Introduction
Momentum indicators in general measure the velocity of a price move. Momentum indicator charts provide supporting evidence to the signals given by price-volume charts. The momentum indicators that will be discussed in this paper are:
1. Rate-of-Change(ROC)
2. Relative Strength Index(RSI)
3. “Moving Average Convergent Divergent” Method(MACD)
4. Stochastics
1. Rate of Change(ROC): Rate of Change compares today’s price of a security with the price(of the same security) of a prior period of specified number of days. For example, a 10-day ROC is calculated by comparing the current day’s price with the price 10 days before. For comparison, subtraction or division is used. When both prices are same, in the case of subtraction, the difference is zero. In this case, zero line is the equilibrium line on the chart. When current price is higher, ROC is positive and when current price is lower ROC is negative.
2. Relative Strength Index(RSI): RSI is calculated for a specified period. For example, 14-day RSI is calculated using prices of 14 consecutive days.
The formula for calculating RSI is as follows:
RSI = 100 – [100/(1+RS)]
RS equals the average of the closes of the UP days divided by the averages of the closes of the DOWN days. The maximum value of the RSI can be 100 when there is no DOWN day in the period. The minimum value is going to be zero when there is no UP day in the period.
3. Moving Average Convergent Divergent Method (MACD): In this method, the difference between a shorter period EMA (Exponential Moving Average) and a longer period EMA is termed as MACD. For example, in a method represented by MACD(12, 26, 9), the 26-day EMA is subtracted from the 12-day EMA. 9-day EMA of MACD is also calculated and plotted. This line is termed as the Signal line or Trigger line. In general one should own a stock only when the MACD line is above the Trigger line.
Important Lines: MACD line : 12-day EMA – 26-day EMA
Signal Line: 9-day EMA of MACD
4.Stochastics: The Stochastic method has a %K line and a %D line.
The formula for the %K is:
%K = 100[(C-Ln)/(Hn-Ln)]
In this equation C is today’s close, Ln the lowest low for the last n days, and Hn is the highest high over the same n day trading period.
The formula for the %D is:
%D = 100(H3/L3)
In this expression, H3 is the 3-period sum of (C-Ln) in the %K calculation, and L3 is the 3-period sum of (Hn-Ln).
%D is also calculated as 3 period moving average of %K.
Lane, the person who developed this indicator, strongly emphasizes that there is only one valid signal that can be obtained from the %D. This is a divergence between %D and the price of the security being monitored.
Weight of Evidence Approach
In technical analysis we are concerned with identifying trend reversals at a relatively early stage. We assume that the new trend will continue and we attempt the riding of that trend until it reverses. The operating assumption is always that the prevailing trend is in force until the weight of the evidence proves otherwise.
We must use several indicators to determine the direction of a trend. When a majority of them are in agreement, we can be more confident that the trend has indeed reversed. .
Different Trends
The most widely followed trends are short, intermediate, and long. These last approximately 3-6 weeks, 6-39 weeks, and 1-2 years. Most of the material later in this chapter will focus on short, and to a lesser extent, intermediate trends (i.e., trends lasting 3-6 and 6-39 weeks, respectively). Nevertheless, it is still of paramount importance for any trader to gain some understanding of the current position of the long-term trend. Just as the unwary swimmer finds it difficult to swim against the tide, the short-term trader will certainly encounter problems if he is swimming against the main trend. Time and again we find that trend-spotting systems often will lead traders to make money-losing decisions based on erroneous signals. These are short and intermediate-term price trends that are swimming against the tide of the primary trend.
Originally posted in Knol http://knol.google.com/ k/ momentum-indicators-oscillators
Principles of Momentum Interpretation
The methods of interpreting momentum indicators can be divided approximately into two broad categories. The first category deals with overbought and oversold conditions and divergences called “momentum characteristics.” The second category deals with interpretive principles that measure reversals in the momentum trend itself, the assumption being that when momentum changes direction prices will follow sooner or later.
Trend-determining techniques such as trend line violations, moving average crossovers, and the like can be applied to momentum as well as to price.
Momentum typically reverses along with price, often with a small time lag. However, just because oscillators change direction doesn’t always mean that prices will also. Normally, a reversal in the momentum trend is evidence of a price-trend reversal signal. In effect this momentum signal acts as a supplementary “witness” in our weight-of-the evidence approach to determine the validity of a trend reversal.
Take special note of this fact: Actual buy and sell signals can come only from a reversal in trend of the actual price, not from a reversal in the momentum series.
Overbought and Oversold Levels
All momentum series have the characteristics of an oscillator as they move from one extreme to another. These extremes are known as “overbought” and “oversold” levels. In my seminars I often equate these zones with leash attached to an unruly dog taking a walk. The animal continually strains at the leash moving from one side of the side walk to the other. Market momentum works in a similar way.
Some indicators, such as the RSI are calculated in such away that they have finite extremes above or below which the momentum cannot go. In these cases there is an established “default” level for the overbought and oversold lines. For the most commonly used 14-day time span, these lines are drawn at 70 for overbought and 30 for oversold.
Other indicators such as the rate-of-change (ROC), have no such theoretical boundaries, at least on the upside. This means that we must judge for ourselves where to establish the levels for overbought and oversold signals. These lines should be drawn so that the space between them includes the bulk of the trading activity. In this case, try to think of the oscillator as a rubber leash that from time to time will be stretched beyond its normal length. What we must do is find the metaphorical equivalent for the end of the leash, that means the point that include most of the rallies and reactions in the market under study.
