Interview with Maxim Kondratyuk (maxfade)

Hello, Maxim. Could you tell us a little about yourself, please? What do you do? Since what time have you been trading?
I’m 33 (practically 34), have a diploma of higher education. I have been interested in working on Forex since 2005. I’m employed in logistics and transportation at the moment. Before that, I was a programmer.

Does your programming background help you in developing of Expert Advisors?
Of course, it does. I don’t experience any problems in coding EAs. It’s much more difficult to invent an idea.

How do you find ideas to be realized?
I usually try to find regularities that are visible in the price chart on different timeframes. Then I write an EA to test them, then check it with the Strategy Tester and optimize by inputs. If all these actions result in significant ranges of values in the optimization graph where the EA trades with profits, it means that we can expect the EA to trade profitably in future if we take values from the center of that range. As to indicators, basically, I use МА and sometimes МАСD now.

You wrote that your EA was using concurrences of averages as a signal to open a position. Such signals are considered to be strongly behind the real situation. Why do you utilize them?
It’s not quite concurrence of two averages. It’s rather when price crosses an average with a certain period (although price can also be considered as an average with a period of 1). We can see in the charts that price often crosses MA and then moves on. It is this movement that my strategy tries to “intercept”.

The contest EA uses some fixed inputs: MA period, the levels of SL and TP. These parameters were chosen out of profitable ranges according to the optimization results. In future, I’m going to seek regularities in order to make the inputs dynamic, newly calculated for each new trade.

What interest do you have in automated trading?
I’ve got some experience in real trading. Automated trading is one of steps to be made to exclude emotions when making decisions, as well as to win some more spare time for oneself. The EA will watch the market and it will do what and when should be done, whereas it does not do anything if the chance has already been “let go by”. The only thing to be done is to support its high availability and good connection to the server.

How often do you intervene into your EA’s operation? In other words, how long can your EA work without your intervention, without any changes in the code?
I haven’t developed an EA yet that I’d use on my real account. However, I launched them on my real account for a very short time in order to test them, to estimate the differences between a demo account and a real one.

What differences were detected?
Let’s put it that the differences can be seen in the manual mode. For example, the execution time for trade orders is still longer on a real account than on a demo one, so I wanted to observe the behavior of my EA in both cases. It worked on real without any special problems, so that was sufficient for me.

What Expert Advisor is an effective one, in your opinion?
I think an effective EA should consist of several subsystems that would overlap at as many as possible different market stages. At that, it should be profitable in “its” area and should be non-losing in the “others’” area.

Your EA is showing good results now. Let’s discuss its algorithm – what is it based on?
It is based on an observation that price often crosses MA and moves on far-forth. A pending order (Buy Limit or Sell Limit) is placed at a price that equals to the value of МА(590-М1). If the price changes, MA value changes, too, so the order is respectively modified, and so on until it triggers.

The order is not modified after it has triggered. Positions are closed by stringent ТР and SL. As soon as the EA “sees” that there are no orders, it places a new pending order. Since it often happens that the price only touches МА and doesn’t move on, I made my SL rather close – 30 points, whereas the TP=70.

Why is the number of short positions almost half as large again the number of long positions? The pair of EURUSD traded by your EA, basically, grew all the time during the Championship. Is your EA a kind of countertrend ones?
The EA tries to catch a “trend” of 100 points. This is why it does not “see” the global trend. And another thing, it hates it when a pair grows without strong rollbacks, since, in this case, it always tries to open a countertrend position and has to close it by SL. This is why the EA, in case the preceding trade has been losing, most probably opens against the trend, its purpose decreases twofold (just a beautiful number) and TP=35 points. This way, the EA has more chances to get a profit on a rollback and open by trend, not against it.

Before November 27, Balance and Equity of your contest account ranged between 10 000 and 20 000 dollars. Only after November 27, when the price of EURUSD fell from 1.4960 to 1.4530, your Balance went up to 40 000 dollars. Was it an accident or an expected change?
It depends… Well, it’s rather a lucky accident. I didn’t get so many consecutive profitable trades during testing. This type of market (large movements in both directions) seems to be the most suitable for my Expert Advisor. I also checked the ratio between profitable and losing trades that time – the ratio was 50/50.

The number of profitable trades is the same as that of losing ones; however, the profit of your account is positive. Could you explain, please, how you achieve this?
This is achieved due to the ratio of TP/SL (70/30>2). It means that even two losing trades will be covered by one profitable trade (it is important that the losing trade frequency does not increase by the same amount of times).

You are not using Trailing Stop for open positions, modifying pending orders instead. This is a rare approach. Why have you chosen it?
Frankly speaking, I didn’t have enough time to study how Trailing Stop influences this strategy.

Your EA is attached to the one-minute chart of EURUSD. What other timeframes does it “track”?
No other timeframes. I explore the fact that МА(590) on a one-minute timeframe is approximately equal to МА(590/60) on a one-hour timeframe. We just cannot use MA on one-hour timeframe with a fractional parameter. This is why I’m using one-minute timeframe. However, if we used the values of MAs of even larger timeframes, we would have to utilize them directly, since the built-in function does not accept too large values of parameters.

You have already mentioned that you’re using MA and MACD. Are you utilizing any other, additional indicators?
No, I’m not. The MACD itself is just a set of simple arithmetic operations with the values of averages, the same are other indicators as a whole (however, I’ll probably discover an indicator for myself, who knows).

Maxim, why did you choose such a name for your EA – Disaster?
To scare misfortune away.

You participated in the last-year Championship, though without such achievements as in the current one. Did the results of Championship 2006 help you in your preparations for this one? What lessons have you learned?
Yes, the results helped me. I started thinking more over the Expert Advisor’s optimization results: Are they random, or is it the result of adapting parameters to history?

The most important lesson was about how to manage the position volume. In the last-year Championship, I used the percentage of the current balance to open positions, so my EA could open maximum 3 positions. After each profitable trade, it increased the volume, so it got a larger loss at the next losing trade than it would get when trading with a fixed amount of lots.

Although I’m using the same method in this present Championship, it does not suit for an Expert Advisor working in real. For example, the working amount of lots may be increased after a series of losing trades.

Moreover, I decided to change for pending orders in this contest EA.

What lessons have you learned from this Championship?
So far, it’s only one lesson – one has to work hard. Better’s success and his huge breakaway was another lesson. This is like a signal – we have to learn everything concerning neural networks. After I had read an article devoted to neural networks on the website of MQL4.com, I didn’t see any prospects in this field of trading. I was obviously wrong.

Reporting the 10th Week of the Championship (3-9 December) Ten weeks of the Championship are behind us. Only two weeks are left until the contest ends.	Reporting the 11th Week of the Championship (10-16 December) Eleven weeks are already behind us. Here is the final straight.
Previous	Next