Speed Resistance Lines
McClellan Oscillator and Summation Index
The "Open" Trading Index
Cycles An Introductory Picture
The theory of stock market cycles contends that stock prices move as a result of a combination of cyclical forces. Fundamental factors also influence stock prices, but their effect is generally a smooth one and unrelated to market timing.
The concept of the chart (intro.cht.1.gif) is from The Art of Independent Investing by Claud A. Cleeton. It depicts a chart which appears very much like a typical stock chart and which most analysts would consider random in character. In other words, most analysts would contend there is little, if any, predictability to the pattern. The fact is that the stock-like chart is an exact representation of a combination of the four sine waves (or cycles) and the straight line below it. It is perfectly predictable. For those who are mathematically inclined, it represents the equation y=12sinx + 4sin3x + 2sin6x + sin24x + .04x, where y represents price or the vertical axis and x represents time or the horizontal axis. No real life stock chart is, of course, a strict mathematical formula, but the principles of general predictability do apply.
Mathematical knowledge is not required to note the similarities between the results of the equation and a typical stock chart. Note that although the largest cycle doesn't bottom until 270 or letter C on the horizontal chart, the actual price bottom occurs just before 225 at point B. Note what appears to be a bottom formation between letters B and D, the formation of a bullish ascending triangle, and the dynamic upside action after the breakout of the pattern. Also note the long descending trend-line from A to B which seems to halt all rally attempts, then the ultimate move above the trend-line after the price bottom at B.
Virtually every technical formation can be explained by some cycle combination. The straight line below the chart represents a combination of fundamentals and much longer cycles. In the above formula, it is represented by the term + .04x. In some cases, it will trend upward at various angles, in others it will trend downward at various angles, and in still others, it could be horizontal. Sudden fundamental changes in a company could shift the straight line suddenly, but would not affect the underlying cycles. Examples of this would be takeover offers, new products, or unexpected earnings reports.
We do not mean to imply that only four cycles determine the movement of all stocks. In fact, there are probably scores of cycles acting simultaneously on the market, making analysis a more difficult procedure than a simple breakdown of mathematical formulas.
Needless to say, cycles in the stock market are an intricate and complex subject. In the next few paragraphs, we will attempt to impart on awareness of the nature of cycles and their effect on the market, but t is impossible to go into cycle theory in any great detail here.
Basically cycle theory contends that the major trends of stocks and stock averages are determined by fundamentals. But fundamental influences affect a stock or a stock average smoothly. The trend is sideways, up, or down at varying angles and is, of course, subject to change when fundamentals change i.e., earnings reports, new products, etc. These smooth fundamental trends are affected by cycles, and the cycles are most important for market timing because they repeat with a good degree of regularity.
The factor which makes cycle analysis most difficult is the fact that all cycles act on averages simultaneously. Some are pushing upward and others are pushing downward at any given time. The ideal buying situation occurs when many cycles bottom together or in proximity to each other.
At this point, we should emphasize that you need have no understanding whatsoever of cycles to take advantage of your subscription. Occasionally, subscribers complain that the letter is too complicated and difficult to understand. We attempt to keep it as simple and readable as possible, but keep in mind that it is impossible to treat a complex subject in grammar school fashion. We not only make mutual fund recommendations, we also discuss the cycles, the market itself and some simple technical indicators. The principal function of the service is to maximize our subscribers' stock market gains and to minimize their potential losses. That problem is confronted in our section on mutual funds and on our telephone updates. These are the only tools the subscriber really needs in order to take advantage of the service, for they tell him what and when to buy and, just as importantly, when to sell.
Included in the technical indicator area are long term weekly charts of the Dow Industrial and Transportation Averages. The weeks on these charts are numbered so that we can refer the subscriber to particular cycles discussed in the letter. These charts are included at the end of this section. Week #0 is the major Dow low of December 9-13, 1974.
There is also a long-term monthly chart of the Dow Industrials dating back to their inception in 1897. This chart can be used as a reference for long term market cycles.