The technical interpretation of overbought and oversold lines is that they represent an intelligent point for anticipating a trend reversal. An overbought condition is where you should consider taking profits or reducing your exposure. For example, if you are holding three gold contracts and the price rallies to where it generates an overbought reading, you might wish to take some partial profits. Even though the trend may continue, or the price has failed to cross below its moving average, or it violates a trendline the overbought reading by itself indicates that the odds of a reversal have increased. If the risks of a top have grown, then it makes sense to reduce your exposure. If press stories concerning the bullish nature of the security are beginning to emerge, and your emotions are telling you to buy more, use these signs as further confirming evidence that it is a good time to begin to decrease rather than increase your exposure.
On the other hand, if you believe that the main trend is down and you have been waiting for a short-term bounce as a time to sell, an overbought reading is as good a time as any. For the same reason, it would normally be a grave mistake to even consider making a purchase when an oscillator signals an overbought condition. The problem with this interpretation is that this is precisely the time when most people have the urge to buy, because rising prices attract optimism, positive news stories, and bullish sentiment.
The opposite is true for an oversold condition. Few people want to buy after prices have been falling and the news is inevitably discouraging. Unfortunately, this is the time when we need to control our shaking hands, pick up the phone and call our friendly broker. This is also the moment when we should overcome (at all costs) the temptation to take a short position. In fact, the correct action is to cover part of any outstanding short positions. At the time, you may think that it is possible to make more money by holding on to your investment, but, believe me, taking some partial profits will put you in a far more objective frame of mind when that inevitable rally gets underway.
When the momentum indicator moves through its overbought or oversold level and then re-crosses it on its way to the equilibrium level, a good buying or selling entry can be made with confidence.
The importance of an overbought/oversold reading will depend on the time frame under consideration. For example, if the period used in constructing the indicator is five days, the implications from extreme readings will be nowhere near as profound as those from a momentum indicator spanning twelve months.
Oscillators that move in the direction of prevailing market trends tend to move to a greater extreme and stay there longer than those that move against the trend. In the case of a bull trend reactions are almost always reversed at the oversold line or sometimes even before the oscillator reaches that point. This trait itself is a primary characteristic of a bull market.
Divergences
When price and momentum are moving in the same direction, they are said to be “in gear.” There is nothing important to be learned from this state of affairs except that the trend is healthy. However, when momentum does not confirm the price, beware: The prevailing trend may be about to reverse. The conflict between momentum and price is known as a “divergence.” It is called a “negative divergence,” when rising prices are supported by weaker and weaker underlying momentum.
In one respect markets are like houses. They take longer to build than they do to tear down. Markets spend most of their time advancing rater than declining. This means that the lead-time characteristics of momentum indicators are usually more pronounced at market peaks than at troughs.
Divergences also occur at market bottoms where they are called “positive,” because momentum hits bottom before price does. In this instance, the technical position is said to be “improving” or getting “stronger”. Indeed, if you think a market is in the process of reaching its bottom and you do not see a divergence, you may want to reconsider your analysis, because most market bottoms are preceded by at least one positive divergence.
It is very important to note that although they indicate either a deteriorating or an improving market-condition, divergence in and of themselves do not signal that the prevailing trend has reversed. That signal can come only from some kind of trend-reversal sign generated by the price itself. This cue could take the form of a price-pattern completion, a moving–average crossover, or some other signal. When this occurs technicians say that the divergence has been “confirmed” by the price.
Another sign of a mature trend occurs when the momentum index moves strongly in one direction but the price fails to follow through with any degree of gusto. This indicates that the price is tired of moving in the direction of the prevailing trend: for despite the strong momentum thrust, prices are unable to respond. This is an unusual but nevertheless powerful phenomenon.
Significance of a Divergence
Generally, the more the divergences that occur, the greater their significance.
The time period separating the divergence is also important. Usually, the greater the time-span between the peak in momentum and the peak in price, the greater the significance.
The type of trend being monitored raises the final point concerning the length of time separating divergences. If a trader is analyzing short-term price movements he would expect the divergences to take place over the course of as week or so at most. On the other hand, an investor is primarily concerned with the primary trend, so he would look for divergences associated with a momentum graph constructed from an intermediate time frame. In this case the divergences in an intermediate oscillator are obviously more significant than the divergences in a short-term momentum series.
It should be clear, that the length of the time span separating divergences is a function of the trend itself. In other words, when there are two, three, or four divergences within one short-term trend, their primary significance rests in their relation to the next short-term trend. Their secondary importance relates to the type of trend under consideration. For example, divergences between intermediate trends have significance for the next primary trend, whereas divergences in short-term momentum are important for the next intermediate trend and so on.
Our final point concerning the importance of divergences concerns the level at which the last divergence takes place. As market peaks, rallies in a momentum indicator that are barely able to move above the zero level are often followed by a very sharp decline. This is one of the few instances in technical analysis when a clue hints at the character of the next move. I must stress that such instances are not always followed by a sharp drop. Remember: Technical analysis is far from perfect. However, in most cases when weak momentum of this nature is confirmed by a trend-break in the price, be on your guard for a larger-than-normal sell-off.
The Divergence Trap
Most of the time divergences proceed in a fairly orderly way. Times, just as you expect the price to drop, a final rally develops out of the blue. Normally, this advance will push the momentum indicator back above at least one of the two peaks (see example given in the text), causing the wary trader to surrender his bearish sentiment. Typically, this latest rally will prove to be a “divergence trap” after which the price will then fall in the manner previously expected. This final burst will probably result from some unanticipated news event that causes short covering. When the short-covering ends, there is very little to support the price and down it goes.
