The following material introduces technical analysis and is intended to be educational. If you are intrigued, do your own reading. The answers are brief and cannot possibly do justice to the topics. The references provide a substantial amount of information.
First, the references; the links point to Amazon, where you can buy the books if you’re really interested.
- John J. Murphy
Technical Analysis of the Futures Markets - Martin J. Pring
Technical Analysis Explained - Stan Weinstein
Stan Weinstein’s Secrets for Profiting in Bull and Bear Markets
Now we’ll introduce technical analysis and explain some commonly mentioned aspects.
- What is technical analysis?
Technical analysis attempts to use past stock price and volume information to predict future price movements. Note the emphasis. It also attempts to time the markets. - Does it have any chance of working, or is it just like reading tea leaves?
There are a couple of plausibility arguments. One is that the chart patterns represent the past behavior of the pool of investors. Since that pool doesn’t change rapidly, one might expect to see similar chart patterns in the future. Another argument is that the chart patterns display the action inherent in an auction market. Since not everyone reacts to information instantly, the chart can provide some predictive value. A third argument is that the chart patterns appear over and over again. Even if I don’t know why they happen, I shouldn’t trade or invest against them. A fourth argument is that investors swing from overly optimistic to excessively pessimistic and back again. Technical analysis can provide some estimates of this situation.A contrary view is that it is just coincidence and there is little, if any, causality present. Or that even if there is some sort of causality process going on, it isn’t strong enough to trade off of.A very contrary view: The past and future performance of a stock may be correlated, but that does not mean or imply causality. So, relying on technical analysis to buy/sell a stock is like relying on the position of the stars in the atmosphere or the phases of the moon to decide whether to buy or sell.
- I am a fundamentalist. Should I know anything about technical analysis?
Perhaps. You should consider delaying purchase of stocks whose chart patterns look bad, no matter how good the fundamentals. The market is telling you something is still awry. Another argument is that the technicians won’t be buying and they will not be helping the stock move up. On the other hand (as the economists say), it makes it easy for you to buy in front of them. And, of course, you can ignore technical analysis viewpoints and rely solely on fundamentals. - What are moving averages?
Observe that a period can be a day, a week, a month, or as little as 1 minute. Stock and mutual fund charts normally are daily postings or weekly postings. An N period (simple) moving average is computed by summing the last N data points and dividing by N. Moving averages are normally simple unless otherwise specified.An exponential moving average is computed slightly differently. Let X[i] be a series of data points. Then the Exponential Moving Average (EMA) is computed by:EMA[i] = (1 - sm) * EMA[i-1] + sm * X[i] where sm = 2/(N+1), and EMA[1] = X[1].“sm” is the smoothing constant for an N period EMA. Note that the EMA provides more weighting to the recent data, less weighting to the old data.
- What is Stage Analysis?
Stan Weinstein [ref 3] developed a theory (based on his observations) that stocks usually go through four stages in order. Stage 1 is a time period where the stock fluctuates in a relatively narrow range. Little or nothing seems to be happening and the stock price will wander back and forth across the 200 day moving average. This period is generally called “base building”. Stage 2 is an advancing stage characterized by the stock rising above the 200 and 50 day moving averages. The stock may drop below the 50 day average and still be considered in Stage 2. Fundamentally, Stage 2 is triggered by a perception of improved conditions with the company. Stage 3 is a “peaking out” of the stock price action. Typically the price will begin to cross the 200 day moving average, and the average may begin to round over on the chart. This is the time to take profits. Finally, the Stage 4 decline begins. The stock price drops below the 50 and 200 day moving averages, and continues down until a new Stage 1 begins. Take the pledge right now: hold up your right hand and say “I will never purchase a stock in Stage 4”. One could have avoided the late 92-93 debacle in IBM by standing aside as it worked its way through a Stage 4 decline. - What is a whipsaw?
This is where you purchase based on a moving average crossing (or some other signal) and then the price moves in the other direction giving a sell signal shortly thereafter, frequently with a loss. Whipsaws can substantially increase your commissions for stocks and excessive mutual fund switching may be prohibited by the fund manager. - Why a 200 day moving average as opposed to 190 or 210?
Moving averages are chosen as a compromise between being too late to catch much move after a change in trend, and getting whipsawed. The shorter the moving average, the more fluctuations it has. There are considerations regarding cyclic stock patterns and which of those are filtered out by the moving average filter. A discussion of filters is far beyond the scope of this FAQ. See Hurst’s book on stock transactions for some discussion. - Explain support and resistance levels, and how to use them.
Suppose a stock drops to a price, say 35, and rebounds. And that this happens a few more times. Then 35 is considered a “support” level. The concept is that there are buyers waiting to buy at that price. Imagine someone who had planned to purchase and his broker talked him out of it. After seeing the price rise, he swears he’s not going to let the stock get away from him again. Similarly, an advance to a price, say 45, which is repeatedly followed by a pullback to lower prices because a “resistance” level. The notion is that there are buyers who purchased at 45 and have watched a deterioration into a loss position. They are now waiting to get out even. Or there are sellers who consider 45 overvalued and want to take their profits.One strategy is to attempt to purchase near support and take profits near resistance. Another is to wait for an “upside breakout” where the stock penetrates a previous resistance level. Purchase on anticipation of a further move up. [See references for more details.]The support level (and subsequent support levels after rises) can provide information for use in setting stops. See the “About Stocks” section of the FAQ for more details.