The following section lists several important market cycles and refers to a chart and the numbers on that chart so that the subscriber can refer to the cycles.
INDUSTRIAL CHART CYCLES
|220.25 weeks||Ideal lows||-650||-210||+11||+224||451||671||892||1112||1333||1553|
Revised October 30, 1987 from 218.76 weeks
|206.6 weeks||Ideal lows||-633||-427||-220||-14||+194||400||607||814||1020||1227||1433||1640|
|108.75 weeks||Ideal lows||173||280||391||497||608||715||826||936||1043||1152||1261|
TRANSPORTATION CHART CYCLES
|212 weeks||Ideal lows||-636||-424||-212||0||+212||424||636||848||1060||1272||1484|
|202 weeks||Ideal lows||-615||-413||-211||-9||+193||395||595||799||1001||1203||1405|
|143 weeks||Ideal lows||-636||-493||-350||-207||-64||+79||222||365||508||651||794|
|140.875 weeks||Ideal lows||-633||-493||-352||-211||-70||+72||213||354||494||635||776|
The Cycle Indicator
|1.095 and above||Very high momentum signaling powerful and sustained market movement.|
|1.035 1.080||High risk intermediate to long term top.|
|.995 1.034||Neutral no important information given.|
|.980 - .995||Intermediate market bottom.|
|.925 965||Major market bottom capable of moving market to new all time market highs.|
|.910 and below||Major market bottom, but new all-time highs improbable on next advance.|
Through our research on market cycles, we have devised an indicator to measure one of the most helpful and accurate market cycles.
The Cycle Indicator very simply measures the average number of daily advancing issues (up stocks) on the N.Y. Exchange over a 189 day (market days) period. Taking holidays into consideration, the 189 market day span most closely measures 39 weeks. That period is used because both the 39 and 78 week cycles are paramount to our market analysis. When we first started analyzing these figures, we discovered that periods of low risk coinciding with market bottoms generally showed the some low Cl readings. Conversely, periods of high risk coinciding with market tops showed the some relatively high Cl readings.
The Cl readings were sufficient to measure the market over the few years of our initial analysis. They worked so well that we become fascinated with the idea of constructing a related indicator which could be compared over any period of market history for which we had advance-decline statistics.
We discovered in the middle to late 1970's that anytime the average daily advances over a 39 week period declined to the vicinity of 680-710, a market bottom was imminent. On the other hand, when the average daily advances reached 750-760, a market top was in prospect. The problem with the indicator arose from the fact that no comparisons could be made with the other market periods for which we had the daily advance-decline statistics. For example, in the early 1930's, a neutral day on the N.Y. Exchange may have seen 170 advances and 170 declines. Obviously the parameters given above of 680-710 average advances for a bottom and 750-760 averages advances for a top would have no significance when only 400 issues traded on the N.Y. Exchange.
For that reason, we devised the Neutral Cycle Indicator (NCI). It allows us to calculate a ratio which, in turn, allows a comparison of any two periods of market history. The NCI simply determines what a neutral day (same number of advances and declines) would consist of in a given time period. For example, on April 1, 1981, there were a total of 148,647 advancing issues over the previous 189 market days. Over that same period, there were a total of 144,441 declining issues. If we divide 148,647 by 189, we get 786 as the average number of daily advances over that 39 week period. Dividing 144,441 by 189 gives us 764 as the average number of daily declines over the some period. We can now say simply that over that 39 week period, the average day showed 786 advances and 764 declines. We can also say that on the average day over that period 1550 Issues (786 plus 764) changed in price i.e. advanced or declined. For purposes of our research we ignore the issues remaining unchanged. The final conclusion is also a simple one. Given the average number of issues changing price each day of 1550, a neutral day (equal advances and declines) over that period would be 775 advances - 775 declines (1550 divided by 2). 775 thus becomes our NCI or Neutral Cycle Indicator.