Complex Divergences
Price trends are determined by the interaction of many different time cycles. Most momentum indicators, however select only one cycle, since they are constructed using a specific time span. One way to solve this problem is to overlay two different momentum indicators constructed from two different time spans and compare them. Normally, both series will move in a broadly similar direction. A divergence in the indicator signals an impending trend reversal.
There are several factors to consider in analyzing complex divergences.
- It is important to compare two time spans that are separated by a long interval. For example, it makes sense to compare two oscillators based on a 10- and 20-day time span for a short-term trend because the indicators are separated by a substantial time span.
- The peak in the longer-term indicator must be substantial in relation to the shorter one. For example, if you are comparing a 13-week oscillator with one constructed from a 26-week time span, the latter should achieve a new high lasting at least six months or longer.
- The oscillator with the shorter time span must be at the equilibrium level or close to it when its longer-term counter part is peaking.
- Complex divergences also occur at market bottoms, but in this case all the conditions are reversed. In other words, the indicator with the longer time span should be registering a bottom of relatively long duration as its shorter-term counterpart is rallying close to the equilibrium level.
- Perhaps most important of all, complex divergences must be confirmed by a reversal in price.
Summary - Principles of Momentum Interpretation
- Momentum is a generic term that comprises a number of different indicators. It measures the rate at which prices rise and fall, often giving advance warning of latent strength and weakness in a specific trend.
- The principles of interpretation apply to all types of momentum indicators to some degree or another.
- Momentum interpretation can be broken down into three primary areas: overbought/oversold, divergences, and momentum trend analysis.
- Extreme readings and momentum divergences do not in and of themselves represent actual buy and sell signals these can come only from a trend-reversal in the price itself. Momentum characteristics do, however, emphasize the significance of price signals when they are given.
PRINCIPLES OF MOMENTUM INTERPRETATION – TRENDS
Momentum, like price, moves in trends. This means that the techniques used for analyzing price trends can be used for appraising momentum trends. Despite this fact, we must still keep in mind that a trend reversal in momentum is usually, but not always, associated with a similar reversal in the price. Occasionally, an analysis of an oscillator trend will accurately tell us that momentum has reversed, and a reversal in price may indeed follow. However, the lag between the signal of a reversal in momentum and the actual turning point in price may be so great that trading decisions based on this signal will be unprofitable.
Trendlines
It is possible to construct a trendline joining several momentum bottoms or tops. When the trendlines are violated, the trend in the momentum is reversed. When both trendlines in price and momentum are penetrated, the market usually reverses trend or, at the very least, consolidates for a while. In effect, the momentum signal represents additional confirmation in our “weight of the evidence’ theory of trend reversals discussed earlier.
Measuring Objectives
In some instances when dealing with trend line violations in price, it is possible to come up with measuring objectives. This concept can also be applied to momentum analysis.
Since the reliability of this technique is somewhat limited it should never be used in isolation as basis for making a forecast.
Price Patterns
One of the basic price-reversal techniques used in technical analysis is price formations, or price patterns. These same techniques also can be applied to momentum analysis. Price formations do not occur very often in oscillators, but when they do it’s time to sit up and pay attention, because the formations carry great portent for future price activity.
It is very important to understand that there is usually a lag between the point when the price formation in the momentum indicator is completed and the actual reversal in the price. The momentum series nearly always gives an advance warning that the underlying technical picture is improving or deteriorating.
In effect, the rule is telling us to use a little common sense, and not to blindly assume that every price-pattern breakout will result in a reliable move. As a general guideline, I tend to ignore upside price breakouts in momentum that occur much above the equilibrium line and downside breakouts that develop much below it. In conclusion, price patterns are a fairly rare occurrence in oscillator series. Provided they are not formed near an overbought or oversold zone, it certainly pays to respect them.
Peak and Trough Analysis
The central building block of technical analysis is the concept that a rising trend consists of a series of rising peaks and troughs and a falling one of declining peaks and troughs. When the prevailing trend of rising tops and bottoms is broken and the price experiences a lower peak and lower trough, it is assumed that a new downward trend is underway, and vice versa. The significance of this new trend will depend directly on the nature of the rallies and reactions. If they are of an intra-day variety, the reversal will be very short-term in duration. On the other hand, if the peaks and troughs are associated with rallies and reactions lasting six weeks or more, then it will be of a primary-trend in nature.
Peak and trough analysis also can be applied to momentum indicators. I should add that although peak and trough analysis is a legitimate technique for both price and momentum analysis, it appears to be more reliable for the former. In most instances trendline analysis is the most reliable method for identifying momentum trend reversals.
Advance Breakdowns and Breakouts
Occasionally, an indication of an impending trend reversal in price occurs when the momentum indicator breaks a series of rising peaks and troughs, but the price indicator does not. What often happens in such instances is that the next rally in price proves to be the final one for that specific trend. I term these momentum failures “advance breakdowns,” because they represent very subtle warnings that the trend in momentum has reversed.
Moving Averages and Momentum Indicators
Because raw momentum indicators are often quite jagged and seemingly random affairs, the practice of smoothing them with moving averages has evolved. This makes it easier to get a better sense of the underlying momentum trend.
There are four ways to use moving averages to greater advantage in momentum analysis. They are: (1) double moving average crossovers, (2) a change in direction of a moving average, (3) a predetermined bench mark crossover, and (4) the construction of a moving average derivative.
