Work for Forex?

DOES THIS WORK FOR FOREX?    

          Since currency trading is a large component of my trading, I wanted to optimize a system for forex. The Defcon model was tested to stand up against a non-correlated investment vehicle to the equity markets. I chose the euro currency market to run a performance test. As we discussed in Chapter 7, due to the computers’ inability to test the forex markets’ data because there is no centralized market and prices are quoted in bid/ask form, we ran the test using the euro currency futures, which trades parallel to forex markets. The test period was conducted using 15-minute intervals during the U.S. open outcry trading session from 8:20 a.m. (ET) until 3 p.m. (ET).

The winning percentages were not as great as in the S&P; but, boy, the bottom-line results showed a healthier profit! Table 11.5 shows the rate of return with 365 percent, but it was based off a recommended starting balance of $16,770. The overall gross profit was $61,275, based on a test period that went back three years. We had 310 trades—the system generated slightly more trades here than in the S&P. This may indicate that the market is more volatile, not less, as some people believe

As for the test period, I believe this was a good time frame because it represents a great sampling of various market conditions, considering that we had several market conditions exist: bull trend, to a consolidation phase, and then a trend reversal or downtrend.

In January 2003, the euro was valued at 1.0500 to the U.S. dollar. It went as high as 1.3660 in December 2004; and as of February 2006, the euro was back at 119.00. During that time, the market conditions changed from bullish to bearish and went into a consolidation phase as well. We want to see how a trading system performs in various market conditions. Notice the three market conditions as indicated by the trend lines drawn in the weekly euro currency chart in Figure 11.8. According to the test results, the system fared pretty well.

Let’s examine both the volatility and the margin requirements. As of February 15, 2006, the daily initial margin requirement to trade one euro currency with a 125,000 contract value (margin requirements are set by the exchanges and are subject to change without notice) was $2,835.00. As you can see, we are trading a highly leveraged market that has an average daily range of approximately 86 PIPs (percentage in points) per day. In the futures, each PIP is $12.50.

That computes to a daily trading range of $1,075 per day. In using strict money management guidelines, we are using only 28 percent of our investment capital. So in the spirit of being a great trader, I want to ask why I would trade only one contract with a starting account of $10,000. The answer is clear, and another great example of why it is important to know what you are doing before you do it. By back-testing a system, I can determine what the worst to expect is. Granted, we did not implement any means for risk management. The Defcon system solely generates a trade triggered by the pivot point moving average approach. You will see that by going back and validating the methodology, we had a single loss in the amount of $4,313. That is the bad news; and more than likely, as with any severe and sudden market loss, it was generated by a shocking news event. The monthly unemployment report on August 6, 2004, was one such event where the euro currency moved almost 260 points in a single day. This was a good definition of a news-driven price shock. As the data in Table 11.3 shows, the profits were bigger in the euro than they were in the e-mini-S&P; but so were the risks. So we have a bigger profit; but along with that greater profit came greater risks. How many times have you heard that before? The bigger the risk, the bigger the rewards. I have heard that a bunch of times, but it is great to see it put in front of me based on a defined set of rules. So if I get the feeling one day to go “all in” like a poker player and use 100 percent of my margin, in the euro currency I could easily do three contracts. In one bad day, not only would I wipe my account, but I could be deficit, meaning I could actually lose every penny in my account and owe money.

Using the back-tested results of a system allows me to see my pro-gram’s strengths and weaknesses. That is the edge I have against the market—knowing when to raise my positions and when not to. Besides learning that my system can make money, I also found out that I need to double my account if I want to double my positions. Using a computer product like Genesis allows me to identify and validate the methods employed. As we look at the equity curve in Figure 11.9, we can see a solid performance; and that should help maintain your confidence to stick with trades generated by the signals. Figure 11.9 also highlights a negative $13,162.50 maximum drawdown from peak profits. Imagine increasing your lot or contract size prematurely during that negative phase. It would certainly ruin your trading day, year, or even career!

In the stacked bar graph in Figure 11.10, we see the results of peak performance; and more important, we see the kink in our armor, the weakest point in our system from a seasonal perspective. By back-testing the system, our diagnosis shows a seasonal weakness in the markets that occurs in January and February and continues into March. We know from the table in Figure 11.1 that we have a string of winners that last on average 10 in a row.

Therefore, if at the end of October I have 10 winners in a row, I think I will be more selective in my trade signals in November! As you can see, the statistics show small drawdowns in November. According to the three- year test period, April through August is the most profitable trading time period (Table 11.6). We can enhance our performance by not trading in April, though, as that was one of the worst-performing months as indicated in Figure 11.6 on page 297.

As you break down the numbers by the statistical results, it is by that data that you can determine the validity of your methods (Figures 11.11 and Table 11.7). When we look at a month-to-month breakdown of the euro currency, we see January 2004 was a whale of a disaster, as shown in Figure 11.11! With this information, we can determine if the cause of the disaster was the system, the methods, or one heck of a wild trading period. The answer should not surprise you: It was the last. There were only two days that had daily trading ranges over 250 points. On January 16, 2004, the market dropped like a hot sack of potatoes, with a high of 1.2610 and a low of 1.2351, for a range of 259 points. Then two days later, it reversed higher, with a low of 1.2345 and a high of 1.2599, for a 254-point range. The market gained back what it lost; but overall, it was simply a very violent trading period, as Figure 11.12 shows. Without some type of intraday risk management method, such as a trailing stop, all traders at some point will be subjected to news-driven price-shock market environments.

No one is immune to them. The good news is that they do not occur frequently, as this data shows.

To summarize, the essence of back-testing is validating your methods and showing the strengths and the weaknesses of your system. Moreover, it will help you define your goals and expectations for performance. Therefore, it can help you achieve the highest trading profits with the lowest risks in most trading market conditions. I want to elaborate that by studying a system, in all market conditions, you at least will be better prepared and less shocked when an eventual negative situation develops in your trading career. The main goal in trading is consistency and staying in the game. Statistics show this!.