Thus, on the date of April 1, 1981, 786 was the Cl reading (average number of daily advances), and 775 was the NCI (average number of daily advances and declines on a neutral day). We can now form a ratio by dividing the Cl by the NCI. The result is 1.014 (786 divided by 775). Because the market moves cyclically, the ratio of the average number of advances (Cl) to the average of both advances and declines (NCI) fluctuates above and below the 1.000 mark. A ratio over 1.000 indicates a rising market (i.e. more advances than a neutral day). Fewer advances than a neutral day would, of course, give a result less than 1.000 and imply a declining market over the prior 39 weeks. Because ratios are used, any period in market history can be compared with any other period for which we have advance decline data.
Although there is theoretically no limit to the number of points the Dow Industrials or any other average can advance over a period of time, the computer-generated ratios have taught us that there are general limits to the swings of advances over declines (up market) and declines over advances (down market) over a 39 week period. Once these limits are reached, the market is bound to change direction. The purpose of the Cl, NCI, and CI-NCI Ratio is to measure the swings and determine the limits. It is not a short term measuring device, but an intermediate to long-term one.
The following have been the readings of important market tops and bottoms since 1932.
As a general rule, important long-term bottoms occur with readings between .925 and .965. Important market tops usually occur with readings between 1.035 and 1.080. These are the normal limits for important tops and bottoms. Intermediate market bottoms generally occur with ratio readings of .980 to .995.
These are the normal limits, but from time to time the market will go to out of limit or out of bounds readings. Out of limit readings are considered to be greater than 1.095 and lower than .910. The readings given above for tops were not the highest readings registered. Some of the interim readings were significantly higher. At first this statistic bothered us but, upon further analysis, we noted that all the uncommonly high readings occurred just after long-term market bases. In our interpretation, they were indications of extraordinary upside momentum required to drive the market into very long-term moves to the upside. For example, the highest CI-NCI Ratio reading in our data occurred in 1943 at 1.143. This extraordinary reading occurred after 10 years of base building by the advance-decline line between 1932 and 1942. It marked the very beginning of an almost uninterrupted 14-year rise in stock prices. The momentum was so great at the inception of the move that the low CI-NCI Ratio for 1943 was 1.031, the low reading for 1944 was 1.016, and the low reading for 1945 was I.051, a ratio generally associated with market tops. Again in 1950, extraordinary momentum was indicated by a CI-NCI Ratio at 1.096. This was followed by a virtually uninterrupted rise to 1956. From 1959 the all-time high point for the A-D line through mid 1983, no reading higher than 1.080 had occurred.
On May 10, 1983, the CI-NCI Ratio reached 1.102, its highest readinq in 32 years. This reading characterized a very rare period of great momentum. There was every reason to believe it had the same implications then as it had in 1943 and 1950 ( i.e. a very long term and consistent market advance). That also proved to be true, for the following 4 years saw a continuation of one of the great historic bull markets.
When out of limit moves occur to the downside, the ratio is simply telling us that downside momentum is so great that no new all-time high will follow in the popular averages at the next major top. There have been five such bottoms since 1932; 1932, 1938, 1966, 1970 and 1974. The 1932 bottom ran to a high of 196 on the Dow, considerably below the 1929 high of 386. The 1938 bottom ran to a high of 158, well below both prior highs. The 1966 bottom ran to a high of 995 in late 1968, below the prior high of 1001. The 1970 bottom ran to the Dow's all-time high of 1067 in early 1973, but this was only 6% higher than its previous high and the Dow Transportation average was nowhere near its prior highs, failing to confirm. Finally, the 1974 bottom ran to a high at 1026, again below prior highs.
Importantly, in March of 1980 and again in March of 1982 we had the first bottom readings in 20 years low enough (.947 and .926) to qualify for a major bottom and also high enough to propel the Dow to new all-time highs over the next few years. The low reading just prior to the massive August 1982 market rally was .955 on August I3.
Notice also that some important market tops have occurred with readings below the norms associated with prior tops e.g. 1973 = .975. Also note that the last three market tops have occurred at lower numbers than those seen at previous tops.
