CYCLES

← Back to Technical Analysis of the Financial Markets Overview

CYCLES

The most intriguing book I’ve ever read on the subject of cycles was written by Edward R. Dewey, one of the pioneers of cyclic analysis, with Og Mandino entitled Cycles: The Mysterious Forces That Trigger Events. Thousands of seemingly unrelated cycles were isolated spanning hundreds and, in some cases, thousands of years. Everything from the 9.6 year cycle in Atlantic salmon abundance to the 22.20 year cycle in international battles from 1415 to 1930 was tracked. An average cycle of sunspot activity since 1527 was found to be 11.11 years. Several economic cycles, including the 18.33 year cycle in real estate activity and a 9.2 year stock market cycle, were presented. (See Figures 14.1 and 14.2.)

Figure 14.1 The 22.2 year cycle of incidence of sunspots. Drought often follows two years after the sunspot minima which last occurred in the early 1970s, and is due again in the mid 1990s. In the chart, the dotted line is the “ideal” cycle, and the solid line is the actual detrended data. (Courtesy of the Foundation for the Study of Cycles, Wayne, PA.)

Figure 14.2 The 22.2 year cycle in international battles was due to top in 1982. In the chart, the dotted line is the “ideal” cycle, and the solid line is the actual detrended data. (Courtesy of the Foundation for the Study of Cycles, Wayne, PA.

Two startling conclusions are discussed by Dewey. First, that many of the cycles of seemingly unrelated phenomena clustered around similar periods. On p. 188 of his book, Dewey listed 37 different examples of the 9.6 year cycle, including caterpillar abundance in New Jersey, coyote abundance in Canada, wheat acreage in the U.S., and cotton prices in the U.S. Why should such unrelated activities show the same cycles?

The second discovery was that these similar cycles acted in synchrony, that is, they turned at the same time. Figure 14.3 shows 12 different examples of the 18.2 year cycle including marriages, immigration, and stock prices in the U.S. Dewey’s startling conclusion was that something “out there” in the universe must be causing these cycles; that there seemed to be a sort of pulse to the universe that accounted for the pervasive presence of these cycles throughout so many areas of human existence.

Figure 14.3 The 18.2 year cycles on parade. (Source: Dewey, Edward R., Cycles: The Mysterious Forces That Trigger Events (New York: Manor Books, 1973.)

In 1941, Dewey organized the Foundation for the Study of Cycles (900 W. Valley Rd., Suite 502, Wayne, PA 19087). It is the oldest organization engaged in cycles research and the recognized leader in the field. The Foundation publishes Cycles magazine, which presents research in many different areas including economics and business. It also publishes a monthly report, Cycle Projections, which applies cyclical analysis to stocks, commodities, real estate and the economy.

Basic Cyclic Concepts

In 1970, J.M. Hurst authored The Profit Magic of Stock Transaction Timing. Although it deals mainly with stock market cycles, this book represents one of the best explanations of cycle theory available in print, and is highly recommended reading. The following diagrams are derived from Hurst’s original work.

First, let’s see what a cycle looks like and discuss its three main

characteristics. Figure 14.4 shows two repetitions of a price cycle. The cycle bottoms are called troughs and the tops referred to as crests. Notice that the two waves shown here are measured from trough to trough. Cyclic analysts prefer to measure cycle lengths from low to low. Measurements can be taken between crests, but they are not considered to be as stable or reliable as those taken between the troughs. Therefore, common practice is to measure the beginning and end of a cyclic wave at a low point, as shown in this example.

The three qualities of a cycle are amplitude, period, and phase. Amplitude measures the height of the wave as shown in Figure 14.5, and is expressed in dollars, cents, or points. The period of a wave, as shown in Figure 14.6, is the time between troughs. In this example, the period is 20 days. The phase is a measure of the time location of a wave trough. In Figure 14.7, the phase difference between two waves is shown. Because there are several different cycles occurring at the same time, phasing allows the cyclic analyst to study the relationships between the different cycle lengths. Phasing is also used to identify the date of the last cycle low. If, for example, a 20 day cycle bottomed 10 days earlier, the date of the next cycle low can be determined. Once the amplitude, period, and phase of a cycle are known, the cycle can theoretically be extrapolated into the future. Assuming the cycle remains fairly constant, it can then be used to estimate future peaks and troughs. That is the basis of the cyclic approach in its simplest form.

