by Joe T » Fri Mar 19, 2010 10:37 am
In my first year of trading, I purchased an EA that sounded like it made sense to me, and then I tested it and it seemed to work pretty well. Then I went live with it and it ended up being something that should have generated a small loss, but because I wasn't properly considering money management at the time it generated a larger loss than it should have. So I went back to smaller lot sizes, and finally after a couple more months realized that it wasn't anything special.
I then went through a phase of trying to learn the programming myself in order to test and test and test and develop an EA that should work. Conceptually, I think you could do this. But I determined that there are enough different kind of market conditions that can occur at any time, and the parameters are so complex, that it was pretty difficult.
I actually developed a couple EAs that increased account value very consistently when backtested over a period of years, but the overall yield on those weren't much higher than a typical mutual fund yield - 10-15% annually. As soon as you got into higher yields, the risk that you couold blow out your entire account seemed to rise exponentially.
I think it can be done, but I finally concluded that the educational time, programming effort, hardware, data reliability and software requirements to properly backtest was something I wasn't willing to pursue. In fact, I developed a couple EA concepts that took varying market conditions into account that I threw out to MQL programmers, thinking maybe they'd be interested in doing it, and when they looked at the requirements they thought it was above their pay grade, so I pretty much knew that I wasn't going to ever learn enough to get there myself.
One thing I learned about EAs - whenever they make claims about yield that requires trading in lot sizes that could blow your account out with a few trades, run away. Sure, they can show a huge yield over time, but if a future market condition is a little different than anything in the past and it takes a little time to figure out that you should turn your EA off, it may be too late. The other thing I learned is that unless an EA does take into account market conditions, then you still need to know enough about the market to know when to run it. Claims that a novice who knows nothing about Forex can make tons of money are very misleading. You probably can make a lot of money from an EA, but it ends up being little different from regular trading. You have to understand the market and the currencies, and then you have to recognize that the conditions are right to run the EA. If you can do that, then it probably can be a valuable tool. On the other hand, by the time you learn that, you can probably recognize when to make manual trades, anyway, and you feel less inclined to run the EA.
The one thing I liked about building my own was that I knew what it was doing and how it was "looking" at the market. Buying an EA blind will take a lot more effort to figure out how to use it with given market conditions.
I've now been trading a completely different strategy for a few months. I haven't used any EAs for a while and really have no future plans of even testing them, unless a lot of people I trust tell me they've found one that is truly different.