The chart (intro.cht.2.gif) depicts two waves of a theoretical 20 week cycle. In the lower half of the chart, the cycle has been moved forward in time exactly one half cycle or, in this case, 10 weeks. The dotted line, which represents the cycle moved forward, intersects the solid original cycle wave exactly half way between the bottom and the top of the cycle. Each time the solid line intersects the advanced dotted line, the solid line continues until the advance or decline equals the distance it has already traveled to the point of intersection. In other words, the point of intersection marks the halfway point for that particular cycle.
This is one of the most important techniques in our cyclic arsenal. The best way to derive these projections is to use tracing paper. Trace the chart you wish to analyze, then move the tracing paper forward in time for varying periods. By studying the crossings, information is gained about the cycles present in that particular stock or stock average. When tracing the price action, use the center or median of each day's price. The crossing points are determined by this average price line.
Daily charts should allow for each calendar day. Spaces would thus be left for Saturdays, Sundays, and holidays. Friday's median price would be connected to the following Monday's median with two spaces in between.
Study the results of moving the tracing forward 10 calendar days. Are there any crossings? Noting a recent high or low price, were there any projections? If the projections are consistently wrong, try 9 days or 11 days. Different stocks or averages will have different nominal 20-day cycles.
The nominal cycles we use for most Dow projections are as follows: 10 day, 20 day, 5 week (35-40 calendar days), 10 week (70-80 calendar days), 20 week, 40 week, 78-80 week, and 48-52 month. For cycles longer than 40 weeks, weekly charts should be used. Remember that the important periods for movement of the tracing paper are the half-spans of the above cycles. For example, for the five week cycle the tracing paper would be moved forward 17 1/2 days to 20 days.
Projections are not infallible. Other cycles than the one being investigated can add or subtract from prices at crucial times giving an incorrect intersection and an incorrect projection. If the prices cross back above or below the tracing paper before a projection is met, the projection is said to be invalidated. For those interested in a further explanation of our cycle projection techniques, refer to an article written for the Encyclopedia of Stock Market Techniques available from Investors Intelligence, New Rochelle, NY. Audio tapes and charts are also available from Dow Jones Telerate at the Internet address .
On our telephone update and in our letters we often refer to nominal cycles. Because many subscribers have inquired as to the meaning of nominal in that context, we will try to clarify the term and hopefully give you a clearer understanding of our projections.
We describe above our technique of making price projections and mention 10 day, 20 day, 5 week (35-40 calendar days), 10 week (70-80 calendar days), 20 week, 40 week, 78-80 week, and 48-52 month cycles. These are our nominal cycles. The dictionary definition of nominal is 1) being such in name only; so-called. This is precisely its definition as we use the terminology. It does not mean that there is an actual cycle of that precise length in the average or the stock being analyzed. It is used as a reference point. When a projection is given for a 10-day cycle, it does not imply that the market advance or decline will halt after reaching that initial projection. The nominal 10-day cycle projection could, if and when reached, trigger a nominal 20 day cycle projection which, in turn, could trigger a nominal 5 week cycle projection, etc. The advance or decline should finally halt when all outstanding projections have been met or, in turn, when an outstanding projection has been invalidated. We seldom give corresponding time periods for those projections to be met. Price is more important to us than time, except when a consistent time cycle bottom appears to be coinciding with a projected price bottom. There are general rules, however, for choosing approximate time periods when projected tops and bottoms should occur. The rules are based on the length of the nominal cycles. For example, a nominal 20-day cycle ideally consists of 10 calendar days up and 10 calendar days down. Theoretically, upside projections are given at the halfway point in time and price of the advance. Conversely, downside projections are given at the halfway point in time and price of the decline. In the case of the nominal 20-day cycle, any upside projection should be given, in theory, around the fifth day from the prior bottom. Once the projection is given, the implication is that the expected price should be reached in around five days.