Figure 14.4 Two cycles of a price wave. A simple, single price wave of the

kind that combines to form stock and commodity price action. Only two cycles of this wave are shown, but the wave itself extends infinitely far to the left and to the right. Such waves repeat themselves cycle after cycle. As a result, once the wave is identified, its value can be determined at any past or future time. It is this characteristic of waves that provides a degree of predictability for equity price action.

Figure 14.5 The amplitude of a wave. In this figure, the wave has an amplitude of ten dollars (from minus five dollars to plus five dollars). Amplitude is always measured from wave trough to wave crest.

Figure 14.6 The period of a wave. In this figure, the wave has a period of 20 days, which is shown measured between two consecutive wave troughs. The period could just as well have been measured between wave crests. But in the case of price waves, the wave troughs are usually more clearly defined than the wave crests for reasons that will be discussed later. Consequently, price wave periods are most often measured from trough to trough.

Figure 14.7 The phase difference between two waves. The phase difference between the two waves shown is 6 days. This phase difference is measured between the troughs of the two waves because, again, wave troughs are the most convenient points to identify in the case of price waves.

Cyclic Principles

Let’s take a look now at some of the principles that underlie the cyclic philosophy. The four most important ones are the Principles of Summation, Harmonicity, Synchronicity, and Proportionality.

The Principle of Summation holds that all price movement is the simple addition of all active cycles. Figure 14.8 demonstrates how the price pattern on the top is formed by simply adding together the two different cycles at the bottom of the chart. Notice, in particular, the appearance of the double top in composite wave C. Cycle theory holds that all price patterns are formed by the interaction of two or more different cycles. We’ll come back to this point again. The Principle of Summation gives us an important insight into the rationale of cyclic forecasting. Let’s assume that all price action is just the sum of different cycle lengths. Assume further that each of those individual cycles could be isolated and measured. Assume also that each of those cycles will continue to fluctuate into the future. Then by simply continuing each cycle into the future and summing them back together again, the future price trend should be the result. Or, so the theory goes.

Figure 14.8 The summation of two waves. The dotted lines show how, at each point in time, the value of wave A is added to the value of wave B to produce the value of composite wave C.

The Principle of Harmonicity simply means that neighboring waves are usually related by a small, whole number. That number is usually two. For example, if a 20 day cycle exists, the next shorter cycle will usually be half its length, or 10 days. The next longer cycle would then be 40 days. If you’ll remember back to the discussion on the 4 week rule (Chapter 9), the principle of harmonics was invoked to explain the validity of using a shorter 2 week rule and a longer 8 weeks.

The Principle of Synchronicity refers to the strong tendency for waves of differing lengths to bottom at about the same time. Figure 14.9 is meant to show both harmonicity and synchronicity. Wave B at the bottom of the chart is half the length of wave A. Wave A includes two repetitions of the smaller wave B, showing harmonicity between the two waves. Notice also that when wave A bottoms, wave B tends to do the same, demonstrating synchronicity between the two. Synchronicity also means that similar cycle lengths of different markets will tend to turn together.

The Principle of Proportionality describes the relationship between cycle period and amplitude. Cycles with longer periods (lengths) should have proportionally wider amplitudes. The amplitude, or height, of a 40 day cycle, for example, should be about double that of a 20 day cycle.

The Principles of Variation and Nominality

There are two other cyclic principles that describe cycle behavior in a more general sense—The Principles of Variation and Nominality.

The Principle of Variation, as the name implies, is a recognition of the fact that all of the other cyclic principles already mentioned—summation, harmonicity, synchronicity, and proportionality—are just strong tendencies and not hard and fast rules. Some “variation” can and usually does occur in the real world.

The Principle of Nominality is based on the premise that, despite the differences that exist in the various markets and allowing for some variation in the implementing of cyclic principles, there seems to be a nominal set of harmonically related cycles that affect all markets. And that nominal model of cycle lengths can be used as a starting point in the analysis of any market. Figure 14.10 shows a simplified version of that nominal model. The model begins with an 18 year cycle and proceeds to each successively lower cycle half its length. The only exception is the relationship between 54 and 18 months which is a third instead of a half.

