This chapter explores how you can take existing ideas and then develop variations that suit your trading style. In this way, you can build a large collection of trading systems that meet different objectives. Every trader has different beliefs, time horizons, equity limitations, profit objectives, and confidence levels. Hence, few traders are executing precisely the same strategy. Thus you can trade even well-known systems without much difficulty. However, you can gain an edge by developing your variations of well-known ideas. As usual, this is not a recommendation of specific ideas for trading, but an attempt to stimulate your creativity. The focus should be on how each variation changes the structure of the trading system and its response to market data.
We first consider the 20-day or 4-week breakout system, perhaps one of the oldest mechanical trading systems in the business. We set the stage by looking at the results of a 20-bar breakout on the close using two different exit strategies. One exit strategy is a simple a "-day trailing exit, but the other is a more complex volatility-based exit. The first variation is using a fixed barrier above the 20-day range. The next step, after
148áDeveloping Trading System Variations
a fixed barrier, is to consider systems with a variable or volatility-based barrier.
In an interesting twist, instead of adding, we will deduct a volatility-based "barrier" from the 20-day channel, and examine the statistical significance of these channel breakout variations. All these variations should convince you that it is possible to take a proven idea and then adapt it to your needs or trading style.
We then look at two models based on the ADX indicator. The average directional index measures the strength of price trends. You will see that the heavily smoothed ADX reduces trades without missing major trends. We first test the effects of using a fixed ADX level to trigger the trades, and next, impose the restriction that the ADX be rising. You will see that the absolute level of the ADX is not as significant as its trend.
You may have heard this bit of trading advice, "Buy on a pullback into support." This advice by itself is not sufficient to build a trading system, but we will discuss a pullback system that works in uptrends as well as downtrends.
This chapter ends with a pattern-based system for long entries only. The deep pass in football inspired the "long bomb" system. With this system you can gather useful patterns off your favorite charts.
We will test these systems without optimization. Model values ??are chosen at random, before any testing, so that the test results do not bias the selection. The systems are tested on the same data set so you can understand the subtle interactions between system rules and market action. You should try to develop your variations of well-known ideas to find the particular combinations that best suit your trading style and meet your goals.
Table 5.1áSummary of markets and test periods used in this chapter
|áBritish pound Coffee Crude oil CottonDeutsche mark Eurodollar GoldJapanese yen U.S. bond Wheat|
|á2 / 75-7 / 95 1 / 75-7 / 95 1 / 83-7 / 95 1 / 75-7 / 95 2 / 75-7 / 95 2 / 82-7 / 95 1 / 75-7 / 95 8 / 76-7 / 95 8 / 77-7 / 95 1 / 75-7 / 95|
Channel Breakout on Close with Trailing Stops149
We chose the markets in Table 5.1 for testing simply because they represent a diverse group that had shown strong trends, periods of volatility, and many consolidations. The tests used an initial stop of $ 1,500 and allowed $ 100 for slippage and commissions.
áChannel Breakout on Close with Volatility Exit | áChannel Breakout with 20-Tick Barrier | áChannel Breakout System with Inside Volatility Barrier | áStatistical Significance of Channel Breakout Variations | áDay Exit Reference System | áTwo ADX Variations | áThe Pullback System | áThe Long Bomb - A Pattern-based System | áSummary | áIntroduction |