With foreign exchange (forex) turnover averaging $1.5 trillion a day, the Global Interbank Currency Market is the world's largest financial market. The four "blue chip" foreign currencies are the Swiss Franc, German Mark, British Pound and Japanese Yen. The US$ price of these currencies fluctuates constantly, 24 hours a day, 7 days a week. You can make money trading foreign currency if you "buy low, sell high". Suppose, for example, you buy one lot (62,500) of British Pounds at $1.6950 per Pound because you believe the price will rise. If the price does indeed rise to, say, $1.6990 a few hours later, you can then sell your lot for a $250 profit. If, however, the price falls to $1.6910 and you decide to sell your lot because you believe price will not recover soon, you lose $250.
The art and science of trading is to predict correctly, more often than not, which direction price is headed to yield a profit. Many books on this subject promote a trading system to predict price movement that works well with 20/20 hindsight in isolated market situations, leaving you to discover how well their past performance predicts tomorrow's uncooperative market. Unfortunately, trading is one endeavor where mistakes cost money. Consider the following case study.
Early this year, my brother Steve read the book High-Impact Day Trading by Robert M. Barnes. He asked me to write the computer software necessary to validate the author's "mountain-valley" (m-v) trading theory. I agreed, and Steve subscribed to the Data Broadcasting Corp real-time market data service and began collecting intraday tick data on the 4 foreign currencies that the author tested in his book. It turns out that there were a few important details the author only made casual reference to or did not discuss at all.
Taking every buy-sell signal around the clock every trading day from 02/11/97 to 05/02/97, m-v did nothing but lose money in our 20/20 hindsight testing. By narrowing our trading hours, 12:30AM to 2:00PM for the Yen for example, taking only the first buy-sell signal, and taking a contrary position we were eventually able to tune m-v to show $23,630 in profit. Impressive for 55 trading days! We then added our broker's 10 point buy-sell price spread, and profit plummeted to $7,364. Finally, we added a 50 point stop loss and profit dropped to under $5,000. Although 80% less than expected, m-v did show a significant profit.
The remaining question, of course, is "How well does past performance predict the future?" To answer the only question that really matters, we had to put real money on the line. We soon discovered another important detail: broker quotes do not span the range of market quotes. A few points per trade add up. During the next 4 weeks, we lost 10% of our trading account.
The next time you consider using a new system for trading foreign currency, separate the marketing hype from reality by using a simulated trading account. Most brokers, such as Currency Management Corp. in London, will set up such an account for you on their real-time electronic trading server in the hope that you will become a client. However, you can only use the account for a limited period of time for real-time testing, and you cannot do any historical testing. For unlimited real-time and historical testing, you can purchase Forex Trader, a foreign currency trading simulator available on CD-ROM for your Windows 3.1 or Windows 95 PC that lets you buy and sell at realistic broker prices. This innovative by-product of our study also feeds the historical tick data supplied on CD-ROM to your trading system just like DBC does for complete simulation of your trading decision environment.