Figure 5.4 conventionalinterpretation of oscillator indicators can issue many signals against the major trend. in this illustration the strong rally is interpreted by stochastic as an overbought condition resulting in the generation of several unprofitable short trades. note the profitable buy signals created as the mid-to late-day correction approaches exhaustion. chart created with TradeStation® 2000iby omega research,inc.

oscillator-type indicators in trending markets. Oscillator indicators are by definition unable to exceed a peak value, in this case 100. In a strongly uptrending day such as this one where new highs are regularly being plotted, this limit is either closely approached or actually reached on a frequent basis. Therefore a system that receives its directions from an oscillator indicator such as this is easily tricked into identifying a strong market, such as the one depicted in Figure 5.4, as overbought. Subscribing to the overbought theory, the system then attempts to sell into the teeth of an extremely strong rally. The result is an unacceptable string of losing trades that would quickly discourage even the most seasoned trader and send him or her scurrying back to the drawing board in search of a more reliable trading approach.

Figure 5.5 illustrates the reaction of our demonstration system under market conditions that have created a mild, although persistent, downtrend for the day in question. This chart is of the same stock issue (CMVT) using the same two-minute data compression as before. This activity occurred a few days previous to the charts used earlier.

Note that although the initial buy signal issued by the system

STOCHASTIC

Figure 5.5 In this instance stochastic has identified a downtrending day as oversold, issuing buy signals that turn into losing trades. note that once again the oscillator has properly identified the exhausted correction against the major downtrend, issuing a sellsignalshortly before the noon hour. chart created with TradeStation® 20001 by omega research, inc.

could have been exited for a smallprofit had one acted quickly, the remainder of the buy signals issued during what turned out to be a down day in the market would have in all probability resulted in losses for at least four of the last five entries. Again, as in the previous chart, note that our system has generated a trading signal that could have produced a significant gain. This short position could take a good profit much more easily than the other signals on this chart, as it is taken in the direction of the major trend of the day. The sell signal that was issued shortly after 11:30 a.m. came at the exhaustion point of the corrective rally, which occurred slightly past the halfway point of the trading session. In Chapter 10, we will identify strategies that will enable the trader to identify this trend early enough to profit from these types of trades while avoiding the losing trades on the buy side.

As mentioned before, staunch believers in oscillator trading have made, and will make again, vigorous attempts to create a profitable strategy from oscillator indicators by working with the variables involved. A long bias can be built into the system simply by raising both the overbought and oversold lines, thereby making it much easier for the system to buy than to sell. The opposite case can also be constructed,

making it easier to sell by lowering both lines. Unfortunately, building in a long or short bias will only make things worse when the system encounters a persistent trend opposite the built-in bias.

One can also vary the results of the system by altering the time compression of the underlying chart and making adjustments to the length input of the indicator, and ultimately the system. Unfortunately, none of these adjustments to the basic structure of an oscillator system can overcome the inherent weakness of such a strategy-the propensity to aggressively sell a rallying market and just as aggressively buy a rapidly falling market.

RELATIVE STRENGTH INDEX

The Relative Strength Index (RSI), created by Welles Wilder many years ago, has stood the test of time and is still one of the most popular, widely used technical indicators available. Like the stochastic indicator discussed earlier, it is also an oscillator-type tool whose calculated values are restricted to the familiar zero to 100 range.

RSI differs from stochastic in that it attempts to measure the market more in relation to its recent movements rather than making its comparisons to the recent highs and lows. It measures market strength and weakness relative to its recent price activity-hence its name, Relative Strength Index. Being more front-weighted in its calculations, it can give a better velocity reading than other trading tools. Since the calculations of RSI are heavily dependent on only the absolute relative positions of the most recent close and the next most recent close, this indicator is less affected by sharp rises or drops in the price of the underlying security. Thus, RSI has the effect of filtering out some of the noise, or random price activity, in the underlying market.

Briefly, RSI is calculated first by comparing the most recent close to the previous closing price. If the price of the most recent close is higher than the preceding close, then the value of the close is added to the "up amount"; if it is lower, the "down amount" is incremented by the value of the recent close. The totals for the up days and down days are then averaged, using the length input for the averaging factor. The length input reflects the number of bars over which the RSI calculations are to be performed. The up day average is then divided by the down day average; the resulting relative strength (RS) average is used to arrive at the value for RSI.

Most oscillator indicators will display increasing volatility in their plots as the length input is decreased, thus reaching the extremes of the 100-point range much more often. RSI seems to be more sensitive to these input changes than other similar trading tools. Therefore, since RSI is more sensitive to shorter length settings, traders may wish to concentrate on length inputs for this tool that are greater than 5 bars.