When we discuss the various cycle lengths in the individual markets, we’ll see that this nominal model does account for most cyclic activity. For now, look at the “Days” column. Notice 40, 20, 10, and 5 days. You’ll recognize immediately that these numbers account for most of the popular moving average lengths. Even the well known 4, 9, and 18 day moving average technique is a variation of the 5, 10, and 20 day numbers. Many oscillators use 5, 10, and 20 days. Weekly rule breakouts use the same numbers translated into 2, 4, and 8 weeks.

Figure 14.9 Harmonicity and synchronicity.

Days	Weeks	Months	Years
			18
			9
		54
		18
	40
	$\overline{20}$
80
40
20
10

HOW CYCLIC CONCEPTS HELP EXPLAIN CHARTING TECHNIQUES

Chapter 3 in Hurst’s book explains in great detail how the standard charting techniques—trendlines and channels, chart patterns, and moving averages can be better understood and used to greater advantage when coordinated with cyclic principles. Figure 14.11 helps explain the existence of trendlines and channels. The flat cycle wave along the bottom becomes a rising price channel when it is summed with a rising line representing the long term uptrend. Notice how much the horizontal cycle along the bottom of the chart resembles an oscillator.

Figure 14.12 from the same chapter shows how a head and shoulders topping pattern is formed by combining two cycle lengths with a rising line representing the sum of all longer duration components. Hurst goes on to explain double tops, triangles, flags, and pennants through the application of cycles. The “V” top or bottom, for example, occurs when an intermediate cycle turns at the exact same time as its next longer and next shorter duration cycles.

Figure 14.11 Channel formation. (Source: Hurst, J.M., The Profit Magic of Stock Transaction Timing [Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1970].)

Figure 14.12a Adding another component. (Source: Hurst, J.M., The Profit Magic of Stock Transaction Timing [Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1970].)

Figure 14.12b The Summation principle applied. (Source: Hurst, J.M., The Profit Magic of Stock Transaction Timing [Englewood Cliffs, N.J.: Prentice Hall. Inc., 19701].)

Hurst also addresses how moving averages can be made more useful if their lengths are synchronized with dominant cycle lengths. Students of traditional charting techniques should gain additional insight into how these popular chart pictures form and maybe even why they work by reading

Hurst’s chapter, entitled “Verify Your Chart Patterns.”

DOMINANT CYCLES

There are many different cycles affecting the financial markets. The only ones of real value for forecasting purposes are the dominant cycles. Dominant cycles are those that consistently affect prices and that can be clearly identified. Most futures markets have at least five dominant cycles. In an earlier chapter on the use of long term charts, it was stressed that all technical analysis should begin with the long term picture, gradually working toward the shorter term. That principle holds true in the study of cycles. The proper procedure is to begin the analysis with a study of long term dominant cycles, which can span several years; then work toward the intermediate, which can be several weeks to several months; finally, the very short term cycles, from several hours to several days, can be used for timing of entry and exit points and to help confirm the turning points of the longer cycles.

Figure 14.13 (Source: The Power of Oscillator/Cycle Combinations by Walt

Bressert.)

Classification of Cycles

The general categories are: long term cycles (2 or more years in length), the seasonal cycle (1 year), the primary or intermediate cycle (9 to 26 weeks), and the trading cycle (4 weeks). The trading cycle breaks down into two shorter alpha and beta cycles, which average 2 weeks each. (The labels Primary, Trading, Alpha, and Beta are used by Walt Bressert to describe the various cycle lengths.) (See Figure 14.13.)

The Kondratieff Wave

There are even longer range cycles at work. Perhaps the best known is the approximate 54 year Kondratieff cycle. This controversial long cycle of economic activity, first discovered by a Russian economist in the 1920s by the name of Nikolai D. Kondratieff, appears to exert a major influence on virtually all stock and commodity prices. In particular, a 54 year cycle has been identified in interest rates, copper, cotton, wheat, stocks, and wholesale commodity prices. Kondratieff tracked his “long wave” from 1789 using such factors as commodity prices, pig iron production, and wages of agricultural workers in England. (See Figure 14.14.) The Kondratieff cycle has become a popular subject of discussion in recent years, primarily owing to the fact that its last top occurred in the 1920s, and its next top is long overdue. Kondratieff himself paid a heavy price for his cyclic view of capitalistic economies. He is believed to have died in a Siberian labor camp. For more information, see The Long Wave Cycle (Kondratieff), translated by Guy Daniels. (Two other books on the subject are The K Wave by David Knox Barker and The Great Cycle by Dick Stoken.)

