three Touch Trend Channel

Requiring three "touches" on the lower trend channel is likely to reduce the number of false starts — although it will take longer to establish the trend channel.

I suggest that you take ignore signals where price fails to follow-through below the preceding touch-low and wait for a confirming break ([break] in the above example) below the preceding low.

Standard Deviation Trend Channel

Alternatively, consider using a Standard Deviation Channel: a trend channel drawn at a set number of standard deviations around a linear regression line. Often these capture trend action more accurately than simply using the lower trendline, especially when drawn through the daily lows.

On the above channel, downward breakout would only be confirmed when the day marked [confirm] falls below the previous day's low.

Raff Regression Channel

If you want an even wider, more forgiving trend channel, increase the number of standard deviations to 2.5 or use a Raff Regression Channel. The Raff Regression Channel draws lines parallel to the linear regression line, but through the most extreme highs or lows.