Interview on Automated Trading Championship 2012 from 16.10.2012.
Francisco, tell us a little about yourself. Where are you from and what do you do?
I'm from a small town Huéscar near Granada. I have a degree in Computer Science at the University of Granada and work by vocation as a school teacher. I enjoy my profession. I have registered at the MQL5.community to participate in the Championship as I wanted to test myself.
So you have some experience in programming?
Yes, I write in many languages: C, C++, Java, PHP, Perl, ASP, LISP.While knowing the general principles well enough, it is not difficult to understand the features of a particular language. Each one of them fits its own range of tasks, you just need to choose the right tool.
How did you discover the world of automated trading?
About 5 years ago I became interested in trading, especially in the ForEx market and since then I have been doing regular self-training in this area. For automated trading, I've tried lots of different platforms, but eventually opted for MetaTrader. MQL4 and MQL5 are responsible for this. In fact, coding - that's what (in my humble opinion) I am good at.
Which language MQL4/MQL5 do you prefer more?
Undoubtedly MQL5! It is much improved compared to the previous version due to the object-oriented programming and event handling. I do not need more.
Why did you start automated trading right away? Have you tried to trade by yourself?
I haven't traded on real accounts yet, but even on demo I was unable to control my emotions. I find it hard to follow a strategy, if you have to keep a loss-making position or to close the terminal without fixing a profit.
I am a reasonable man and prefer not to go on an adventure, so I decided for myself that I will program trading strategies. This is the only way to eliminate influence of emotions, which is detrimental for your deposit. For this reason, MQL5 language if perfect for me as a working tool.
How do you create trading robots? What is the most difficult for you?
The hardest part I think is finding a successful trading strategy. There are many strategies that work well for certain currency pairs, but the problem is to select specific values of parameters. It is very difficult to optimize a system to have no adjustment.
If you adjust parameters for the maximum profit on the history data, Expert Advisor will fail in the future, that's for sure. Forward Testing solves this problem. Implementing a strategy in form of EA's source code is the easiest part for me.
How did you decide to participate in the Championship?
I have tried many strategies, but last year I did not have time to prepare my EA for the Championship. This year I decided to try my wings and began to prepare in advance. After some hesitation, I have chosen one of the two strategies interesting to me.
How did you use Strategy Tester?
I have backtested my Expert Advisor on two year history and then arranged forward testing. I didn't use the computational resources of MQL5 Cloud Network as I wanted to test EA in another way. But I highly recommend this service to everyone who program Expert Advisors.
Why your EA trades only three currency pairs GBPUSD, USDCHF, EURJPY? Why haven't you taken the popular EURUSD?
I had to exclude many currency pairs from the final set, since they do not make a profit or profit was only taken in a short period of time.
Tell us about the strategy your Expert Advisor is trading
EA trades on D1 timeframe for each pair every two days. Indicators are not used. Instead my EA compares the current price with OHLC prices of previous days and places pending orders. Market will show which ones will trigger.
I do not use trailing stops, instead I sometimes move Stop Loss to a winning position. If you want to win the contest you have to take risks. My money management takes into account the daily movement as well as different open positions.
Since my Expert Advisor is multicurrency, I carefully handled the lack of capital problem - the volumes of pending orders are dynamically changed due to OnTrade event. Losses are suppressed by other pending orders that act like Stop Loss.
D1 is not the most popular timeframe. Why did you select it?
Lesser timeframes generate more price noise, thus more false signals. And the bigger ones do not allow to trade aggressively. Don't forget that the Championship is held for less than three months. Therefore D1 is best for my strategy.
Automated tests have revealed some errors in your EA. How did you cope?
Yes, during the test I encountered a wrong behavior of my trading robot - it traded volumes inadmissible by the rules and placed incorrect stop levels. I want to say thanks you for the How to Fix Errors in Trading Robots and Trading Rules of the Automated Trading Championship 2012 articles, they really helped me.
The biggest trouble was with stops. Unfortunately, I made a mistake in the algorithm of sending trade requests, and as far as I know now I'm not the first who run into this problem. I hope the upcoming news will shed some light on this issue.
Do you use MQL5.community services? What do you think of the newly released Signals?
First of all, community is the best source of information about MQL5 language. I read articles, observe publications in Code Base and constantly refer to the documentation. The Signals service seems very interesting and promising to me.
What can you recommend to those who dare to engage in algorithmic trading?
First of all, remember that Expert Advisors are systems that can't always be profitable. You have to constantly improve your trading robots. I recommend beginners to start with the documentation and then try to create a simple EA based on moving average crossover.
Thank you for the interview and good luck in the Championship!
I want to thank MetaQuotes Software for the excellent trading platform and for the MQL5.community. I wish good luck to all other participants.
Translated from Russian by MetaQuotes Software Corp.
Original article: https://www.mql5.com/ru/articles/563