There are virtually an unlimited number of technical indicators, but only a few have widespread use. Technical indicators (aka chart indicators) are mathematical formulas or conditions applied to specific market data that the trader uses to try to either characterize the present market or to forecast immediate future market movements, and to generate specific buy or sell signals to enter or exit a position. Since technical indicators are based on specific formulas or conditions, they help remove the emotion from trading decisions. Indicators can be displayed in their own charts below the charts for price and volume information for easier comparison.
Indicators are also classified as to their use. A lagging indicator shows past market activity. For instance, the moving average shows what average prices were over a specified number of days. A coincident indicator shows the current state of the market, such as current volume or prices, and a leading indicator is used to forecast market activity.
Most common indicators are based on price and volume data on particular securities or the market in general over some lookback period. The lookback period is often used in smoothing market data to eliminate smaller market moves of lesser significance — so-called noise. Smoothing is achieved by calculating the average of the market data over the lookback period to display trends more clearly. Many indicators can also be modified by parameters or input variables, such as changing the lookback period or other type of information used to calculate the indicator.
The most common type of indicator that also illustrates the above mentioned qualities of technical indicators is the moving average, which is the average of security prices or the value of an index over the lookback period, which can be the prior 5, 10, 20, 50, or 200 days, or any other number of days that the trader wishes to use. Other input variables to the moving average include the price that is averaged or smoothed: high, low, open, close, or even the mean price. The moving average reduces the constant price fluctuations of the market, and the longer the lookback period, the more smoothed the data.
Trend Indicators and Oscillators
There are 2 major types of indicators: trend indicators and oscillators. Trend indicators indicate price and index trends, including short-term, intermediate-term, and long-term trends. The best known indicator of trends is the moving average.
Oscillators are indicators that show whether the indicator is above or below some average, which is usually interpreted as either an overbought or oversold condition. Oscillators are designed to show these 2 basic states; hence, they are often normalized so that they vary from 0 to 100 or from -100 to +100, to indicate an extreme price variation. If the price moves beyond a certain percentage of the oscillator range, such as 30% for the lower range and 70% for the upper range, then it has reached an extreme price range and is likely to move back toward the average.
Oscillators are used to make frequent trades to make small profits in a market that is trending sideways, what is sometimes called a whipsaw market.
Using Multiple Indicators: Crossovers, Convergence, and Divergence
Indicators are sometimes combined to generate better trading signals. A crossover occurs when 2 or more indicator lines cross over the others. For example, the beginning of a trend is indicated when a short-term moving average, such as the 10-day moving average, crosses a longer-term moving average, such as the 50-day moving average.
A convergence occurs when 2 or more indicator lines converge, or move closer, toward each other. A divergence occurs when 2 or more indicator lines diverge, or move farther apart. Sometimes, the convergence or divergence of an indicator with prices rather than with other indicators is measured. Convergence and divergence are usually used to indicate that a crossover may or may not occur, which gives the trader an earlier signal than waiting for the crossover.
False Signals, Filters, and Confirmation
Technical indicators are used to forecast future price moves; however, no indicator is foolproof, especially in a whipsaw market. To minimize false signals, where subsequent price movement differs from what was expected from the indicator, they are often combined with other tests, called filters, or with other indicators and market data to increase reliability, called a confirmation of the signal.
The most common filters can be classified in the following categories:
- Time: the signal must be present for a specified amount of time. For instance, a 5-day moving average must be above a 20-day moving average for at least 2 trading days.
- Magnitude: the signal must be in a specified range. Example: an oscillator must be more than 70% or less than 30%.
- Volume: indicators usually have more significance when they are based on higher volume, which reduces the influence of outliers which are more likely to give false signals.
Technical indicators are often chosen to comport with the trader's goals. Some indicators work better over specific time horizons, so it makes sense to choose those indicators that give good results for your investment horizon.
The frequency of trading signals is also a factor to consider. If you are a day-trader, you would want indicators that generate many buy and sell signals in a day, whereas a longer-term investor would be more interested in fewer signals. Using a combination of indicators and market data usually results in fewer signals and lower trading costs.