Figure 14.14 Kondratieff’s long wave. For more information, see The Long Wave Cycle by Nikolai Kondratieff, translated by Guy Daniels (New York: Richardson and Snyder, 1984). That translation is the first ever from the original Russian text. (Copyright ©1984 by The New York Times Company. Reprinted by permission [May 27, 1984, p. F11.])

COMBINING CYCLE LENGTHS

As a general rule, long term and seasonal cycles determine the major trend of a market. Obviously, if a two year cycle has bottomed, it can be expected to advance for at least a year, measured from its trough to its crest. Therefore, the long term cycle exerts major influence on market direction. Markets also have annual seasonal patterns, meaning that they tend to peak or trough at certain times of the year. Grain markets, for example, usually hit their low point around harvest time and rally from there. Seasonal moves usually last for several months.

For trading purposes, the weekly primary cycle is the most useful. The 3 to 6 month primary cycle is the equivalent of the intermediate trend, and generally determines which side of a market to trade. The next shorter cycle, the 4 week trading cycle, is used to establish entry and exit points in the direction of the primary trend. If the primary trend is up, troughs in the trading cycle are used for purchases. If the primary trend is down, crests in the trading cycles should be sold short. The 10 day alpha and beta cycles can be used for further fine tuning. (See Figure 14.13.)

THE IMPORTANCE OF TREND

The concept of trading in the direction of the trend is stressed throughout the body of technical analysis. In an earlier chapter, it was suggested that short term dips should be used for purchases if the intermediate trend was up, and that short term bulges be sold in downtrends. In the chapter on Elliott Wave Theory, it was pointed out that five wave moves only take place in the direction of the next larger trend. Therefore, it is necessary when using any short term trend for timing purposes to first determine the direction of the next longer trend and then trade in the direction of that longer trend. That concept holds true in cycles. The trend of each cycle is determined by the direction of its next longer cycle. Or stated the other way, once the trend of a longer cycle is established, the trend of the next shorter cycle is known.

The 28 Day Trading Cycle in Commodities

There is one important short term cycle that tends to influence most commodity markets—the 28 day trading cycle. In other words, most markets have a tendency to form a trading cycle low every 4 weeks. One possible explanation for this strong cyclic tendency throughout all commodity markets is the lunar cycle. Burton Pugh studied the 28 day cycle in the wheat market in the 1930s (Science and Secrets of Wheat Trading, Lambert-Gann, Pomeroy, WA, 1978, orig., 1933) and concluded that the moon had some influence on market turning points. His theory was that wheat should be bought on a full moon and sold on a new moon. Pugh acknowledged, however, that the lunar effects were mild and could be overridden by the effects of longer cycles or important news events.

Whether or not the moon has anything to do with it, the average 28 day cycle does exist and explains many of the numbers used in the development of shorter term indicators and trading systems. First of all, the 28 day cycle is based on calendar days. Translated into actual trading days, the number becomes 20. We’ve already commented on how many popular moving averages, oscillators, and weekly rules are based on the number 20 and its harmonically related shorter cycles, 10 and 5. The 5, 10, and 20 day moving averages are widely used along with their derivatives, 4, 9, and 18. Many traders use 10 and 40 day moving averages, with the number 40 being the next harmonically related longer cycle at twice the length of 20.

In Chapter 9, we discussed the profitability of the 4 week rule developed by Richard Donchian. Buy signals were generated when a market set new 4 week highs and a sell signal when a 4 week low was established. Knowledge of the existence of a 4 week trading cycle gives a better insight into the significance of that number and helps us to understand why the 4 week rule has worked so well over the years. When a market exceeds the high of the

previous 4 weeks, cycle logic tells us that, at the very least, the next longer cycle (the 8 week cycle) has bottomed and turned up.

LEFT AND RIGHT TRANSLATION

The concept of translation may very well be the most useful aspect of cycle analysis. Left and right translation refers to the shifting of the cycle peaks either to the left or the right of the ideal cycle midpoint. For example, a 20 day trading cycle is measured from low to low. The ideal peak should occur 10 days into the cycle, or at the halfway point. That would allow for a 10 day advance followed by a 10 day decline. Ideal cycle peaks, however, rarely occur. Most variations in cycles occur at the peaks (or crests) and not at the troughs. That’s why cycle troughs are considered more reliable and are used to measure cycle lengths.