- What would cause these levels to be penetrated?
Abrupt changes in a company’s prospects will be reacted to in the stock market almost immediately. If the news is extreme enough, the reaction will appear as a jump or gap in prices. More modest changes will result, in general, in more modest changes in price. - What is an “upside breakout”?
If a stock has traded in a narrow range for some time (i.e. built a base) and then advances above the resistance level, this is said to be an “upside breakout”. Breakouts are suspect if they do not occur on high volume (compared to average daily volume). Some traders use a “buy stop” which calls for purchase when a stock rises above a certain price. - Is there a “downside breakout”?
Not by that name — the opposite of upside breakout is called “penetration of support” or “breakdown”. Corresponding to “buy stops,” a trader can set a “sell stop” to exit a position on breakdown. - Explain breadth measurements and how to use them.
A breadth measurement is something taken across a market. For example, looking at the number of advancing stocks compared to declining stocks on the NYSE is a breadth measurement. Or looking at the number of stocks above their 200 day moving average. Or looking at the percentage of stocks in Stage 1 and 2 configurations. In general, a technically healthy market should see a lot of stocks advancing, not just the Dow 30. If the breadth measurements are poor in an advancing sense and the market has been advancing for some time, then this can indicate a market turning point (assuming that the advancing breadth is declining) and you should consider taking profits, not entering new long positions, and/or tightening stops. (See the divergence discussion.) - What is a divergence? What is the significance?
In general, a divergence is said to occur when two readings are not moving generally together when they would be expected to. For example, if the DJIA moves up a lot but the S&P 500 moves very little or even declines, a divergence is created. Divergences can signify turning points in the market. At a major market low, the “blue chip” stocks tend to move up first as investors becoming willing to purchase quality. Hence the S&P 500 may be advancing while the NYSE composite is moving very little. Divergences, like everything else, are not 100 per cent reliable. But they do provide yellow or red alerts. And the bigger the divergence, the stronger the signal. Divergence and breadth are related concepts. (See the breadth discussion.) - How much are charting services and what ones are available?
Commercial services aren’t cheap. Daily Graphs (weekly charts with daily prices) is $465 for the NYSE edition, $432 for the AMEX/OTC edition. Somewhat cheaper for biweekly or monthly. Mansfield charts are weekly with weekly prices. Mansfield shows about 2.5 years of action, Daily Graphs shows 1 year or 6 months for the less active stocks. Of course there are many charts on the web. See the article elsewhere in the technical analysis section of this FAQ about free charts.S&P Trendline Chart Guide is about $145 per year. It provides over 4,000 charts. These charts show one year of weekly price/volume data and do not provide nearly the detail that Daily Graphs do. You get what you pay for. There are other charting services available. These are merely representative examples. - Can I get charts with a PC program?
Yes. There are many programs available for various prices. Daily quotes run about $35 or so a month from Dial Data, for example. Or you can manually enter the data from the newspaper. - What would a PC program do that a charting service doesn’t?
Programs provide a wide range of technical analysis computations in addition to moving averages. RSI, MACD, Stochastics, etc., are routinely included. See Murphy’s book [Ref 1] for definitions. Frequently you can change the length of the moving averages or other parameters. As another example, AIQ StockExpert provides an “expert rating” suggesting purchase or short depending on the rating. Intermediate values of the rating are less conclusive. - What does a charting service do that a PC doesn’t?
Charts generally contain a fair amount of fundamental information such as sales, dividends, prior growth rates, institutional ownership. - Can I draw my own charts?
Of course. For example, if you only want to follow a handful of mutual funds of stocks, charting on a weekly basis is easy enough. EMAs are also easy enough to compute, but will take a while to overcome the lack of a suitable starting value. - What about wedges, exhaustion gaps, breakaway gaps, coils, saucer bottoms, and all those other weird formations?
The answer is beyond the scope of this FAQ article. Such patterns can be seen, particularly if you have a good imagination. Many believe they are not reliable. There is some discussion in Murphy [ref 1]. - Are there any aspects of technical analysis that don’t seem quite so much like hokum or tea leaf reading?
The oscillator set known as “stochastics” (a bit of a misnomer) is based on the observation that a stock which is advancing will tend to close nearer to the high of the day than the low. The reverse is true for declining stocks. It compares today’s close to the highest high and lowest low of the last five days. This indicator attempts to provide a number which will indicate where you are in the declining/advancing stage. - Can I develop my own technical indicators?
Yes. The problem is validating them via some sort of backtesting procedure. This requires data and work. One suggestion is to split the data into two time periods. Develop your indicator on one half and then see if it still works on the other half. If you aren’t careful, you end up “curve fitting” your system to the data.
Article Credits:
Contributed-By: Maurice Suhre, Neil Johnson