Interview with Andrey Barinov (ATC 2012)

Interview on Automated Trading Championship 2012 from 09.10.2012.

It was on Friday of the Championship's first week that the trading robot of Andrey Barinov (Wahoo) occupied the fifth place in TOP-10. Andrey is a newcomer in the Championship but he has already managed to execute more than 100 orders in Jobs and develop a dozen of products for Market. We have arranged an interview with him and learned that the development of a "simple multicurrency Expert Advisor" is not an easy but a fairly easy task.

Hello, Andrey! Why do you have such a strange nickname Wahoo?

I think, it is no stranger than any other. I was trying to invent some non-existent word initially. But it turned out to be quite a challenge. Later, I found out that Wahoo is a name of a very strong and fast fish difficult to catch. I decided that strength and speed are the things I want to be associated with. "How do you call a boat, so it will float."

Very interesting! How did you start your trading activity and programming of trading robots?

I started real trading a few years ago. I traded manually but the results were not so good. Then I took a rather long break. I accidentally learned about the Championship a few months ago and decided to participate. To do this, I had to master MQL5 to develop my Expert Advisor. Then I became interested in "Jobs" service and realized that I can help other people to write Expert Advisors based on their strategies. However, I also had to clarify some questions concerning MQL4, as it is still a very popular language. That is how the next stage of my Forex activity started 3 months ago.

What prevented you from successful trading - lack of knowledge, experience or something else?

I think, it was a strong desire to make money as soon as possible. It is common enough for many traders. I've made plenty of mistakes caused by emotions. Big profits have made it difficult to think soberly leading to even greater losses. You should be fully confident in your strategy and avoid any hesitation if you want to trade manually. I was not so good at that.

You've said, you had not known MQL4 up to then. How did you perform your trading operations and test your strategy?

I do not know whether it is appropriate to mention competing products here but there are some other platforms for manual trading apart from МetaТrader. I used these products both for manual trading and for my strategy testing.

You've mentioned a full confidence in strategy. How can it be obtained?

First of all, it should be checked on history using the tester. Then it should be tried on demo and real accounts for a sufficient amount of time. If the balance is growing slowly but steadily, the strategy can be called promising. I've noticed that the lower the risks, the greater the confidence. When risks go beyond all limits, you are lost in doubts, as any deal may turn out to be the last one for your deposit. Therefore, your confidence is stronger and you feel better when the risks are low. However, the profit is also limited in the latter case. Traders should select the type of trading activity and risk management that suit them best.

So, the necessity to test your strategies has led you to the algorithmic trading? What is the first strategy you have programmed?

Well, maybe you are right. The first program was a simple multicurrency Expert Advisor based on different use of the moving averages. I combined the most popular options in one Expert Advisor with the ability to switch between the strategies and optimize them.

Excellent answer! I've especially liked the beginning about "a simple multicurrency Expert Advisor" followed by "with the ability to switch between the strategies and optimize them". So, what Expert Advisor can you call complex then?

I think that a "complex" Expert Advisor requires implementation of additional features apart from the ones present in MQL. For example, neural networks.

I see. Let's talk about your activity in Jobs service. Do you often receive some interesting orders? What are the most popular job types?

Of course, some orders turn out to be quite interesting. I especially like orders aimed at detecting market models or based on pure analysis of the price behavior without using indicators. However, most orders are of the opposite nature. Ideas based on various indicators and different order grids are the most popular ones. Besides, there is often the need to modify an existing Expert Advisor or an indicator. These are the least interesting tasks.

What language is the most popular among the customers - MQL4 or MQL5?

MQL4 is much more popular but I think, that will change pretty soon.

Do you need some specific knowledge to develop such a complex algorithms? What is your background - education and work experience?

I have a higher technical education and I've always been good at Math and logic. Besides, I am constantly learning. The more orders you execute, the easier each successive one becomes.

Do you experience any problems with orders or customers? What are they if you do?

Nothing serious. Sometimes, I face misunderstandings that can lead to tense situations. But I've been able to handle them so far. If customers are not satisfied, I'm always willing to allow them not to pay even if the work has already been done. However, I have not had to do this so far. All issues have been resolved amicably and by mutual agreement.