The cycle crests act differently depending on the trend of the next longer cycle. If the trend is up, the cycle crest shifts to the right of the ideal midpoint, causing right translation. If the longer trend is down, the cycle crest shifts to the left of the midpoint, causing left translation. Therefore, right translation is bullish and left translation is bearish. Stop to think about it. All we’re saying here is that in a bull trend, prices will spend more time going up than down. In a bear trend, prices spend more time going down than up. Isn’t that the basic definition of a trend? Only, in this case, we’re talking about time instead of price. (See Figure 14.15.)

HOW TO ISOLATE CYCLES

In order to study the various cycles affecting any given market, it is necessary to first isolate each dominant cycle. There are various ways of accomplishing this task. The simplest is by visual inspection. By studying daily bar charts, for example, it is possible to identify obvious tops and bottoms in a market. By taking the average time periods between those cyclic tops and bottoms, certain average lengths can be found.

There are tools available to make that task a bit easier. One such tool is the Ehrlich Cycle Finder, named after its inventor, Stan Ehrlich (ECF, 112 Vida Court, Novato, CA 94947 [415] 892-1183). The Cycle Finder is an accordion-like device that can be placed on the price chart for visual inspection. The distance between the points is always equidistant and can be expanded or contracted to fit any cycle length. By plotting a distance between any two obvious cycle lows, it can be quickly determined if other cycle lows of the same length exist. An electronic version of that device, called the

Ehrlich Cycle Forecaster, is now available as an analysis technique on Omega Research’s Trade Station and Super Charts (#Omega Research, 8700 West Flagler Street, Suite 250, Miami, FL 33174, [305] 551-9991, www.omegaresearch.com). (See Figures 14.16-14.18.)

Figure 14.15 Example of left and right translation. Figure A shows a simple cycle. Figure B shows the trend of the larger cycle. Figure C shows the combined effect. When the longer trend is up, the midpeak shifts to the right. When the longer trend is down, the midpeak shifts to the left. Right translation is bullish, left translation is bearish. (Source: The Power of Oscillator/Cycle Combination by Walt Bressert.)

Computers can help you find cycles by visual inspection. The user first puts a price chart on the screen. The next step is to pick a prominent bottom

on the chart as a starting point. Once that is done, vertical lines (or arcs) appear every 10 days (the default value). The cycle periods can be lengthened, shortened, or moved left or right to find the right cycle fit on the chart. (See Figures 14.19 and 14.20.)

Figure 14.16 The 4 year presidential cycle is clearly identified with the Ehrlich Cycle Forecaster (see vertical lines). If the cycle is still working, the next major low would be expected to occur during 1998.

Figure 14.17 The Ehrlich Cycle Forecaster has identified a 49 day trading cycle in S&P 500 futures prices (see vertical lines). The ECF estimates that the next cycle low will be formed 49 days from the last cycle low, which would be on March 30, 1998.

Figure 14.18 The ECF has uncovered a 133 day cycle in Boeing (see vertical lines). Since the last cycle low occurred during November, 1997, the ECF estimates that the next cycle low is due to occur 133 days later on June 3, 1998.

Figure 14.19a The bottoms in the cycle arcs coincide with important reaction lows in the Dow when spaced 40 weeks apart. That suggests a 40 week cycle in the Dow. The last two cycle troughs were in the spring of 1997 and the start

of 1998 (see arrows).

Figure 14.19b The daily cycle arcs reveal the presence of 50 day cycle bottoms in the Dow during the second half of 1997 and the start of 1998. The idea is to shift the arcs until their lows coincide with a number of reaction lows on the price chart.

Figure 14.20a Beginning with the major bottom in 1981, the cycle finder arcs reveal that bonds have shown a tendency to form important bottoms every 75 months (6.25 years). These numbers may shift with time, but still provide useful trading information.

Figure 14.20b Applied to this daily chart, the cycle arcs showed a tendency for bond prices to bottom every 55 trading days during this time span (see arrows).

