|
|
|
|
|
|
|
As you come up with your own trading methodology, you should refer back to these steps. Unfortunately, designing a computer trading system that accurately reflects your trading methodology demands a lot of time. It is not something that you can put together over a weekend. However, the huge advantage is that when you are done, you have accomplished something that 98 percent of all traders never do. Consequently you will see your trades produce consistent results. Once you are producing consistent results, either profitable or not, you can begin working on your ability to perceive the market better. As your perception increases, your ability to produce consistent profits will increase. It should be stressed that in order to have the internal beliefs required to do the research necessary to develop your methodology, you must make a decision to become responsible for all your beliefs. |
|
|
|
|
|
|
|
|
At some point all traders must confront how they will use the power of the computer to optimize their trading methodology. For those traders who are unfamiliar with optimization, allow me to briefly describe it. Optimization is achieved when you have written a mathematical formula or theory that describes the market action (in part or wholly) through variables. By programming the computer to literally perform all the mathematical permutations possible on the variables, and then correlating these permutations to the profitability, you can determine the variables that created the most profit. In other words, by determining the best combination of variables to maximize profitability, you can create a highly profitable methodology. The only problem is that it is good only for historical data; it is absolutely worthless in real time. |
|
|
|
|
|
|
|
|
Say we have a simple moving-average crossover trading system. Our rules are very simple. First, if the short-period moving average goes above the longer-term moving average, go long. Second, if the shorter moving average goes under the longer moving average, go short. Consequently we are always long or short. By writing (or buying) a program, we can specify that we want to vary the shorter moving average from a period of 2 to 20, and the longer period from 21 to 60. Then by allowing the computer to test all the permutations that could occur by varying the periods of the shorter and the longer moving average, and by keeping track of the profitability, we can determine the most profitable short-term and long-term moving average. For example, the computer might indicate that a shorter moving average of 15 days and a longer-term moving average of 57 days generates the best profit. Typically the second most profitable combination of variables will generate less than half as much profits as the most profitable combination! |
|
|
|
|
|
|
|
|
At this point most beginning traders are very excited, convinced that they have just found the holy grail! There is a huge problem here. These |
|
|
|
|
|