Double Moving average Crossover
Signals are generated when the short-term average crosses above and below its long-term counter part.
Moving Average Directional Change
If the moving average itself is considered, it is apparent that a change in its direction offers a promising approach for providing reliable and timely signals for momentum-trend reversals
Overbought/Oversold Crossovers
Overbought and oversold levels can be established and can be used to generate signals. The levels are established on a trial and error basis with the specific security being monitored. The buy and sell indications are given when the average moves through one of the extremes and then recrosses it on its return journey to the equilibrium level.
Equilibrium Crossover Signals
Buy alerts are given when the indicator(moving average) moves above zero and sell signals when it moves below zero. Unless the selection of a moving average is made very carefully there is a chance that most of the signals could be quite late. It is also possible to use this technique for unsmoothed oscillator data but unless the time span is chosen very carefully, you will find that there will be substantial number of whipsaws.
Derivatives of Moving Averages
This process involves a “double smoothing,” which is a moving average of a moving average. It is even possible to take the process a step further by smoothing the result for a third time. There is no end to the possibilities, but remember, there is no Holy Grail. It is usually best to keep things as simple as possible,
Research is the key. Examine several markets over a long period of time – 2-3 years of daily data, 10-20 years of weekly figures, and 20-30 years for monthly statistics-to make sure that the method you have chosen works in practice.
DETAILED COVERAGE OF MOMENTUM INDICATORS
Rate of Change (ROC)
Introduction
Rate-of-change (ROC) is probably the easiest momentum indicator to construct. Many people believe that a mathematically simple formula is inferior to a complex one because of its simplicity. The facts, however, disprove this belief.
Construction
Rate-of-change compares today’s price with yesterday’s. For example, a 10-day ROC is calculated by comparing the price today with the price of 10 days ago. The result is then plotted as continuous series that oscillates above and below the equilibrium level.
Scaling for the ROC takes one or two forms. In the first method, equilibrium line is plotted at zero. Positive values appear as +1, +2, etc., while negative values are represented as minus numbers. The alternative is for the equilibrium level to be plotted as 100 and positive and negative numbers to appear as percentages. Readings in excess of 100 indicate a rising trend, and those below it indicate a negative one. From an interpretive point of view it is immaterial which method of scaling is used because the general movements are identical. I prefer to use positive and negative numbers rather than the percentage method, since this gives a better sense of bullish and bearish tendencies. The comparison in the ROC calculation can be done either by subtraction of division. My personal preference is to use the division method, because the division calculation gives you a sense of proportion.
Overbought and Oversold Levels
The ROC indicator lends itself handsomely to overbought and oversold interpretation. The problem is that there are no hard-and- fast rules about where the lines should be drawn, since the magnitude of the oscillations will vary according to the volatility of the underlying security and time span being considered. Generally, the longer the time span, the greater the higher and lower readings, or extreme, in the oscillator.
For this reason overbought and oversold lines are constructed on the basis of judgment. Wherever possible it is important to place them equidistant from the equilibrium level. This is because fear and greed tend to move in proportion and should be represented graphically to reflect this fact.
Short-Term Trends
Excluding hourly intraday trends, short-term market movements typically extend 3-6 weeks can be as short as 5 days and be as long as 45 days. Since this covers a fairly wide range, there are several possible time frames that can be chosen. The most popular are spans of 5, 10, 12, 14, 25, 28,and 30 days. The 5-day ROC is used to monitor extremely short-term price movements lasting between 7 and 10-days. The 12-day span is used for trends extending between 15- and 25- days. I prefer a 10-day span because it encompasses two weeks of normal trading data. The 25, 28 and 30-day spans are used for short-term trends in excess of 25 days.
A reason why it makes good sense to compare several ROC time spans is that some series may be giving extremely strong signals of an impending trend reversal while others may not. If the analysis is limited to just one oscillator measuring one time span, the overall picture may be missing a key element. However if several time frames are considered it may well turn out that one of the oscillators is in the process of completing a price formation, a quadruple divergence, or a very significant trend line break. Whatever the signal, if it is a significant one, it will give us a more explicit indication that the prevailing trend may be about to reverse.
Intermediate-term trends
Technical lore has it that intermediate trends range in length from as short as three weeks to as long as six months, sometimes a little longer. I like to use ROC time spans of 6, 13, and 26 weeks. The intermediate time frame can also be expanded to include a 39-week ROC, but this length of time is better utilized to monitor long-term trends that are below average duration. Calendar quarters are useful, I believe, because they reflect the seasonality of the year and therefore the dominant cycles that influence intermediate trends.
The same concept of comparing several ROC indicators of varying time spans discussed previously can be applied to intermediate a analysis.
Long-Term trends
Most people understand the long-term trend to mean a primary bull or bear market. This implies a cycle that extends for approximately four years from trough to trough. In fact, the 4-year cycle, which has been modeled on the U.S. Business cycle since the early part of the 19th Century, is 41 months.
The most commonly used long-term time frame is a 12-month ROC. The annualized rate-of-change is a useful measure because it eliminates all seasonal variations. I have found that the 39-week (9-month) time span is often a very useful interval to follow. Some analysts have argued that this is an important emotional period because it is the length of the natural cycle in a woman’s pregnancy. In many instances, I have found that this 9-month period brings out momentum characteristics that are superior to those highlighted by a 26 or 52-week ROC. Commonly used long-term ROC spans also include 18 and 24 months. Multiples of annualized rates-of-change in the form of 36 and 48-month spans occasionally supplement the analysis.
