Vladimir Tsyrulnik: The Essense of my program is improvisation! (ATC 2010)

Interview on Automated Trading Championship 2010 from 27.10.2010.

Vladimir Tsyrulnik is the holder of one of the brightest highs of the current Championship. By the end of the third trading week Vladimir's Expert Advisor was on the sixth position. The IMEX algorithm the Expert Advisor is based on was developed by Vladimir. To learn more about this algorithm, we had an interview with Vladimir. 

Vladimir, people get acquainted with automated trading under different circumstances. Can you tell us about your first acquaintance?

In 2003 my brother Alexander who is also participating in the Championship (Tsyrus), showed me the МetaTrader 3 trading terminal and told me about its features. At that time I already new something about Forex, but my first opinion of it was very simplistic and, therefore, skeptical. I am an engineer-physicist by education. I specialized in cybernetics and automated data processing systems, i.e. I was programming much at that time. Probably this circumstance together with the trading terminal possibility to use the built-in program code for automating the trading process prompted me to have a try. I began with attempts to write simple robots. At that time I was positioning myself as a developer, not as a trader. The fact that it was a C-like language helped much.

About two years I hadn't traded, only occasionally tested my applications on demo accounts. I was not particularly worried with the results. More interesting and meaningful to me was the gradual understanding of the trading process as of a physical process, which, in my opinion, could be formalized in one form or another. And I wondered if it was reasonable to combined development, modeling and trading. Today I know my answer to this question. My first attempts to trade with Alpari were not very successful. Together with automated trading I was trading manually - I had small deposits. I was only trying, but made certain conclusions. The main of them was: if trading, then only using automated software complex. In short, I had a means, but and to use it I needed to understand the nature of the ongoing process. And I chose the path of physical analysis. Today I have about a dozen and a half developments - from classical Williams' to my own know-how.

Do you sell your applications or do you develop them only for your own use?

About 3 years I've been using my products in my own trading without any definite system though. The results were quite good. Two years ago I was invited to a company who were interested in my models. In my opinion, this is the team with the necessary and reasonable division of functions. I am deeply convinced that the trading process is not only trading of the buy-sell kind, with hedging or something like that. For me it is implementation of a certain technology of using a complex mechanism for solving technical, organizational, commercial, analytical and other tasks. Only in this form the systematic moving towards success and implementing of long-term working projects is possible. And the built-in programming languages are so useful in this task.

Although, someone like to feel alone with the nature. And I admit that it is possible to achieve some success while going alone without using any additional tools. But it is a short-term success as a rule. Once working with one of dealing centers, I had the pleasure of earning $10,000 starting with $50 within a month. But this is not the trading process.

You participated in the Championships of 2007 and 2008 with different success. What was your experience of the competitions?

To participate in these two Championships I used my very first modeled idea and the same Expert Advisor, though different versions of it adjusted to the Championship rules. In 2007 I used quite a moderate risk management, but perhaps I hadn't completed the selection of working parameters to the required degree. The following year I made an unfortunate mistake in the logics of my Expert Advisor. In general, at that time I, probably like many participants, was thinking more about a trading algorithm, not about the sporting aspect. Personally I had imagined the upcoming process as a trading in live account conditions.

By 2009 I already had some confidence in the possibility o implementing an acceptable algorithm for participating in the contest, but the Championship was not held. Of course, the Championships give a certain impulse to self-development. But for me it is primarily the promotion, in the good sense, on sporting principles.

What is the reason for choosing USD/JPY?

When choosing a symbol I was trying to answer the question: which of the elements (categories) of the process that I was describing, would be easier to model? Analyzing the nature of currency pairs and their dynamics, I determined that USDJPY is the closest one to an ideal category clearly describing the sort of systematic relationship between two most influential and most mutually dependent economic systems (the USA and Japan). By the way, even in social and historical processes this relationship is very strong. As a rule, the strongest relations are more susceptible to description.

All movements of assets affect these two financial poles. All other financial conglomerates derive from the first two ones. Simply put, there is risk and its absence, and between them all assets are moving. In physics it is called the difference of potentials. In this case, a unit (asset) behaves the most expected way. Today the USA and Japan are two potential-poles in a certain context. The reflection of the relationship is in the appropriate currency pair.

Most of your trades were closed by SL and TP, though some trades were closed without stop orders. On what principle are they closed?

In any physical process there is a time point that differs from all other points in that at this moment the system (process) goes into a new state. For example a pendulum. Empirically I've defined such a moment – this is a certain live time of a bar after which the system gets new qualitative characteristics. In our case we see the beginning of a new reverse movement. The expert Advisor simply closes a position and opens a reverse one. However, there can be several points in the whole time interval of the process.

