The advantages of algorithmic trading and the importance of market data analysis

Annotation: The article describes the basic concepts of algorithmic trading that is widely used in the banking sector and cryptocurrencies market. The advantages and crucial points of hedge funds, platforms for trading are also depicted in this work.

Key words: algotrading, algorithmic trading, traders, hedge funds.

The concept of algorithmic trading

The concept of algorithmic trading has two basic definitions:

1) Algotrading. It is an autosystem, that can trade without a trader in accordance with the algorithm set for it. Such system is used to obtain profit by means of automatic market analysis and position opening. This algorithm is also called "trading robot" or "advisor".

2) Algorithmic trading. It is used to describe the process of large orders execution on the market, when they are automatically divided into parts and gradually opened according to specified rules. [1]

To simplify all facts mentioned above, algorithmic trading is the automation of day-to-day operations performed by traders, which reduces the time required to analyze stock information, calculate mathematical models, and conduct transactions.

The process automation solves the most important problem of human factor: emotionality, speculation, intuition, wrong forecasts, thinking errors. All those factors can hamper the process of gaining profit. 

The key issue of algorithmic trading is to choose and develop the rules for position opening and robot families’ settings. Those rules can be:

• manual - based on mathematics and physical models and performed by the researcher.
• automatic - necessary for mass search of rules and testing within the program.
• genetic - in this case rules are developed by a program with elements of artificial intelligence.[2]

According to ZeroHedge estimates, 84% of transactions at global exchanges are conducted with the help of high-frequency trading tools (high-frequency trading) – being the main type of algorithmic trading where specialized programs automatically search for earning opportunities, sell and buy positions within few seconds. [6]

Hedge Funds

Investment banks and hedge funds are pioneers in this area, and they need large order execution automation more than anyone else. They have successfully invested a lot of money in developing such algorithms, and as a result, there are various systems that influence the market.

The Renaissance Institutional Equities Fund (RIEF) is one of the largest hedge funds that uses algorithmic trading. It was launched by Renaissance Technologies Corp., an American investment company founded in 1982 by mathematician James Harris Simons. The Financial Times in 2006 awarded Simons the title of “the smartest of billionaires”.

Another fund, Bridgewater Associates, founded by Ray Dalio, manages $160 billion in assets, based on the quantitative investing. The company's investors' annual profit was $5 billion. [3]

The main official participants in high-frequency trading are Citadel LLC, ATD, Hill, Virtu Financial, Tradebot, Timber Chicago Trading and GETCO. However, the most active in this area are the HFT-divisions of the largest financial institutions - Deutsche Bank, Goldman Sachs, Morgan Stanley and others.

Now there are more than $ 3.5 trillion concentrated at hedge-funds - a figure comparable to the GDP of Germany and almost one and a half times more than the GDP of Great Britain. At the same time, approximately 50% of assets are concentrated in the first hundred hedge funds, which constitute a cohort of the biggest names in the industry. For example, Bridgewater Associates now has $122 billion, AQR Capital Management has $70 billion, and Two Sigma has $53 billion.

Hedge funds are widely known all over the world. Their clients are large institutional investors: pension and sovereign funds, insurance companies and other large financial institutions. Moreover, hedge funds are popular among wealthy clients who have the opportunity to invest in them through premium banks and family offices.

These funds usually attract clients with their risk-return ratio. For example, one of the large and reputable algorithmic funds, Two Sigma Spectrum, has shown the same returns as the S&P500 stock index over three years, but with much less risk. While this index was extremely volatile in some periods, the hedge fund's yields not only held the blow, but also grew. If you look at the chart since 2005, the moment when the fund was created, you can see that the Two Sigma Spectrum strategy has significantly outperformed the S&P 500 indicator. [4]

Applications and markets

The use of automatic robots is widespread on the interbank foreign exchange market. Trading advisors earned their popularity thanks to the platform MetaTrader 4 and MQL4 programming language, which allows algorithmic trading on Forex even for beginners:

• even ordinary users are able to work with his language, as a consequence, there is an algotrading guide for beginners with a full description of the language functions.

• programmed automatic advisors can be immediately launched and compiled in the format of the terminal.

• those robots do not require large computing power, a desktop computer is enough.

• the terminal has a wide range of tools for testing the robot on a large time scale.

Exchange organizations are considered to be the most interested in the development of algorithmic trading.

The most popular platforms for algorithmic trading can be presented in the following list:

1) TSLab, that has the ability to create complex algorithmic systems and to offer a practical visual range and the possibility of editing and watching the script work.

2) Wealth-Lab, with various advantages, such as implementing trading systems with the built-in strategy master, multi-systems building, development in any .NET language, strategy checking for all trading instruments.

3) MetaStock/TradeScrip – offer a big library of indicators and formulas, large number of program modules at a high speed of work.

Most of the brokerage API have interfaces in C++ and/or Java. The frequency of trade operations is the most important element of the trading engine algorithm. The robot can send hundreds of orders per minute, so the performance of the system is extremely important. If the system is not well implemented, it is inevitable that there will be a significant slippage between the price at which the order was placed and the price at which it was actually executed. This can have a dramatic effect on profitability.

Programming languages like C++/Java are considered to act as the best option in making a trading engine, but they create issues in terms of development time, ease of testing and code support. When speed is important (for example, in HFT-trading), efficient low-level languages such as C++ and even pure C are used.

Basically, two types of trading robots are developed with C++:

• Trading engine - an accessible and simple system responsible for performing easy tasks.

• Trading robot for settings control - this system is responsible for managing algorithms and editing user interface, it includes mechanisms for presenting trading results. [7]

Effectiveness of algorithmic trading in cryptocurrencies

Algorithmic trading in cryptocurrencies is gaining momentum today. Major and most reliable exchanges, including Bitfinex and Poloniex, encourage automated trading. This happens because they receive a fee on every transaction, regardless of whether the client loses or gains money.

There are various strategies that are used in crypto trading. The main ones are arbitrage, which involves making money on the difference in the price of an asset on different markets (e.g., two exchanges), and market-making, that is, playing on the rates of coins and their derivatives.

Algotrading systems are used both by professionals, including financial institutions, and by amateurs - ordinary owners of cryptocurrencies trying to multiply their capital. The solutions of such class are different in their complexity and in the principles of their construction. There are three main categories of software for working with cryptocurrency exchanges:

• bots with prewritten logic.

• learned trading robots based on AI and machine learning technologies.

• robot-advisors.

High-frequency exchange trading giants, including Jump Trading and Tower Research, have entered the world of cryptocurrencies, moreover, trading platforms based on artificial intelligence are constantly improving. [6]

Advantages and disadvantages of algotrading

The main advantage of algotrading is, first of all, the lack of disadvantages of manual trading.

There is a list of advantages of algotrading: 1) Automation of processes. 2) Absence of physical limitations and human factor. 3) Strict following of the specified program.

However, with all advantages, algotrading has certain disadvantages as well:

1) Program errors. If the programmer makes a mistake, the robot will steadily follow the wrong program and lose money. 2) Sufficient complexity of programs. When working out the algorithms it is necessary to understand not only the programming, but also the trading. It requires specialized skills and experience. 3) The lack of information in free access. 4) The lack of flexibility concerning the market changes. It is easier to adapt to rapid changes in a manual mode, than to change the entire algorithm in the program.

In conclusion, it should be noted that algotrading allows not only to increase the profit from trading, but also to reduce the trader's workload. Algotrading, along with its many variations, can be used both on currency and stock markets. Robots have their own problems, but they are still less significant than the disadvantages of the manual form of trading.

Ulangazy Askarbekov, Director of Quotex (HK) Limited