These thoughts are designed to be a review with a few twists for experienced system builders and as a possible framework for those who want to build a trading system or are just starting. The beauty of system building is that the extremely complex inner workings can be boiled down to simple, easy-to-use, trading screens.
A simple definition for “trading system” is a group of rules/parameters that determine entry and exit points for a given kind of trading vehicle, whether it is currencies, commodities, equities, etc. The points are most efficiently plotted on a chart in real-time and prompt the immediate execution of the trade.
The main indicator groups that lend themselves to basic systems are trend/counter trend indicators such as moving average convergence/divergence (MACD), moving averages, Parabolic SAR. Another group of indicators measures momentum. This group includes the commodity channel index, relative strength index and stochastic oscillator. The volatility group measures change in volatility, which can lead to change of price direction – examples are average true range, Bollinger bands and price channels. The volume group is used primarily to confirm direction or breakouts. Chaikin money flow and a moving average of volume can be effective tools used for this purpose. The event driven group monitors events that take place before a trade, such as buying on the 3rd consecutive down day, holding, and selling on the 3rd consecutive up day.
As you can imagine, event driven choices are limitless. On top of these indicator groups come a plethora of ways to control the risk of the trade as an overlay, whether it is trailing stops, hedging through the accumulation of the counter-trade, Fibonacci retracement, etc. As you probably already know, the manipulation of the data within each of the previously mentioned indicators and other popular indicators that are not mentioned is also limitless.
With every trading system, each has its own advantages and disadvantages. One of the main reasons for using a trading system is that it takes the emotional side away from carrying out any business. In trading (and to a large degree life in general), it is in human nature to worry, to hope, to let ego dominate, to blindly speculate, to not admit a mistake and let losers run, etc. These traits are absent in trading systems.
A trading system makes all calculations basing only on facts and clear logic. Another advantage is that much time is saved because if automated for trading, the system performs work that a human did previously. If the system is not automated for trading, it at least saves a great deal of time by alerting the user to the trades that need to take place. Thus, the convenience and time saving factors allow the trader to engage in other activities that could improve the business or personal life.
A main disadvantage of trading systems is that they can be complex structures. Robust systems require a lot of time and effort looking at many different combinations of indicators, factoring, determining allocation structure, risk containment, etc. and delivered in an elegant enough fashion that a single user or multiple people in a firm can use comfortably.
Before designing a system, one must determine the size of the universe, minimum and maximum allocations, number of indicators, average holding period, number of trades, costs, slippage, target returns and drawdown.
• Size of the universe – many individual traders and system builders concentrate on the EUR/USD currency pair because it is the most liquid market comprising about 28% of all Forex trades. However, it is possible to explore great trading opportunities beyond the euro and U.S. dollar. In order behind the U.S. dollar and euro, the Japanese yen, Pound sterling, Australian dollar, Swiss franc, Canadian dollar, Swedish krona, and New Zealand dollar represent about 91% of all currency distribution in the Forex market, so the combination of these currencies provides enough “trading food” for the trader.
If you want to create different return streams, one method is to work with more pairs. Of course, it is up to the trader to design the pairs to trade based on their personal preference. But for the purposes of this article, assume that we are using the eight non-U.S. dollar currencies listed against the U.S. dollar.
• Minimum and maximum allocation – if you want the universe of eight pairs represented at all times in the portfolio, one can limit the minimum allocation slot to greater than zero. Also, if a trader wants to reflect their quantitative analysis via weighting, then an allocation reward system can be implemented. As an example, if the EUR/USD pair exhibits higher quant scores than USD/JPY, then allocation can favor a much greater weighting to the EUR/USD.
However, some systems simply reward allocation to performing pairs in an equal weighted fashion. For example, the 8 pairs in our universe can be divided into 12.5% allocations and there is simply a change in pair direction with each cycle of strength and weakness.
• Number of indicators – an indicator can simply mean one simple measurement as in price crossing above a moving average, or, it can be a measurement of the culmination of many indicators that is expressed in an easy to read fashion. If combining indicators into one, it makes sense to use like-type indicators with others. As an example, one could build an interesting price oscillator combining RSI with stochastics, commodity channel index, with percentage of movement above a low channel line and percentage of movement below a high channel line.
Maybe an interesting moving average indicator would be combining a direction of Hull moving average, a price oscillator of simple moving averages and an adaptive moving average. As a minimum, a trader should run 10 concurrent indicators per pair. They can be the same indicators for each pair, or, if someone desires more complexity different indicators for different pairs.
This is an excerpt from May 2011 issue of Forex Journal.