You have more than 100 completed works! How do you manage to cope with such a large amount? Do you have some professional tricks?

Nothing special. Of course, I had to develop my own function library allowing me to assemble Expert Advisors brick by brick. Besides, I have a template containing basic functionality. All other things are developed individually for each new task. But, as I've already said, the more orders you execute, the easier each successive one becomes.

Do you have regular customers? Do they recommend your services to other community members?

Yes, I have regular customers, as well as new ones. There is always plenty of work for me to do and I always have a few projects in the queue.

Do you have any suggestions concerning Jobs service? How can you evaluate the service using the five-mark grading system?

In its current form I would mark it with 5. If users could change the price and stop the work by mutual agreement after conforming a technical assignment, it would be 5+.

Let's go back to the concept of "a simple multicurrency Expert Advisor". This is exactly the type of Expert Advisors you offer in Market. Why?

МetaТrader 5 allows traders to test multicurrency strategies and I think, this feature should not be ignored. Besides, it provides more opportunities for managing risks encouraging the development of more interesting strategies. They can be implemented for МetaТrader 4 as well. But it is impossible to test them there. I think that the multicurrency tester is one of the main advantages of МetaТrader 5 over МetaТrader 4.

I see. No wonder you have submitted the multicurrency Expert Advisor for the Championship. Tell us a little about it, please.

The Expert Advisor works with three pairs: EUR/USD, USD/JPY and EUR/JPY. Market entry is performed by the trend. Trend is determined by the slope angle of the linear regression line. Trading is performed not very frequently. Market exit is performed by stop loss, trailing stop or at a break-even point. Stop levels of the current position are changed in case of an opposite signal. This version of the Expert Advisor is similar to the one I have submitted for Market. I've made only slight modifications for the Championship.

What profit did it make during the automatic tests? What was the optimization criterion?

It made $300 000 for the first 7 months starting from $10 000 with the maximum volume of 5 lots per one pair. The final balance was used as the optimization criterion.

Where do you collect ideas for development of trading robots? Do you have your own inventions or everything can be found in books and on the Internet?

It is always nice to realize that you have invented something. But in most cases it turns out that your invention already exists, though in a slightly different form. I check the ideas that seem to be interesting to me. There are various areas to look ideas for. For example, I do not believe in the indicators, so I look for ideas in other areas. It is very difficult to invent something new in this field but you can combine the best elements and look at the results - it's an interesting challenge.

Is it your first Championship? Have you used experience of the previous years' participants when developing your Expert Advisor?

Yes, it is my first Championship. When examining the previous years' results, I realized that it is not necessary to perform a lot of deals to win and many things depend on luck.

What are the most complex and the simplest things when developing a trading robot according to your experience?

The most complex thing is to formulate and formalize your strategy clearly, so that it does not contain any ambiguities and covers all possible situations. The simplest thing is to write a code.

Now you can assemble robots one by one like on an assembly line. How many Expert Advisors have you made?

This is not the case when quantity changes into quality. I mean, the number of Expert Advisors does not increase the profitability of each of them. I have developed about 10 different robots to check my ideas and about 50 ones for customers. The robots made for customers are not checked for profitability, they are only checked for their compliance with a technical assignment.

According to the Championship's statistics, members of the first week's TOP-10 fall out of the race long before its final days. Do you feel any uneasiness that your trading robot is at the top of the table right now?

I'm not afraid of that fact. However, I am not too excited about it as well. I have no illusions, as I'm fully aware that TOP-10 is likely to completely change each week. Probably, it will fully change by the moment this interview is published.

Are you going to root for someone? Perhaps, you have some familiar developers among the participants?

I'll root for our guys :). Besides, there is one participant from another country whom I helped with an Expert Advisor for the Championship. I'll root for him as well.

What could you wish for those who want to do algorithmic trading? Where should they start from?

First of all, develop your strategy on paper. Then you can start coding it. You don't need to do everything alone. You can seek help from professional developers. They will help you with the first version of your Expert Advisor. Then you can modify it yourself if necessary. It all starts from a strategy. If you have a strategy, all the rest is a matter of technique. You don't have to know everything. You just have to know where to find someone who can help. MQL5.community is the very place for that.