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]

44
 Table 14.1 Top 15 Stocks from Database Symbol Average High CHRP \$6.84 ITWO 6.63 AFFX 6.61 PMCS 6.43 BRCM 6.37 5.91 RMBS 5.85 BRCD 5.72 INKT 5.50 QCOM 4.89 YHOO 4.72 QLGC 4.56 JDSU 4.47 4.25 SEBL 4.24

personality of the stock in question. The timing of the breakout entry as well as specific target and stop levels may be determined either individually or in combination with each other using this technology.

Initially, entry routines are selected by testing the system using a series of breakout times ranging from 30 minutes to one and a half hours after the open. Data generated from this test will usually report a range of breakout times that will be acceptable for the system when applied to an individual issue. The actual selection of a specific breakout time will be selected from this range after ranges for other system parameters are available.

We must now set exit parameters. While there are a number of strategies that are useful for both profit taking and stop loss protection, I have chosen to place a calculated target for purposes of taking profit and a calculated stop loss point to protect the position.

Think about a profit target from the larger perspective of trading over an entire calendar year. If you were able to take a profit of only 1 percent of the value of a stock each week, you would have an annual rate of return of 50 percent on your trading capital. Not bad. Assuming conservatively that half our trades will be profitable, we should

try to gain 2 percent per week or better to achieve our goal of 1 percent. Therefore, lets try for a profit on each trade of 2 to 3 percent of the value of the underlying security. If we can do this, then our goal of 1 percent profit per week should be conservatively within reach.

In many ways it is much easier to speak of profit targets and stop losses as a percent of the value of the item being traded than to try to take \$1 or \$2 or \$3 per share from each trade. While it might be possible to take a \$1 or \$2 profit regularly from trading a \$250 stock, it would be quite a feat to make the same gain on a \$25 stock. A percentage target is also one that can be more uniformly applied across an entire trading portfolio.

The same is true of a stop loss determination. If we are to take a rather small 2 to 3 percent profit we cannot afford to have our losses amount to much more than that, especially if we are assuming that half of the trades will be profitable.

In much the same manner as was used to set the breakout time for trade entry, an automated system can be utilized to establish a range of effective profit targets and stop loss levels. When these lists have been created, this same program can then be used to select the combination of breakout time, profit target, and stop loss level that is most profitable considering the trading style of the user.

Now, lets recap our decisions. We will set up the system to trade the breakout of the early range sometime during the first one and a half hours of the session. We will attempt to take 2 to 3 percent of the value of the underlying security as a profit target, risking about the same amount as our target level.

Wait a minute. Dont we need to be a bit more specific in these parameter determinations? The answer is both yes and no. We will definitely need to have more rigid parameters for each stock issue when we actually trade. But for now, still in the testing stage, it is adequate for our purposes to define the ranges for each parameter more loosely. Remember that each issue will respond a bit differently in its most profitable configuration as far as the system goes. Systematic testing of all three of our trading parameters can now define the settings for each security.

For instance, it could be determined by such system testing that PMCS responds most favorably to a breakout strategy applied 75 minutes after the open with a profit target of \$4.25 and with a stop loss placed \$3 from the entry point. BRCD may be found to be most prof-

USING AN AUTOMATED SYSTEM

itable trading the 45-minute breakout with a \$3 target and a \$3 stop loss. It is also a distinct possibility that an issue such as CHKP or ITWO may require, to be profitable, a large stop loss that would make trading this system against these issues impractical considering the risk-carrying ability of the trader. (Please be aware that these system parameters are presented here as examples only. These issues have not been tested against this system.) Other issues will be found to respond in much the same manner, all with different parameters.

Although this testing routine could be accomplished manually, the use of automated system testing software-such as Omegas TradeStation line of products, which has been used throughout this book-certainly simplifies the task while also reducing the possibility of error. An incredible amount of time would be required to design a system individually adapted by hand to each stock or commodity contract to be included in a trading portfolio.

Testing each issue using the out-of-sample routine will give values for our system settings that can be employed confidently in real-time trading.

CHAPTER REVIEW

1. Computerized trading systems offer the advantage of actual generation of trades from market data.

2. The advantage of these routines is the ability to observe the behavior of a system on historical data prior to actual trading.

