Interview with Alexei Ermolaev (Zyx40)

Alexei Ermolaev (Zyx40) is a newbie in automated trading. He does not have much experence in it. He works with statistical approaches to analysis, so he decided to apply these methods in developing his own EA. "I began trading with ready-made indicator-based Expert Advisors, but I learned very soon that they did not have prospects. The matter is that an indicator sees events in the market very limitedly. It does not work at the points where it diverges from the market... I think that success depends on the combination: indicator + statistics."

Hello, Alexei. Could you, please, tell us a little about yourself? How did you start trading, when and why did you become interested in it?
I became interested in trading Forex six months ago. I work as a maintenance engineer and research statistical approaches to data analysis. By the way, sI use some of those methods in my historical data analyses on Forex. The reason why I concern automated trading is that the methods I use can be implemented only through a computer and only using relevant software.

What success have you achieved? Do you trade live?
Well, it's early to speak about success. Yes, I do trade live, but this is more for testing my strategies than for earning money. So far, generally speaking, I'm still in the hole.

How long have you been trading automatically?
Since about May, 2007. I began trading with ready-made indicator-based Expert Advisors, but I learned very soon that they did not have prospects. The matter is that an indicator sees events in the market very limitedly. It does not work at the points where it diverges from the market. However, the market can develop in many scenarios, so these diverged scenarios result in losses. In a sense, I'm still in a search of a method or a combination of methods for my trading.

What could be the way out, in your opinion?
I think the multipronged attack on the problem would be a way out. I'm now researching statistical techniques, doing my best to get positive results. And I also suppose to attach some indicators in future. There are many fields to be researched. Generally, I think that success depends on the combination: indicator + statistics.

Well, I think everything is clear about indicators. Statistics will help to watch in history, what price changes at certain moments result in this price's movements in the nearest future. This was the basis, by the way, on which I built my contest EA. Plus some parameters that improve the results.

What do you think of statistical reports published on the Championship website? Are they helpful, in your opinion?
No doubt. The statistics given in the reports shows both visible data, such as profits and losses and the Expert Advisors' behavior during trading. The only thing I'd add to these reports is a kind of Expert Advisor's coefficient of efficiency. It is very possible that, to this index, absolutely other EAs will be in the top five than those now being in it.

Let's go to your EA now. What is its analytical part? What is it based on?
Since my EA trades using statistical data, it is necessary to obtain these data preliminarily. I process data starting from year 2000. So I specify constant inputs, currencies, trading time, position holding time. Then I define the period of history, on which the condition of price difference is met. If the condition is met, enter a short or a long position at SL and TP computed in the same way.

What signals are used to open positions?
Buy, if the difference between bar 2 and bar 1 exceeds delta. Sell if the difference between bar 1 and bar 2 exceeds delta. Figures 2 and 1 are, of course, conditional.

The strictly defined levels of Stop Loss and Take Profit - is this optimization result? Why did you choose this ratio between SL and TP?
Yes, this was optimization that resulted in these parameters. They are crucial. As of the moment, I have the ratio between SL and TP as close to one.

How are short and long positions opened - independently of each other? Or does the EA check the direction of the position already opened?
New positions are opened independently of those already existing.

Is the result of the last closed trade considered when making the decision of opening a new position?
I thought I should consider these interrelations between new positions and those already closed. However, I did not manage to realize this in my EA because of the lack of time. The Expert Advisor had been submitted only 5 minutes before registration was closed. I have already made some conclusions about the direction, in what I could develop my EA. This is what I'm doing now.

Can you describe this direction in more details?
Yes, of course. If the time is set strictly (18:00), you cannot trade during the rest of the day. Having created several entering points like this - say, 4 or more - you can establish interrelations between trades. Having Buy as the latest position, we get a new signal. If the signals are the same, we buy some more. If they are opposite, we can leave our position unaltered or make a reversal. Here, we have to consider what is better in each specific situation.

Why is there no money-management system in your Expert Advisor? Why is the volume set strictly for 4 lots?
As I mentioned before, the ratio TP/SL is crucial. If I had implemented a money-management system in my EA, it would have completely lost its deposit at such ratio. An Expert Advisor without money management and with the volume of 4 lots minimizes the risk in time, i.e., the longer it trades the smaller is the risk to lose my deposit. Generally, it was aimed at reaching the profit of the last-year Winner. The task is now to keep the gain unchanged. Well, perhaps, even to increase it a bit.

All trades are opened at 18:00, server time. Why?
It's a small matter. At my first calculations, the EA having such inputs was more profitable. Knowing that trading sessions start at different times throughout the world and that trades are of different intensity, you can, so to say, see a "wave" running through trading floors, and follow it. Everything in this world is interrelated. So, perhaps, the changes happened a few hours ago may result in changes happening now or those that will happen relatively soon.

Who of your competitors in the Championship do you consider to be your arch-rivals?
Well, I may be mistaken, but these are not today's leaders at all. I would like to particularly note ttauzo and Damian. In my opinion, their coefficients of efficiency are higher than the others'.

What is your opinion about MiltonWaddams?
His trading method is based on long position holding period. This is called long-term investment. The decision had not been made by the EA, I think, but by MiltonWaddams himself just before he submitted his Expert Advisor. He has real chances to win, in case there won't be any rollbacks on his currency pair on the 20th of December, as it is the case now.

How do you estimate your chances to win the Championship?
Of course, I hope to. However, I see that there are developers there who are much stronger than I am. In any case, as a newbie, I consider this Championship to be a good test for my ideas.

Thank you very much for your interesting interview, Alexei.

Created: 2007.11.16  Author: MetaQuotes
Statistical Report #2

We go on collecting and analyzing statistical data obtained in the course of the Automated Trading Championship 2007. The results of the second statistical report clearly indicate the changes in views of Expert Advisors' writers after the Automated Trading Championship 2006.

Reporting the 7th Week of the Championship (12-18 November)

The 7th week of the Championship ended with several surprises. The first place has been taken by a Participant from Ukraine winning “by a head”. The last-weeks leader has given his place to others and gone somewhere beyond the first page.

Previous Next
To add comments, please, log in or register
2007.11.19 13:32