SEASONAL CYCLES

All markets are affected to some extent by an annual seasonal cycle. The seasonal cycle refers to the tendency for markets to move in a given direction at certain times of the year. The most obvious seasonals involve the grain markets where seasonal lows usually occur around harvest time when supply is most plentiful. In soybeans, for example, most seasonal tops occur between April and June with seasonal bottoms taking place between August and October. (See Figure 14.21.) One well known seasonal pattern is the “February Break” where grain and soybean prices usually drop from late December or early January into February.

Figure 14.21 Soybeans usually peak in May and bottom in October.

Although the reasons for seasonal tops and bottoms are more obvious in the agricultural markets, virtually all markets experience seasonal patterns. Copper, for example, shows a strong seasonal uptrend from the January/February period with a tendency to top in March or April. (See Figure 14.22.) Silver has a low in January with higher prices into March. Gold shows a tendency to bottom in August. Petroleum products have a tendency to peak during October and usually don’t bottom until the end of the winter. (See Figure 14.23.) Financial markets also have seasonal patterns.

Figure 14.22 Copper usually bottoms during October and February, but peaks during the April-May period.

Figure 14.23 Crude oil prices peak during October and turn up during

March.

The U.S. Dollar has a tendency to bottom during January. (See Figure 14.24.) Treasury Bond prices usually hit important highs during January. Over the entire year, Treasury Bond prices are usually weaker during the first half of the year and stronger during the second half. (See Figure 14.25.) The examples of seasonal charts are provided by the Moore Research Center (Moore Research Center, 321 West 13th Avenue, Eugene, OR 97401, (800) 927-7259), which specializes in seasonal analysis of futures markets.

Figure 14.24 The peak in the German mark during January coincides with a lowpoint in the U.S. dollar that usually occurs at the start of the new year.

Figure 14.25 Treasury Bonds prices usually peak around the new year, and then remain weak for most of the first half. The second half of the year is better for bond bulls.

STOCK MARKET CYCLES

Did you know that the strongest three month span for the stock market is November through January? February is then weaker, but is followed by a strong March and April. After a soft June, the market turns strong during July (the start of the traditional summer rally). The weakest month of the year is September. The strongest month is December (ending with the well known Santa Claus rally just after Christmas). That information, and a whole lot more about stock market cycles, can be found in Yale Hirsch’s annual Stock Trader’s Almanac (The Hirsch Organization, 184 Central Avenue, Old Tappen, NJ 07675).

THE JANUARY BAROMETER

According to Hirsch: “as January goes, so goes the year.” The well known January Barometer holds that what the S&P 500 does during January will determine what kind of year the market as a whole will have. Another variation on that theme is the belief that the direction of the S&P 500 during the first 5 trading days of the year gives some hint of what’s ahead for the year. The January Barometer shouldn’t be confused with the January Effect, which is the tendency for smaller stocks to outperform larger stocks during

January.

THE PRESIDENTIAL CYCLE

Another well known cycle that affects stock market behavior is the 4 year cycle, also called the Presidential Cycle, because it coincides with the elected term of U.S. presidents. Each of the 4 years has a different historical return. The election year (1) is normally strong. The postelection and midyears (2 and 3) are normally weak. The preelection year (4) is normally strong. According to Hirsch’s Trader’s Almanac, election years since 1904 have seen averages gains of 224%; postelection years, gains of 72%; midterm years, gains of 63%; and preelection years, gains of 217%. (See Figure 14.16.)

COMBINING CYCLES WITH OTHER TECHNICAL TOOLS

Two of the most promising areas of overlap between cycles and traditional technical indicators are in the use of moving averages and oscillators. It is believed that the usefulness of both indicators can be enhanced if the time periods used are tied to each market’s dominant cycles. Let’s assume that a market has a dominant 20 day trading cycle. Normally, when constructing an oscillator, it’s best to use half the length of the cycle. In this case, the oscillator period would be 10 days. To trade a 40 day cycle, use a 20 day oscillator. Walt Bressert discusses in his book, The Power of Oscillator/Cycle Combinations, how cycles can be used to adjust time spans for the Commodity Channel Index, the Relative Strength Index, Stochastics, and Moving Average Convergence Divergence (MACD).

Moving averages can also be tied to cycles. You could use different moving averages to track different cycle lengths. To generate a moving average crossover system for a 40 day cycle, you could use a 40 day moving average in conjunction with a 20 day average (one-half of the 40 day cycle) or a 10 day average (one-quarter of the 40 day cycle). The main problem with this approach is determining what the dominant cycles are at a particular point in time.

