Market and
partners news

Features of algorithmic trading

23 July 2020

Amid technological advances, stock markets have adapted very well to the exciting world of algorithms.

#Reference: Algorithmic trading is a method of executing a large order (too large to be executed at once), when, using special algorithmic instructions, a parent order is divided into several sub-orders (child orders) with its own characteristics of price and volume, and each of the sub-orders is sent at a certain time to the market for execution. Such algorithms were invented so that traders do not have to constantly monitor quotes and manually divide a large order into small ones.

To dive into the principles of algorithmic strategies in more detail, we talked with the candidate of physical and mathematical sciences, the chairman of the State Attestation Commission at the Financial University under the Government of the Russian Federation, Andrei Valentinovich Batunin.

Aravana CM: Why are you interested in this direction? Please describe your strategy.

A.V. Batunin: I am a theoretical physicist by education, graduated from the Physics Department of Moscow State University, the Department of High Energy Physics and Elementary Particles. I defended my Ph.D. thesis on hadron interaction. In the early 90s, I was interested in a new topic at that time – fractals (infinite self-similar sets with non-integer dimensions) and their manifestations in physical phenomena.

At that time, I was still engaged in physics, although many of my colleagues either worked abroad in their specialty, or changed their field of activity in Russia because of the very low salaries of scientific workers. I chose the second of these opportunities – classmates from the Physics Department of Moscow State University founded the investment and financial corporation “Russian Company” and invited me to the leadership.

In the early to mid-90s, the securities market in Russia practically did not exist (excluding vouchers), therefore, we had to gain experience with specific Russian financial instruments – treasury bonds, tax exemptions, government short-term bonds. In parallel, the company was engaged in buying up shares of privatized enterprises.

For self-education, I read a lot of translated literature on the stock market, completely unknown to me. Then, for the first time, I had the idea of the possibility of applying the methods of studying fractals to the study of the patterns of behavior of the securities market as a nonlinear dynamic system.

The idea is simple – if we can find correspondences between the characteristics of fractal sets well studied by mathematicians and the characteristics of the stock market, then we can apply the conclusions from the theory of fractals to predicting the behavior of stock indices or stock prices.

The new thing that I propose in my method is to consider the phase trajectories of a dynamic system as a subject of research, that is, the dependence of the rate of change of the stock market indicator on its current value.

The explicit dependence on time disappears here, but it remains in the form of time-ordered points (states of a dynamical system) on the phase trajectory. This is in stark contrast to technical analysis, when time series of stock quotes are considered, trends are looked for, etc. The phase trajectory, in principle, is the very fractal (its final approximation) to which the corresponding research methods will be applied – histograms of the density of points on the phase trajectory, Poincaré sections, etc. will be built.

In my articles in 2001-2005 in the magazines “Finance and Credit” and “Digest-Finance”, these studies were carried out using the example of the American stock index S&P 500, and a number of previously unknown patterns were revealed.

Our efforts are now focused on identifying explicit indicators in the language of phase trajectories, warning investors of impending stock market shocks. Such indicators can be qualitative changes in phase trajectories: splitting of one line into several, transition of a trajectory to another region of phase space, appearance of loops in Poincaré sections, etc.

Aravana CM: What advantages and disadvantages can you highlight in algorithmic trading versus conventional market strategies?

A.V. Batunin: The advantages of algorithmic (high-frequency) trading are well known. The first is to eliminate human emotions when deciding whether to buy, hold or sell. Robots are not subject to senses. The second is the speed of decision making and its execution on the exchange. It takes a man seconds at best, robots take a millisecond. Algorithmic strategies are good in a fast changing volatile market in standard situations. Non-standard situations that are not spelled out in the program can lead to failures and losses in money. Therefore, the most correct, in my opinion, is the combination of tactical operations of robots and strategic control of an experienced specialist over the situation in the market, who sees the whole picture of what is happening as a whole.

Aravana CM: Who, in your opinion, will be interested in algorithmic trading? Why?

A.V. Batunin: It will be of interest to investors who have their own original strategy that needs to be implemented into an algorithm. Or those who believe in an already existing and well-proven algorithm and want to make money with it. In any case, the investor must understand that the impact of serious cataclysms (sudden military conflicts, imposition of sanctions, environmental disasters) on the stock market cannot be accounted for by any algorithm.

In this case, human intervention is necessary namely a strategic assessment of what is happening and the appropriate conclusions and actions – to stop the algorithm or correct it.

Aravana CM: How do you see the future of this direction on a global scale? Who, in your opinion, is more efficient in the financial market – a robot or a human?

A.V. Batunin: A person writes a program for a machine, so a competition between robots is still a competition between humans and humans. The reaction speed of robots is the same, small nuances in speed are introduced by the programming language in which the program and platforms are written. The main contribution to the effectiveness (profitability) of the strategy is still made by the algorithm itself, its “zest”, which is created by a person.

How do I see the future of this direction? Artificial intelligence will be able, independently of a person, to create and introduce fake news into the global media that can bring down the stock market and neutralize the algorithmic strategies operating on it. At this moment, it turns on its own algorithm, sharpened for this situation, and skims all the cream. The main thing will be to predict exactly how the bulk of investors will react to this news – in which direction the market will swing.

Aravana CM: At the same time, you have been at the head of the State Attestation Commission (SAC) of the Financial University under the Government of the Russian Federation since 2012. What thoughts about the development of the university did you have then, and how much do they coincide with what you came to after 8 years?

A.V. Batunin: I have been cooperating with the Financial University under the Government of the Russian Federation since the late 90s, when it was still called the Financial Academy. The memory of this remained in the name of his site – fa.ru. I gave a course of lectures “The securities market” at the faculty of second higher education, then a more complex (more mathematics) course “Derivative financial instruments”.

My students were already self-made people who already had their first higher education and worked in large corporations, in ministries, in the Central Bank. I saw that the knowledge that I give them was interesting and needed in their work. Naturally, I believed that a second higher education would always be in demand, and we would not have a shortage of students. Unfortunately, I was wrong. The current economic situation does not require an increase in the number of financial specialists. A striking example: the number of banks in 20 years has decreased from one and a half thousand to four hundred.

Aravana CM: How does the work in the university resonate with your fondness to the algorithmicization of market instruments?

A.V. Batunin: It is very rare to defend theses, but there are topics related to the algorithmicization of market strategies. I remember two works that touched on the topic of fractal sets. With these students after the defense, we discussed in detail the points of interest to me, I gave them links to my articles, and recommended to continue research in this area.