The idea of monitoring several indicators based on different time spans is just as relevant to long-term trends as it is short and intermediate trends.
Trend line Construction
ROC indicators lend themselves to trendline construction as much as any other momentum indicator. It is better to construct trendlines at a less extreme angle of descent, provided it has been touched or approached on at least two occasions. These same principles would hold in reverse following an unusually severe selling climax.
The joint penetration of ROC and price trendlines often gives valuable buy and sell signals. Always remember that a price trend violation often occurs well after the momentum violation. More reliable signals seem to develop when the trendline for either price or momentum is violated simultaneously with the completion of a price formation.
Price Patterns
With the possible exception of the RSI, price formations probably occur more in ROC oscillators than in any other indicator. A completion pattern that develops near an overbought level in a rising trend is far less likely to work than if it takes place at an oversold or even neutral level. The opposite is true for a reversal in a declining market.
The principle of “commonality” can also have an important effect on price patterns. The commonality principle states that the more a specific characteristic can be observed, the greater the significance when the trend reverses. For, example if the ROC indicator on one stock in an industry group was tracing out a head-and-shoulder formation, this would imply either that the company itself was heading for trouble or that the whole industry was in for bad times. On the other hand, if the majority of the stocks in the industry were showing the same signs, there would be little doubt that this particular sector was due for correction.
We could take it one step further and state that if a huge number of equities in the stock universe were experiencing “toppy” momentum then the stock market itself would be vulnerable.
Divergences
Divergences represent an integral part of ROC analysis. Divergence occurs when the price makes a new high or new low for the move, but the oscillator-after one negative or positive divergence-is barely able to rally above or slip below the equilibrium point.
Technical analysis rarely gives us clues to the character or steepness of an impending move, but divergences do just that. For this reason, it is well worthwhile to watch for them, because when they materialize and are also confirmed by a trend break in the price series, expect an unusually powerful move to follow.
Complex Divergences
Complex divergences work quite well with the rate-of-change concept. It is important to make sure that them time spans for the two oscillators are sufficiently far enough to reflect two different cycles and that when a complex divergence appears, it is also confirmed by some kind of reversal in the price trend.
______________
______________
Moving averages of ROC Indicators
The ROC lends itself very easily to the smoothing process. The volatility common to short-term time spans makes these oscillators subject to a significant amount of whipsaw crossovers. A better approach is to use the moving average itself as a signaling device. Examples: Changes in the direction of the 12-month moving average of a 12-month ROC can be used as the signaling points. A 12-day moving average of a 12-day ROC and an 8-day moving average of both the Moving Average and ROC are used. Signals are generated when the shorter-term averages crosses above or below the longer-term one. A 12-month ROC smoothed with a 6-month moving average.
A variation of this approach involves an additional smoothing. The advantage of the double-smoothing is that it filters out several false or temporary changes in the direction of the unsmoothed version. The disadvantage is that the signals appear at a later date and can occur well after the signal you are waiting for. This drawback becomes a major disadvantage if this approach is being used as part of a mechanical trading system in conjunction with a signal confirming a reversal in the price trend. On the other hand, if the chosen combination offers reliable but sometimes untimely signals, just use the timely one for the purpose of making a trade or an investment. This method will not, by definition, be able to offer a suitable exit point, since you will not know whether the countervailing signal will be timely or late. However, if we work on the assumption that the timely entry point will be followed by a worthwhile price move, it should be a fairly simple matter to use some other method to signal a good time to take profits (for example profit protection stops).
This idea of “smoothing” an ROC indicator at least once opens up a whole series of possible variations. These include choosing time spans for specific rate-of-change indicators, picking various time spans for the smoothing, and, of course, deciding whether to smooth once, twice, or three times, and so forth. Obviously, the potential combinations are innumerable. I would add, though, that it is a wise philosophy to keep things as simple as possible. Most people believe that they will not be successful unless they try to out as many possibilities and combinations as they can. In the days when the calculations were done with pen and paper these possibilities were limited, but in this age of powerful personal computers, it is very easy to toy with a thousand variations looking for the perfection that I guarantee you does not exist. In my view the better approach is the simple and thoughtful one; Spend your time and effort on studying and analysis.
THE RSI INDICATOR
The RSI indicator, commonly known as the “Relative Strength Indicator,” was introduced by Welles Wilder in his 1978 book New Concepts in technical trading.
RSI Calculation
The formula for the RSI is as follows:
RSI = 100 – [100/(1+RS)]
RS equal the average of the closes of the up days divided by the average of the closes of the down days. In the standard calculation originally presented by Wilder, the time span for the total was set at 14. Since the introduction of the RSI, traders have experimented with a number of alternative time spans, thus creating the very popular 9- and 22-day periods.
Pros and Cons of the RSI
The RSI has several advantages over a simple ROC calculation. They do not make this method a better momentum indicator, because the ROC has some benefits that the RSI lacks. Nevertheless, the RSI offers many advantages:
- It is less volatile than the ROC. Because the RSI takes the average of the up and down days. Its results are less affected by a sharp dip or rise on a specific day. As result, this method tends to be more stable momentum indicator than the ROC calculation.
- The extreme levels of RSI are set at 100 and 0, although in practice these extremes are rarely attained. With the RSI it is possible to gauge whether one security is more volatile than another by comparing their up and down movements. A more important benefit that flows from this characteristic is that it is much easier to establish universal standards for the overbought and oversold benchmarks. Using the 14-day default, these levels are traditionally set at 30 for oversold and 70 for overbought.