Fundamental Indicators and Conditional Indicators
Some people have extended the concept of the indicator to include fundamental data — such as analysts' recommendations, movement of interest rates, or trading by corporate insiders — and chart patterns. Buttressing technical indicators with fundamental analysis is certainly a wise decision, but fundamental indicators are different from technical indicators in that they do not depend only on market data, such as prices and volume, which is the foundation of technical analysis. Fundamental indicators can't be graphed and compared with prices like technical indicators.
Conditional indicators are also a different type of indicator. A conditional indicator generates a trading signal if certain conditions are met. Many brokerages and charting software can be programmed to generate trading signals based on market patterns or behavior. These trading rules can be very specific and complex. For instance, if the last 3 closing prices were at least 10% higher than the previous high, then buy. While technical indicators can and are used to generate trading signals, technical indicators do not have to be used that way. As a mathematical formula, they can be calculated and graphed independently of any trading signals. On the other hand, a conditional indicator serves no other purpose other than providing a trading signal.
Optimization and Back-Testing
Optimization is finding the best combination of indicators, parameters, and other market data to see what would have generated the most profits. This is done by back-testing the combination on historical data.
The problem with back-testing is that what worked in the past may not work in the future and it will almost certainly differ from actual trading. Although general patterns do recur, exact patterns rarely do.
The other problem with back-testing is that only trading prices are used in the historical data. However, market prices depend on the number of offers and bids at the time that the trade actually takes place and the number of orders in the queue and on how many shares you want to trade.
For instance, suppose you wanted to sell XYZ stock at the market, and the best bid price was $10 per share for 100 shares of stock, and the next best bid price was for 900 shares at $9.50 per share. If you sold 100 shares at market, you would get the $10 per share assuming that there were no orders ahead of yours. But if you sold 1,000 shares of stock, then you would sell 100 shares for $10 and 900 for $9.50, again, assuming that there were no orders ahead of yours. Hence, your average price for the 1,000 shares is lower than if you had sold only 100 shares. But what if there was a market order ahead of yours, and that person bought 1,000 shares, and the next best bid was for 1,000 shares at $9 per share. Then your shares will sell for $9 per share.
Because most historical data does not include the number of bids and offers at a particular time, and because actual prices depend on how many orders are ahead of yours, back-testing cannot be accurate, but only an approximation. Furthermore, the accuracy of the approximation cannot be ascertained. However, it is easy to conjecture that if the back test had actually been traded, then the more shares traded, the less likely it would have been at the historical price, and that the price would have been worse. For instance, in most market conditions, a market order to buy 100 shares will be at a lower per-share price than a market order for 1,000 shares, which will probably have a lower per-share price than a market order for 2,000 shares, etc. Vice versa for a sell order. In other words, actually buying or selling the stock impacts the price negatively. Simple economics — the more shares demanded, the higher the price; the more shares supplied, the lower the price.
- Back-testing will always show greater profits and smaller losses than if the actual trades had taken place.
- Commission and slippage costs will lower profits and increase losses even more, unless they were accounted for in the back test.
- Therefore, in a back test:
- If an optimization doesn't work, it won't work in actual trading.
- If an optimization is only marginally profitable, then it will probably be unprofitable in actual trading.
- If an optimization is very profitable, then it may be profitable in actual trading.
With the widespread use of computers in trading, many brokerages and financial services, such as TradeStation, provide software that can be used to create custom indicators, where you can construct your own conditions to generate buy and sell signals, then back-test the indicator on historical data. Is it wise to use custom indicators for trading?
While custom indicators are, no doubt, a great marketing gimmick for the companies that offer them, it doesn't seem likely that they will be profitable. Why?
Many question how technical analysis can even work. A common answer is that many technical analysts, using the same information, respond similarly, buying or selling together, thus increasing prices by their buying or lowering prices by their aggregate selling. So some have argued that technical analysis is a self-fulfilling prophecy.
However, a custom indicator is exactly that — custom. Since the trader developed it individually, no one else is using the indicator. So no one is following the custom signal other than the trader who developed it. If the trader is the only one buying and selling based on the custom indicator, then there is no market support from other traders that would come from using a common indicator, so there would seem to be no reasonable basis for believing that it will work other than by chance.
For the same reason, the more complex the custom indicator, the less likely it will be reliable and accurate. Of course, complexity lowers the probability that any indicator will be reliable and effective because fewer people will be following it, fewer signals will be generated, and it will generally be more difficult to interpret. Keep that in mind when choosing an indicator for your trading.