MAXIMUM ENTROPY SPECTRAL ANALYSIS

The search for the right dominant cycles in any market is complicated by the belief that cycle lengths aren’t static; in other words, they keep changing over time. What worked a month ago may not work a month from now. In his book, MESA and Trading Market Cycles, John Ehlers uses a statistical approach called Maximum Entropy Spectral Analysis (MESA). Ehlers explains that one of the main advantages of MESA is its high-resolution measurement of cycles with relatively small time periods, which is crucial for shorter term trading. Ehlers also explains how cycles can be used to optimize moving average lengths and many of the oscillator-type indicators we’ve already mentioned. Uncovering cycles allows for the dynamic adjusting of technical indicators to fit current market conditions. Ehlers also addresses the problem of distinguishing between a market in a cycle mode versus one that is in a trend mode. When a market is in a trend mode, a trend-following indicator like a moving average is needed to implement trades. A cycle mode would favor the use of oscillator-type indicators. Cycle measurement can help determine which mode the market is currently in, and which type of technical indicator is more appropriate to use for trading strategies.

CYCLE READING AND SOFTWARE

Most of the books referred to in this chapter on cycles can be obtained through mail order firms like Traders Press (see reference in previous chapter) or Traders’ Library, P.O. Box 2466, Ellicott City, MD 21041, [800] 272- 2855). There’s also a lot more software to help you perform cycle analysis with your computer. The Ehrlich Cycle Forecaster and Walt Bressert’s CycleTrader are both available as add-on options to run with charting software provided by Omega Research. Bressert’s CycleTrader integrates the concepts he describes in his book, The Power of Oscillator/Cycle Combinations. (Bressert Marketing Group, 100 East Walton, Suite 200, Chicago, IL 60611 (312) 867-8701). More information on the MESA computer program can be gotten from John Ehlers (Box 1801, Goleta, CA 93116 (805) 969-6478). For ongoing cycle research and analysis, don’t forget the Foundation for the Study of Cycles.

INTRODUCTION

The computer has played an increasingly important role in the field of technical analysis. In this chapter, we’ll see how the computer can make the technical trader’s task a good deal easier by providing quick and easy access to an arsenal of technical tools and studies that would have required an enormous amount of work just a few years earlier. This assumes, of course, that the trader knows how to use these tools, which brings us to one of the disadvantages of the computer.

The trader not properly schooled in the concepts that underlie the various indicators, and who is not comfortable with how each indicator is interpreted, may find him- or herself overwhelmed with the vast array of computer software currently available. Even worse, the amount of impressive technical data at one’s fingertips sometimes fosters a false sense of security and competence. Traders mistakenly assume that they are automatically better simply because they have access to so much computer power.

The theme emphasized in this discussion is that the computer is an extremely valuable tool in the hands of a technically oriented trader who has already done his or her basic homework. When we review many of the routines available in the computer, you’ll see that a fair number of the tools and indicators are quite basic and have already been covered in previous chapters. There are, of course, more sophisticated tools that require more advanced charting software.

Much of the work involved in technical analysis can be performed without the computer. Certain functions can be more easily performed with a simple chart and ruler than with a computer printout. Some types of longer

range analysis don’t require a computer. As useful as it is, the computer is only a tool. It can make a good technical analyst even better. It won’t, however, turn a poor technician into a good one.

Charting Software

Several of the technical routines available in charting software have been covered in previous chapters. We’ll review some of the tools and indicators currently available. We’ll then address some additional features such as the ability to automate the various functions chosen by the user. In addition to providing us with the various technical studies, the computer also enables us to test various studies for profitability, which may be the most valuable feature of the program. Some software allows the user, with little or no programming background, to construct indicators and systems.

Welles Wilder’s Directional Movement and Parabolic Systems

We’ll take a close look at a couple of Welles Wilder’s more popular systems, the Directional Movement System and the Parabolic System. We’ll use those two systems in our discussion of the relative merits of relying on mechanical trading systems. It will be demonstrated that mechanical trend following systems only work well in certain types of market environments. It will also be shown how a mechanical system can be incorporated into one’s market analysis and used simply as a confirming technical indicator.