Interpreting the RSI
Wilder lists five basic interpretations of the RSI. They are as follows:
- Tops and Bottoms: Tops occur when the indicator moves above 70, bottoms when it falls below 30. This, of course, is another way of expressing the overbought and oversold characteristics of described earlier. Since momentum typically turns ahead of the price, these “tops’ and “bottoms” often give advance warning of a strengthening or deterioration in the underlying technical structure.
It is important to remember that the longer the time span in the RSI calculation, the shallower the swing in momentum. Of course, the opposite is also true. Consequently, the 70/30 combination is inappropriate when the span is either shorter or longer than the standard 14-day period. In the case of 5-day RSI, 80/20 combination gives a much better feel for the overbought/oversold extreme than the 70/30 default value. On the other hand in case of 30-day RSI, a more appropriate point for the overbought/oversold line appears to be 65 and 35.
- Chart Formations: The RSI is one of the few indicators than can be used to chart pattern construction, Such configurations do not appear to form as often as in the ROC indicator, but they nevertheless represent a useful addition to the RSI analysis. In New Concepts in Technical Trading, Wilder points out that the RSI is subject to “head-and-shoulders tops or bottoms, pennants or triangles often show up on the index.”
- Failure Swings. Failure swings occur at both tops and bottoms. The accepted wisdom is that failure swings are most significant after the RSI has moved through an overbought or oversold level. Once the peak reaches overbought level or higher, a reaction sets in. A failure swing occurs when the next rally fails to surpass its predecessor and the second reaction pushes the RSI below the previous low. The second, or failing rally can take the form of one large movement or several small ones. The key to determining whether a failure swing has been signaled lies in the fact that the second rally does not exceed the first. Failure swings also occur at bottoms where the exact opposite set of conditions appear. This would involve a decline blow the oversold level, a subsequent rally, a successful ‘test” of the previous low, and finally a rally that takes the index above the previous high. Generally, the more extreme the reading at the time of failure swing, the grater its significance. This failure-swing principle implies, that it is perfectly correct assumption that peak-and-trough analysis can be applied justifiably to RSI interpretation.
- Support and resistance. Horizontal support and resistance areas can also be applied to momentum. Wilder emphasizes that the support and resistance lines constructed from the index often correspond to trend lines drawn on the price series. I do not consider support-and-resistance characteristics to be particularly helpful in charting momentum.
- Divergence. One of the most useful functions of the RSI is to point out divergences between the price and momentum.
- Trendlines. Experience in the market-place demonstrates that the concept of RSI trendline construction is an unquestionably valid approach to plotting momentum. Arguably, it is the most important of all.
- RSI and Moving Averages. It is possible to extend the idea of moving-average crossovers to the RSI. A moving average serves to smooth the RSI calculations, making it a more reliable alternative. The raw data is then discarded and the moving average is used to trigger buy-and sell alerts. For example, the 14-day RSI of the gold price has been smoothed with a 10-day simple moving average. Signals are given when the average reverses direction from an extreme zone and crosses back on its way towards the 50 level, or level of equilibrium.
Equilibrium Crossovers
In the Encyclopedia of Technical indicators, Colby and Meyers tested various overbought/oversold combinations. They based their research on weekly periods using several decades of stock market data, but they were unable to find any combination of time span and overbought/oversold level that resulted in significant profits.
Given what we have learned about the characteristics of bull and bear market momentum, these results are not surprising. Technical analysis is an art, not a science. That means it is very important to consider the manner in which an indicator moves into an overbought or oversold zone rather than treat the signal as a mechanical device requiring no interpretation.
Indeed, Colby and Meyers found that the best results came from a simple equilibrium (50) crossover: Buy when above 50, sell, when below. The best time spans were clustered on either side of the 20 weeks, with the 21-week span providing the highest profit of all. Of the 75 trades 21 of them, or 28%, produced a profit.
Comparing RSI Time Spans
Comparison of different spans can be useful from two aspects. The first deals with perspective. For example, a 14-day RSI gives you little inclination as to the direction or maturity of the main trend. On the other hand, a 12-month RSI often can warn of an imminent turning point. Short-term traders entering a long position in the market are better able to do so when both the short- and long term RSIs are in an oversold condition. As discussed earlier, if a trader is going to make a mistake it is much easier to do so when going against the primary trend. Monitoring the action of a long-term RSI will not provide all the answers by any means, but it will certainly help to warn a trader that the prevailing trend is very mature and that perhaps the odds are against a short-term position.
The second point of time-span comparisons revolves around the fact that at any juncture there are a number of different cycles simultaneously operating on the price. If we just consider a 5-day RSI, we completely overlook any information that might be gleaned from other time spans. In some cases it might be possible to spot a price pattern in one period that does not show up in another. Trendlines and divergences are additional characteristics that may appear more prominently in some series than in others.
TREND DEVIATION AND THE MACD INDICATOR
So far the discussion has been limited to indicators that are constructed from a comparison of the current price to a previous one. Another possibility is to relate the current price to some form of trend measurement. This concept works on the assumption that while prices move in trends, they do not move in straight line. Rather, they fluctuate around that trend. These fluctuations form the basis for trend-deviation momentum oscillators.
Trend Deviation and a Moving Average
The simplest form of calculating a trend deviation involves the relationship between the current price and a moving average. The oscillator is constructed by comparing the latest price by the average. Calculating a trend-deviation oscillator can be done using either subtraction or division. For very short-term trends there is very little difference, but for longer-term price movements the division calculation is much preferred because it is better reflecting any proportionate price movements.
