back start next

[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [ 10 ] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52]


nal is created each time the line crosses below the overbought threshold. As before, this is a demonstration vehicle more than a system that would be actively traded.

To demonstrate the behavior of the RSI-based system, we will use the identical chart that was used to demonstrate the same feature of the stochastic system.

Figure 5.8 displays the trading of this system for the day in question. The area of the chart used to detail the buy and sell signals in the preceding figure is highlighted by the gray rectangle.

Note the fairly accurate placement of the respective buy and sell signals on our chart. Since this was a relatively sideways day for CMVT, we would expect the signals from any oscillator-based system to be fairly accurate, as the plots of the indicator in question rarely run very far beyond the bounds placed by our overbought and oversold territories.

Figure 5.9 is a duplication of Figure 5.3, the October 18, 2000, chart of CMVT that was used to demonstrate the buy and sell signals generated from our stochastic system during a relatively trendless trading session.

Comparing the two charts one will note that, although there are

!I(d deS,alon rha(l " ICMVT Comveise Technology Ire LA5f-2

ASS&JSSSSSKm °="5.760 119.750 L=111.563 W186S100

Classic RSI Signals Nontrending Day

......... j-----r


t?b»i, .,...,i«-4, ./..........



9G.00C 96.000 .00 92.000 90.000 S8.000

Figure 5.8 Duriag relntivelytrendless dnys, oscillntor indicators can be quite effective. Note the buy and sell signnls issued on this chart courtesy of the RSI indicntor.

chart created with TradeStation® 20001 by omega research, inc.

Figure 5.9 The upward-pointing arrows on this cbartdefinetbe placement of buy signnls, while the down arrows designnte selling points as taken by the RSI system. Once agnin note the relntively successful trades generated by an oscillator indicntor on n sidewnys, trendless dny.

chart created nlth TradeStation® 2000l by omega research,inc.

certainly some differences in the placement of buy and sell signals, generally the trades are appearing about as one would desire. Although some of the trades were entered a bit early on both the buy and sell sides of this market, most of the trades would have generated a positive result in a rather short period of time. Again the emphasis from both of these charts is to illustrate that, on such sideways, trendless days, oscillator-based systems can be relatively effective trading strategies.

Figure 5.10 demonstrates the activity of the identical system on an entirely different type of trading day.

Using the same uptrending chart from the trade in CMVT on October 19, 2000 (Figure 5.4), we again are able to observe the significant limitation of an oscillator system during a trending phase of the market. It would have been very difficult if not impossible to extract any kind of positive result from any of the first 11 trades generated on the sell side of an uptrending market. Conversely, note that again our system was able to identify an almost perfect series of entries at the conclusion of the market correction that occurred near the end of the noon hour. These trades, taken in the direction of the major trend of the day, are certainly preferable.

Figure 5.10 The signals on the chart demonstrate the limitation of oscillators to give reliable trades during strongly trending days. Note also that the RSI system has successfully placed buy orders as the correction against the major-trend is completed shortly before 1:00 p.m. Chart created with TradeStation® 2000i by Omega Research, Inc.

Now lets look at the RSI system on a relatively downtrending day (see Figure 5.11). For purposes of comparison we will use the same chart that was used to demonstrate stochastic activity under the same conditions (Figure 5.5).

In much the same manner as the stochastic-based strategy, £hei£§I formula has identified several areas on the chart as being oversold. With this determination in place, buy signals are issued in an attempt to get on board an expected uptrend. Unfortunately for the system., there is a downtrend in place for most of the day that causes the system to take multiple losses from its long positions. Obviously, the short positions, taken in the direction of the major trend of the day, are able to extract significant profits from their designated entry points on the chart.


Percent R, developed by Larry Williams and first published in his 1979 book How I Made One Million Dollars Last Year Trading Commodities, remains one of the most popular oscillator indicators in use today. In what by now is a familiar pattern, this indicator also oscil-

CW/TLAST-Jmlr. 1< ) /2000 0-98.625 -10.625 -15.6B* 0=118fii5 119.875 l =114.000 V=16097C0

; rirchnnlatjy Irtc: l I ?.