3. The disadvantage is the temptation to curve fit a system to such an extent as to give an unrealistic picture of expected system performance.

4. System testing affords the trader the ability to determine the best system settings for use on each individual issue to be traded.

USING ONLINE CHARTING SERVICES

As mentioned in the introduction to this book, there now exist multiple web sites and online trading centers that offer as a portion of their services the use of the oscillator indicators we have covered in this book.

New sites and centers appear regularly on the Internet, almost on a weekly basis. While it is not possible or practical to cover a significant number of these resources here, I will attempt to familiarize the reader with a few items that are important to the use of these tools.

While these indicators are readily available and in all likelihood are calculated by the same basic formula, they will appear a bit differently on each site. For instance, stochastic could be referred to as "fast % k," "fast %d," "slow k," "fast d,""fast stochastic," or any combina tion of these. Percent R could appear as "Range Percent," "Williams Pet R," or others. RSI may be represented as "relative strength" or "Wilders RSI" after the originator of the study. Be careful here, as the term "relative strength" is often used when comparing the activity of a group of stocks to another group or to actually describe the relationship of individual stocks to each other within a given group.

Also, on some services it is not possible to apply indicators to the same chart using multiple sensitivity settings. On such occasions it

USING ONLINE CHARTING SERVICES

will be necessary first to apply the slow indicator to the chart, and switch to the faster setting when the slow setting has been satisfied and your attention then changes to that parameter.

Following are images of some of the currently useful sites that exist at the time of publication of this book. While this is by no means comprehensive coverage of the large number of useful trading sites, you can become somewhat familiar with the layout and capabilities you will encounter when using these items.

Trade Signal Corporation Ltd. (www.tradesignals.com), founded in 1997, publishes quantitative technical analysis, daily news, advice, commentary, trade alerts, real-time quotes, dynamic charts, intraday and end-of-day interactive Java charting, and quotes on the World Wide Web and by e-mail.

TradeSignals.com also provides a full range of trading tools. Stochastic is represented both as FastK/FastD and Quick Stochastic. It is possible on this site to represent both of our dual stochastic settings on a single subgraph using the FastK/FastD indicator as indicated in Figure 15.1. The RSI indicator is also on this chart. TradeSignals also offers the ability to chart two settings of this indicator on the same subgraph, as indicated.

To demonstrate the use of these indicator combinations on longer time frames, I have applied these tools to a X-minute S&P futures chart that spans several days. Buy and sell points as issued by the combinations of both indicators are identified on the chart. For easy reference I have labeled the overbought and oversold thresholds on each indicator subgraph.

Figure 15.2 describes the method by which the Directional Day Filter may be utilized on a TradeSignals.com chart. Since this package provides a trend line drawing tool, one can simply calculate the average of the first five minutes of trading and draw a horizontal trend line on the chart at that point. One can then also draw the vertical component of this filter, again using the trend line tool, placing the line at the appropriate chart location as shown.

Note that the filter is indicating that long signals would be most appropriate for this trading session. Dual settings of the stochastic indicator and a 50-period setting of Percent R identify individual buying points.

Figure 15.3 is by eSignal, which provides Internet-delivered,

USING ONLINE CHARTING SERVICES

Figure 15.1 RSI and stochastic plots are giving sell signals early in the day.

Chart courtesy of Trade Signal Corporation, Ltd.

SPZD (Intraday) 1 min

11/28/2000 11.2GET 1343.00 H 1344.50 L 1343 00 134450 Volume 0 Change 2.00 v-vas.e - M

 60 Minn te Line nuz-- A „ III -.....x .....- -----------Btree n-jo - - I" -fP f-.jtUi-. .... .
 FastK/FastD f%K %D1 -7: 3 1 7 -45" : ----- 80.0% 30% 10% Williams %R -50: El 3"32 ...... i;:.->.....;:\- 20.0% 80.0%

Figure 15.2 Use of the Directional Day Filter, stochastic, and Percent R are demonstrated on a popular online chatting package.

Chart courtesy of Trade Signal Corporation, Ltd.

[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]