Quite simply, when the price and the average are identical, the oscillator is plotted at zero. When it is above the average, the momentum series is in positive territory, and vice versa. Zero equilibrium crossovers, therefore, indicate when the price crosses above and below its moving average. This approach has an advantage over the ROC method because zero crossovers, by definition, also offer a price trend reversal signal. In effect, simple trend deviation oscillators represent both momentum and trend reversal signals in one indicator.
Moving Averages
A “moving average” is a technique that helps to smooth out or eliminate random day-to-day fluctuations in price. In effect, it is the technician’s way of trying to graphically come up with a mechanism that can represent the underlying trend.
In their book, The Encyclopedia of Technical Market Indicators, Robert Colby and Thomas A,.Meyers point out that in the 19 years leading up to 1980 there was no really significant difference between simple, weighted, and EMA crossovers when tested for a range of time spans (1-75 weeks) using weekly data. During that period the best weighted average crossover (69 weeks) turned in the best points profit, 118 points in the S & P Composite. This result is to be compared with 112 and 11 for a 42 – week EMA and a 45 – week simple moving average, respectively.
Always use indicators you feel comfortable with and have confidence in. If you lack confidence in your indicators, you will have no staying power when the markets turn against you.
One of the best ways to gain faith in an indicator is to test it to your own satisfaction. Don’t take my word for it, prove it to your self first; after all, it is you who will be losing money if you are wrong. If the indicators you use worked reasonably well in the past and are not overly complex, they are likely to help you in the future.
The most simple trend-deviation method is to compare the closing price to a moving average.
The MACD Indicator
MACD stands for “Moving Average Convergent Divergent” method. This method is simply another way of expressing a trend-deviation oscillator. The system obtains its name from the fact that the two moving averages used in the calculation are continually converging with and diverging from each other. Normally, the two averages are calculated on an exponential basis.
The zero line represents those periods when the two EMAs are identical. When the MACD is above the equilibrium line, the shorter average is above the longer one and vice versa. A moving average of the MACD is used as the signal line. It gets this name because MACD crossovers of signal line generate buy and sell signals. In my experience I have not found these crossovers to be particularly reliable, and I regard them as overrated. I prefer to use the MACD from the point of view of trendline violations, or even price-pattern construction. Another possibility is to use the MACD signal-line crossovers as an alert that some other oscillators based on a longer time span may be poised to give a signal.
Obviously, an MACD can be constructed from many different combinations. Gerald Appel of Signalert, arguably its chief proponent, has done a substantial amount of work on the indicator. He recommends a combination of 8-, 17-, and 9-day EMAs, but he believes that sell signals are more reliable using a 12-25-9-day combination.
STOCHASTICS
The “Stochastic Indicator,” invented by George Lane (Investment Educators, P.O.Box 2354, Des Plaines, IL), measures the relative position of the closing price within a given interval. The period is normally a daily one, but a stochastic indicator can be constructed for any time period as long as high, low, and closing data are available.
The stochastic method rests on the assumption that prices tend to close near the upper part of the trading range during an uptrend and near the lower part during a downtrend. As the trend approaches a turning point, the price closes further away from its extreme (i.e., away from their daily high in a rising market and from the daily low in a declining one). The objective of the Stochastic formula is to identify these points in an advancing market when the closes are clustered nearer to the lows than to the highs, since this indicates that a trend reversal is at hand. For down markets , the process is reversed.
The indicator is plotted in the form of two lines, known as “Percent D” and “Percent K.” In a June 1984 article in Technical Analysis of Stocks and Commodities, Lane explains that he experimented with 28 different oscillators each using a different letter of the alphabet. It happened that “D” and “K” turned out to be the best.
The %K is the more sensitive of Lane’s two oscillators, but it is the %D line that carries the greater weight and gives the major signals. The formula for the %K is as follows:
%K = 100{(C-Ln)/(Hn-Ln)]
In this equation C is today’s close, Ln the lowest low for the last n days, and Hn is the highest high over the same n day trading period. For short-term trading purposes Lane recommends that n should be 3. A 5-period %K line has also become quite popular. In the march 1991 edition of Technical Analysis of Stocks and Commodities, editor Thom Hartle uses a 14-day span and makes the point that some traders extend the period as far as twenty-eight days.
The %D line is a smoothed version of %K. Its calculation uses the following formula:
%D = 100[H3/L3]
In this expression, H3 is the 3-period sum of (C-Ln) in the %K calculation, and L3 is the 3-period sum of (Hn-Ln). In a sense you could equate the %D line with a 3-period smoothing of %K.
The Slow Stochastic
The Slow Stochastic is a smoothed variation of the regular series. In this calculation the original %K line is eliminated and the old %D substituted. This renamed or “slowed” %K is then averaged by three days to form the %D slow. The resulting indicator is less volatile and subject to whipsaws.
The Stochastic indicator, therefore, takes the form of two oscillators, The %K is usually plotted as a solid line, and its slower %D counterpart is expressed as a dashed or dotted line. When plotted, the Stochastic indicator always falls in the range of 0 to 100, like the RSI. A reading near 80 is generally regarded as overbought and 20 as oversold.