Another way of figuring the same general rule is to divide the nominal cycle length by 4. The result shows the general time period from the projection date for the expected projection to occur. A quick example Let's assume a market bottom is hit on January 1. Five days later on January 6, a nominal 20 day cycle projection is given; because the cycle has already advanced for 5 days, its ideal path would be 5 more days of advance, then a 10 day decline. The general area to expect the price projection to be met would be January 11. An alternate rule guideline can also be used. Let's assume that the nominal 20-day cycle projection was given on January 3 after a January 1 bottom. If we used the 1/4 cycle length rule in this case, it would give us the date of January 8. The alternate rule simply assumes the cycle will ideally move 10 days up and 10 days down. Adding 10 days to the prior bottom of January 1 gives us an expected achievement date of January 11, 3 days later than the above date of January 8. In such a case, the projection would be expected between January 8 and 11. We emphasize again that these are general rules and price is more important to us than time.
The converse of the explanation of upward price projections applies to downward price projections. In the case of the alternate rule guideline, of course, days would be counted from the prior top, not the bottom, to establish an approximate date for the downward projection.
We reiterate that the fulfillment of short-term price projections is a minimum expectation and does not imply the end of the advance or decline. Such an implication comes only after all price projections are reached without setting off longer nominal cycle projections.
Speed Resistance Lines
The purpose of speed resistance lines (hereafter called SRL) is to indicate in advance where stocks and market averages are liable to find support and encounter resistance. They are useful for the short, intermediate and long term.
An SRL measures the slope of an ascent or decline. To construct an SRL, you must draw a line horizontally from the base of an up-move or the peak of a down-move. Then draw a vertical line from that baseline to the highest point of an advance, if you are measuring an up-move. Divide that vertical line by 3, placing points at the division marks (1/3 and 2/3 up the line). Finally, draw lines from the start of the move you are measuring through those points. The steeper of the lines is the rising 2/3 SRL, the less steep is the rising 1/3 SRL.
To measure the slopes of down-moves, the reverse procedure takes place. Draw a vertical line from the start of a down-move to the lowest point of the decline. Divide that line into thirds and draw lines from the origin of the down-move through the 1/3 and 2/3 points. The steeper is the declining 2/3 SRL, the less steep is the declining 1/3 SRL.
First it is important that each time a new high or new low is made, the SRL must be reconstructed.
Stocks and market averages will tend to find support at rising 2/3 SRL, and this is generally a good spot to buy. However, if the rising 2/3 SRL is violated decisively, a rapid decline to the rising SRL can be anticipated. The 1/3 SRL usually acts as a long-term support line. If the rising 1/3 SRL is violated decisively, look for a decline back to previous lows.
The opposite is true in declining markets. Rallies will frequently halt at the declining 2/3 SRL. If that line is penetrated, you can expect a rally to the declining 1/3SRL. Upon penetration of that line, a move to the old high or higher can be anticipated.
The chart shows graphically the construction of both rising and declining SRL.
Looking first at the period between May of 1970 through November of 1971, we can see the beauty of SRL. A horizontal line is drawn at the base of the market in 1970 at 627. By April of 1971, the market had topped at 958. From this point we draw a vertical line down to our base at 627. This vertical line represents 331 Dow points. Dividing it by 3 gives us 1101/3 points. We add this to 627 to get our point for the 1/3 SRL (737.3). We subtract this from 958 to get the point for our 2/3 SRL (847.7). Note that when the 2/3 SRL was decisively broken in July of 1971, this was an indication the market would decline to the 1/3 SRL. In November of 1971, after coming down almost precisely to the 1/3 SRL another major up leg began.
From January of 1973 through October of 1974, we have constructed declining SRL's. We leave the arithmetic to you, but note how, after the 2/3 SRL was penetrated, the market continued directly to the 1/3 declining SRL, where it was halted. A move above this line would have indicated the end of the bear market.