Those using varying settings of RSI to set their final trade entry parameters must realize the effect that a lower length input will have on their system. Many more trades will result from such use of an RSI indicator. Traders whose personal style dictates a more aggressive, frequent trading approach will find this RSI feature quite helpful. Although many more trades will be generated, it will be necessary to filter out a number of the trading signals to maintain the level of accuracy necessary to create a profitable strategy. One of the most effective methods to filter out such trades is to use multiple settings of the same indicator, using a slower setting as a qualifier of sorts. This concept is discussed in further detail in Chapter 6, "Multiple Indicator Sensitivity Settings,

Figure 5.6 illustrates the conventionally applied RSI indicator. This chart, for comparison purposes, is identical to Figure 5.1, the chart used to demonstrate the properties of the stochastic indicator in the previous section of this chapter. Note that in this case the overbought and oversold lines are placed at 70 and 30, respectively Although this seems to be the most popular conventional setting, these lines may be varied up to 85 or down to 15 depending on the degree of sensitivity desired for the application at hand. This setting will vary considerably with the data compression of the underlying market, the length input being used, and the frequency of trade generation desired by the end user.

Note that a single line in this case represents the output of the RSI indicator as opposed to the two plots used by stochastic. Although this is the conventional application, it is also possible to plot varying lengths of the RSI indicator on the same chart and trade the crossovers of these two lines in much the same manner as was demonstrated with stochastic.

The usual interpretation of RSI is to generate a buy signal when the indicator first finds itself below the oversold line and then crosses above this line. The buy is taken at the point at which the line actually crosses above the oversold designation. Sell signals arise when the RSI crosses back below the overbought line after first passing into overbought territory. As with the sell signal, the order is placed

" lm«Ji.-f>Litrnn . - (CMVT) CumviMSC I <r<;hnt)lmw Inc \.ahv-2 r

CMVT LAST- 2 min

I, .........IA..-",,,,:.,

.... .LfMti, ...

!,•%t l,\v-,ln ., -

96.000 94 000

92.000 -90.000

Figure 5.6 Relative Strength Index issues trading signalsin much the same

manner as stochastic. Note in this case that the overbought threshold is

placed at70 while the oversold level is at 30, whichisthe more conventional

placement of these lines.

Chart created with TradeStation® 2000i by Omega Research, Inc.

when the RSI actually crosses below the line representing the overbought threshold.

Again, for demonstration purposes, I am using an automated trading system that is designed to indicate exactly where the buy and sell signals are generated by our RSI indicator. Figure 5.7 illustrates the placement of these trades. Note that the downward-pointing arrow designating a sell signal appears just as the RSI plot, which has first risen into overbought territory, crosses below our overbought threshold line.

Three buy signals are created in a short period of time as the RSI plot wanders above and below the oversold line as the market is preparing for the next rally. Signals are actually given when the RSI line crosses above the oversold line after being in the oversold territory for at least one bar. Since, for demonstration purposes, our system is set to signal each buying opportunity that arises, several buys appear here in rapid succession. In reality, only one signal would be taken from this group for actual trading. Systems operating in their normal mode would issue a single buy order at this point. The numbers printed below the upward-pointing arrows identify successive signals on the chart.

CMVTLAST-2 mtn 10.12f200C C=B9000 -29.625 -2*.97% RSI

ti"-" S.iul,

RSI-

99.000 96.000 94.000 92.000 90000 88.000

RSI<CloseI7,30-70.black-black)2813 30.00 overbought-

10/12 B:S2 9:12 932

52 10:12 10:32

Figure 5.7 In our example, RSI issues a sellsignalwhen the indicator plot crosses down through the overbought line and a buy signal when the plot crosses up through the oversold line. Arrows are placed on the price bars to further define the timing ofthesetrades. Chart created with TradeStation® 2000i by Omega Research,Inc.

Our automated system has issued the first two buy signals in this sequence from the appropriate crossovers that are not as apparent to the casual observer as the final crossover in the group of three. The fact that the first two crosses are not as evident as the last is of little consequence, though, because even those observing these chart patterns without the aid of a computerized system could still recognize the final triggering event and be able to enter the long position in time to realize a gain from this trade. Although this position would have been entered effectively, the point remains that even the most minor of crossovers are legitimate as valid entry signals.

This point is further emphasized by the next sell signal on our chart. Notice that the RSI plot does not venture far into overbought territory before turning south to generate a new short position. If one is to use these tools without the aid of an automated system, close observation is necessary for proper trade placement.

As before, no money management stops or trailing stops of any nature are included in the system. Also, the system is designed to take all trades generated and will often enter several consecutive trades in the same direction. Thus, a buy signal is created each time the RSI line crosses above the oversold threshold line, and a sell sig-