R8l -RS

........... Classic RSI Signals

Downtrending Day RSI ,.. : .

-2 . RSI


RSi t:

4 RSI:


RSI(Cttse,;blai:Kl- 57.0& 30.00- 70.00

104.000 -102.000 1 D0.000 98.000 -96.000


iniha 3:02 9:32 l0:02 10:32 11:02 11:32 12:02 12:32 1:02 1:32 202 2:32

Figure 5.11 The RSI system once again demonstrates the effectiveness by which Oscillator indicators are abletoidentify exhausted corrections within the major trend. Note the two Sell signals issued during the 11:00 a.m. hour. Also notethatthe system persistently issues orders against the major trend as the Oscillator is tricked into designatiag oversold areas on the chart.

Chart created with TradeStation® 2000iby Omega Research, Inc.

lates its calculated value between zero, where the underlying market is considered to be oversold, and 100, which again represents an overbought condition.

Percent R simply compares the position of the current close to the range of the underlying issue for the previous number of bars as specified by the length input. If the close of the current bar is equal to the highest high of the number of bars back specified by the input, the Percent R reading is 100 percent. The indicator expects that any market that is approaching a historical high will have some difficulty continuing its uptrend as this level is reached, because this is a natural area of resistance. Hence, a chart in this position is considered to be overbought and ripe for a correction by the Percent R indicator.

Conversely, as the close of the current bar approaches the lowest low of the last number of specified bars, the Percent R reading approaches zero, where the chart is considered to be oversold.

Historical highs, as used by the Percent R indicator, are defined as the highest high of the last "length" input number of bars. Obviously this historical high will change frequently as time moves forward. The

important concept is that the close of the current bar is being compared to rather recent highs and lows.

Similar to users of the stochastic tool, most users of this trading tool select the 80 percent level as the overbought threshold while considering any reading under 20 percent to be oversold.

Figure 5.12 illustrates the conventional plot of the Percent R oscillator indicator. For purposes of a meaningful comparison, we will use the same initial chart as was used for the same purpose for both stochastic and the Relative Strength Index.

Once again we are presented with the familiar plot of the oscillator-type indicators below the price charts. There is one significant difference between the plot of this indicator and the other plots we have examined here. These differences are detailed in Figure 5.13.

Note the flat areas of the Percent R plot highlighted by gray ellipses. These points are commonly seen with this indicator when the price is making new highs for the period identified by the "length" input value. As the closing price makes its new highs it is also placing a new value for the highest high of the period in question, resulting in these new highs always registering a maximum reading of 100 percent for this point on the chart. These flat areas are seen only when the closes of the bars in this area are at the very high of the bar in an

Figure 5.12

Percent R issues buy and sell signals in a fashion quite similar to stochastic and RSI.

i Chart (CMVT) Converse Inc LAST-2 ,.



100.500 99.000 97.50C 96.000

32 12X2 12 32 l:02 1:32 2:02 2:32

Figure 5.13 The highlighted flat areas of the Percent R plot are no more significantthan the points on the chart where the plotonlytouchesthe overbought or oversold extremes. Ome could easily mistake such a plot as being extremely overbought or oversold when, in actuality, market movements following these flat areas are of no greater magnitude than other signals from this indicator.

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

overbought situation or at the low of the bar in an oversold area. Therefore the trader should avoid placing additional significance on these plateaulike areas. It would be easy to regard these areas as extremely oversold or overbought when the only real significance happens to be the fact that the close is at one extreme of the bar or the other. Rarely is the price activity following such a formation significantly different from a chart point that merely touched the 100 percent level. The opposite is also true during an extended down move to the 0 percent level as also highlighted on the chart.

There are two methods by which the Percent R formula can be used to generate actual buy and sell signals. They are quite similar; both obviously make use of the defined overbought and oversold zones as previously identified.

The first scenario waits for the indicator plot to completely max out at either 100 percent or 0 percent, depending on the individual situation. The signal to buy or sell is then given as soon as a bar closes off from the extreme level. For instance, a particular stock is identified as

[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [ 10 ] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52]