High-frequency trading refers to computerized trading, which seeks profits from extremely short-term market changes that people cannot take advantage of.
For example, there is a slight difference between the purchase price of securities and the purchase price, or there is a slight difference in the stock price between different exchanges.
The speed of this transaction is so fast that some trading institutions put their "server farms" near the computers of the exchange to shorten the distance for trading orders to reach the exchange.
2. Characteristics of high-frequency trading system
(1) Trading instructions are completely sent by computers, and the response delay to market data is in the microsecond level, and some even in the nanosecond level;
(2) The system consists of special software and hardware;
(3) The hardware of the system needs to be placed near the switching host, which is called co-location.
3. Two core elements of high-frequency trading.
(1) One is a trading strategy that generates high-frequency trading signals;
(2) The second is the algorithm to optimize the transaction execution process.
Extended data
1, characteristics of high frequency trading system
High frequency system is a very characteristic computer application. At the input-output level, the data is relatively simple.
The input is all market data, including the tick level and even finer granularity, such as the data in the order book.
The output is to report to the exchange, and the frequency will be relatively high at the implementation level, and it is possible to report to the exchange frequently in large quantities. When the system is running, the signal source is the real-time market broadcast by the exchange, which needs the fastest speed to disassemble, calculate and output the signal, and also requires high real-time computing ability of the system.
At the same time, the general high-frequency trading system is relatively simple from the logical level.
2, the choice of programming language
At present, the most mainstream high-frequency trading system is C/C++ language.
This is an advantage and an excellent language. Compared with JAVA and Python which rely on virtual machines, C/C++ is a development language very close to the underlying hardware. Compared with other languages, it has more control and flexibility on hardware operation and stronger control on performance.
However, its grammar is quite complicated and difficult to learn, and it is difficult for developers without systematic programming training to master it.
At the same time, using C/C++ programming can also achieve excellent performance, which is very important for high-frequency trading systems! Moreover, most domestic exchanges provide C++ class libraries, and only by developing with C++ can the system be easily docked.