Interpretation
In the article referred to earlier, Lane strongly emphasizes that there is only one valid signal that can be obtained from the %D. This is a divergence between %D and the price of the security being monitored. All other signals, he points out, are merely “guideposts or warnings that an important signal is near.” My view is that all signals from any momentum indicator should be used as an alert or warning of an impending trend reversal. Remember that momentum measures velocity, not price trends. Price measures price trends. Momentum signals are useful primarily because they emphasize the importance of a price trend-reversal signal. As such, they can and should be used in a prudent way for part profit taking or partial entry of a new position, never as the sole basis for a major decision. In most cases a good momentum signal, such as a %D divergence, will be associated with a trend reversal. However, in a large number of “inconvenient” situations the price continues to make one more high or low, bringing on a “stopped out” position.
Divergence
The main difference between a Stochastic and an RSI divergence is that there are usually fewer divergences using the Stochastic Indicator. In fact, it is probably true to say that in the vast majority of cases, the %D experiences only one or, at the most, two divergences.
According to Lane, the stochastic signal requires action when the %K crosses from the right-hand side of the peak in the %D line. Buy signals are triggered when the %K crosses the right-hand side of the low point of the %D line at market bottoms. In this interpretation the left-hand crossover occurs before the turning point in the %D line, and the right-hand crossover occurs after it. Lane emphasizes that the right-hand crossovers are more reliable.
The Hinge
The “hinge” is a slowing down in the velocity of either line. This implies a reversal in the next trading period, which is the next day for daily data, the next week for weekly data, and so forth.
Warning
In the case of a rising market, a warning occurs when the %K line has been rising for a while, and then one day (or week or month, depending on the time frame being used) reverses sharply. In these instances a sharp reversal is defined as one that falls in the range of 2% - 12%. This represents a warning that only one or two more days of rising movement are likely prior to a reversal in trend.
%K reaching an Extreme
Normally, when an indicator reaches an overbought or oversold extreme it indicates a possible trend reversal. However, when the %K line moves to the extreme of 0%, Lane emphasizes that this signals “pronounced weakness.” Typically, from there %K will bounce up to 20% to 25% from the zero level, later falling back toward zero again. He stresses that the odds are extremely good that this testing process will, in fact, take place. Lane estimates that it normally takes 2-4 days(or weeks or months) for the testing process to run its course, after which a small rally should be expected.
The opposite is true of rising markets in relation to the 100% extreme. Lane goes out of his way to underscore the point that the initial 0% and 100% readings in the %K do not signal a top or bottom in the price. In fact, he goes on to point out that they mean the “exact opposite.” At most, they indicate that a slight pause or hesitation in the prevailing trend will develop, shortly to be followed by a resumption of that trend.
The Set-Up
In a rising market the “Set-Up” occurs as the price makes a low simultaneously with the %D. Both series then go on to make new highs, but on the subsequent reaction the %D breaks below its previous low while the security does not. The implication is that the next rally in price will probably turn out to be an important top. This is known as a “Bear Set-up.”
Failure
Rising and falling peaks and troughs reflect the underlying technical strength or weakness in a market. Lane’s Stochastic “failure” is no exception to this principles. At market bottoms, failure occurs when the %K line crosses above the %D and then falls back for a couple of days while still managing to remain above the %D line. It represents a kind of test that if successful, indicates that the new uptrend is likely to continue.
Trendlines
Clearly, the Stochastic method works well with divergence and moving-average crossover analysis. However, it is not normally suited for trendline construction.
How Doest It Test?
In Encyclopedia of Technical market Indicators Colby and Meyers tested the Stochastic Indicator for the effectiveness of a number of rules and parameters, and found that the vast majority of combinations were unprofitable. This research was based not only on 1977-86 weekly data for the NYSE, but also on the experience of Schwager and Strahm (Technical Analysis of Stocks and Commodities, July 1986). Colby and Meyers finally discovered a profitable combination, although it did not compare favourably with most of the other indicators they tested. The rules they established called for a buy signal when the 39-week unsmoothed %K line (n= 39 with no moving average) crosses above 50%, and the %K and closing price are both above their previous closing levels. They decided to both sell and sell short when these conditions are reversed (i.e., when %K moves below 50% and the %K and closing prices are below their previous week’s levels.) They note that fairly consistent profits were achieved for periods (in other words, the n value) ranging from 38 to 66 weeks, but they reported that the 39-week span proved to be the most profitable period used. Note that the %D was not used in this particular rule.
How to Use the Stochastic Effectively
The Stochastic is most effective when you keep daily series in conjunction with weekly and monthly. In this way the daily indicator monitors the short-term trend, the weekly indicator the intermediate term, and the monthly the long-term trend.
Trading in the direction of the main trend is very important. Trading in the direction of the main trend means that buy signals in the daily and weekly stochastics should be ignored if the monthly series is in a topping-out of or declining phase. If the decline in the monthly Stochastic is very mature and showing signs of bottoming, and if this condition is confirmed by other long-term characteristics of a primary trend reversal, buy signals in the daily and weekly series will probably result in a profitable trade.
Investors can reverse this process by using the two shorter series for more precise timing of primary low. For example, every bear market ends when the first short-term rally in the new bull market takes place. Consequently, if the monthly series is showing signs of bottoming, wait for the intermediate and short-term Stochastics to reverse direction to the upside. The working assumption is that the strength in the implied short or intermediate rallies will be sufficient to shift the balance for the monthly series as well.
Source for the knol.
This Knol is a summary on the topic that I prepared based on
Martin Pring on Momentum
Martin J. Pring, 1994
Published by Vision Books Pvt. Ltd., New Delhi in arrangement with Probus Publishing Company.
___________________________________________________________________________________________
Knol Number 600 top 150 knol