Stop orders are placed based on the calculated range from the current daily ATR using a special correction factor.

It seems your Expert Advisor is trading inside a channel, isn't it?

I gave up a classical channel interpretation of price movement. Generally speaking, the linear representation is no good in the kinematics of prices. The quantum approach and presentation is closer to me. But, alas, it is still easier in the case of my Expert Advisor. There are no channels. By the way, Pivot-levels are no other than a metaphysical representation of the same time points discussed above. In short, these are levels, which attract that pendulum. This is a metaphysical representation of such native concepts.

In the description of your Expert Advisor, you've written that it works best on higher timeframes. Why?

Smaller categories that describe the system, give a greater percentage of "noise." That is, to describe you should choose an object commensurate with the size of the goal. If TP is greater or smaller than daily ATR, then the information carrier should have at least daily information. It is difficult to predict rain for tomorrow (goal), if we analyze the pressure, temperature, and the like within an hour. We need a whole-day information.

How does your system ignores the noise and spurious motion?

My Expert Advisor ignores the movements that have taken place between the time points mentioned above. It is just waiting for the execution of its prognosis. It's like a hunter or predator determines the direction and speed of an animal by the characteristics of its tracks. And it makes a decision using only the time factor. These "noises" must not necessarily be weak in their amplitude. But we are talking about a probabilistic movement in the current bar, that is, the only question is, what color Daily0 (current daily bar) would be.

You are always in the market. is this done intentionally?

Most of the bar lifetime the Expert Advisor is in. But there is a time limit, after which the EA assumes the forecast as outdated and residual and does not open a position. Typically, this is after 3/4 of the lifetime of a bar.

At the beginning of the Championship your Expert Advisor had a large drawdown (maximum drawdown is more  than 50%). Did you expect that?

Of course my heart sank, but I was absolutely confident in the stability of my EA. Stops worked, and they worked several times. This happened the first day of the contest on Nonfarm Payrolls (the price moved in the direction where the EA was directed, but "noise" was high at the beginning). Further events only confirmed my belief. And after all this is I started with less than $7,000.

Judging by the description of your Expert Advisor, it is based on the IMEX index. Can you tell us more about it, please?

IMEX is my working title for the index of market expectations of (Index of Market EXpectation). This is my derivative that is based and counted on the usual Bulls/Bears Power. Though it may seem simple, it is worth noting that its comparative value determines the direction of market expectations. It is based on the principle of "traces". The same hunter or predator analyzes traces (in our case, the indicators) and makes a decision based on them.

My Expert Advisor determines the set of "traces" in a digital format in the certain permitted time point, translates it into an index representation, using the formula, and then issues a "verdict" regarding the prognosis of the system: Buy or Sell. Note the importance of the time moment. For in an arbitrary time the system parameters do not determine its status. More precisely, this parameter is time. That's why doctors measure the temperature of the patient at strictly defined times. That's why meteorologists determine the required parameters of pressure, humidity and temperature also at strictly defined times. In a word, it is all physics!

By the way, instead of the mentioned indicators you can use some others. For example, the volume (OBV or Volumes) and oscillator (MACD, Stochastic).

Mean values of losses and profits of your EA are equal. What money management system does it use?

I use the same MM system as the current leader Boris. This is the classical DUD - dynamic use of deposit. The risks are fixed. The volume of a traded lot depends on the dynamics of the deposit.

Now regarding the average level of losses and wins. The strategy forces to follow "the law of the genre" quite the opposite. I increase SL relative to TP about 3 to 2 to ensure a comfortable absorption of those "noises" (the SL size is up to 70% of ATR). And with a high index of forecast performance probability we have numerical superiority of the number of profitable trades. And then goes arithmetics.

The rate of your profitable trades is now high. Is this due to the analytical part of your EA& Did you have the same results during tests&?

Yes, I had similar results in tests. On a standard test period it showed a drawdown of less than 15% and the number of profitable trades was about 80%, which is actually almost what it has today. With regard to analytics, it is just a mathematical interpretation of describing the behavior of both current and potential (the forecast - that is the main result of the mathematical apparatus of the program). All the analytics is digitization (if I may say so) of the "pictures of the process", that is what we see. This action is contained in the formula of IMEX calculation.

Why do you use the word Jazz in the name of your Expert Advisor?

Frankly speaking, I was waiting for this question. The very essence of the program is improvisation. I'm sure the formula that lies in the heart of the algorithm is unique. And its mother is the improvisation!

How can you evaluate your chances of winning the Championship?

If we assume the chance of staying on the first page as = 100%, then being the 19th = 100/2 (50%), 18th = 100/3, 17th = 100/4, … the 1st = 100/20=5%. So it would be great!

Good luck in the Championship, Vladimir!