McClellan Oscillator and Summation Index
From time to time in the newsletter and on our telephone updates, we mention the McClellan Oscillator and the McClellan Summation Index, devised by Sherman and Marian McClellan. The McClellan Oscillator and Summation Index calculations require only the net advances over declines on the New York Stock Exchange as the raw data. The Oscillator is the difference between two exponential advance/decline oscillators, what the McClellans call the Ten Percent Trend and Five percent Trend. In order to calculate the McClellan Oscillator, begin by figuring the 18 day moving average (MA) of the net advances over declines (obviously a knowledge of negative numbers is important) through yesterday and call that result T1. Then calculate a 36-day MA of net advances over declines through yesterday. Call that result F1. Let's call the daily net differential between today's advances and declines =N2. The formula then becomes:
(10% Trend) T2 = 9T1 + .1N2
(5%Trend) F2 = .95 F1 + .05N2
(Today's Oscillator) O2 = T2 - F2
Summation Index = S1 + O2
S1 is yesterday's Summation Index
O2 is today's Oscillator reading
There is one other important formula for those of you who are technically oriented. Theoretically, the Summation Index can start at any arbitrary number but there is a formula to neutralize the Summation Index so that any two technicians will work with the same standards and numbers. The McClellans have chosen to neutralize the Summation Index at the plus 1,000 level. According to a letter from Marian McClellan to Ken Gammage, Jim Miekka devised the following formula: the Summation Index neutral level equals Summation Index minus Oscillator plus ten times 10% Trend minus 20 times 5% Trend. It can be used to neutralize your Summation Index to a plus 1,000 level if you do the daily calculations. The data we report in our service has been neutralized to plus 1,000.
THE OPEN TRADING INDEX
Although we did not originate the formula for the "Open" Trading Index, we did give the indicator its name. We had been researching different techniques of using Trading Index (TRIN) data. One that has proven very effective is what we call the "Open" TRIN. Rather than calculating the TRIN each day and then calculating a moving average using each day's individual reading, the individual components of TRIN are broken down over the period of the moving average used. For example, to calculate the Open 10 TRIN, you would add the last 10 days of up volume (label the result A), the last 10 days of down volume (Label B), the last 10 days of advancing stocks (label C), and the last 10 days of declining stocks (label D). The formula for the Open 10 TRIN would then be: C divided by D = X, A divided by B = Y, X divided by Y = Open 10 TRIN.
This formula gives us a volume weighted Trading Index and it has been very effective in helping to identify market bottoms. For example, from May of 1984 through September of 1985, every reading greater than 1.00 on the Open 10 TRIN was followed within 1-3 days by an important market bottom. Those parameters would not work in a bear market, of course, but we will continue to research the indicator for specific guidelines.
Another technical indicator we have been experimenting with is what we call the New 10 TRIN or Trading Index (Arms Index). It is different from the Open 10 TRIN which we named and popularized several years ago (no we didn't invent it). What we have attempted to do is take the best features of the Simple 10-day TRIN and the Open TRIN and combine them. The strength of the Open 10 TRIN is that it is in effect volume weighted and an extreme high or low daily reading of the TRIN will have little effect on the Open 10 TRIN unless there is heavy volume accompanying the reading. The strength of the simple 10 day TRIN is that very high single day readings seem to have significance regardless of volume. The New 10 TRIN is calculated by taking a ratio of 10 day down volume to 10 day up volume, then dividing that ratio by the 10 day moving average of the Trading Index. The results usually range from a low of around .60 to .70 to a high of around 1.50 to 1.70.
One other version of the TRIN we have been researching lately is what we call the Combination 10 Trading Index or Combo 10 TRIN. It is constructed by taking the average of the simple 10-day TRIN, the Open 10, and the New 10 TRIN.
Recently we have discovered that a move from below .80 to above .80 on both the New 10 TRIN and the Combo 10 TRIN, has been an effective short term sell signal, but it should be acted on only in association with other indicators.
Revised June 8, 1984 from 140.064.
Rather than provide cycle charts of shorter-term cycles, we feel that they can be more clearly explained and understood within the body of the letter itself when we deem them more important.