Why is this necessary and what are they used for?
Jay Yang is a well-known developer of software for algo trading. The code of its trading robots is often used when developing its own systems.
In this article, we will consider its classification of algorithms.
To read more nesessery like broker reviews follow https://scampond.com/blockchain-trading-structure-of-database
Young identified 4 basic types of algorithms:
-algorithm for executing transactions for the purchase and sale of an asset;
-behavior factor algorithm;
-algorithm for high-frequency scalping of transactions;
-predictive algorithm.
Execution of orders.
The need to reallocate equity between the various savings in the portfolio forces large funds to apply suitable algorithms for this purpose. For many years, the algorithms TWAP — the average price, which is weighted in the time interval, and VWAP — the average price of the volume have been showing themselves effectively.
The general principle of operation of such algorithms:
-evaluate the market phase for a given time;
-place buy and sell orders sequentially.
To avoid dependence on a sharp jump in price, it is much more effective to break the entire capital into several parts.
The behavior factor algorithm.
Here we study the reaction of competing traders who work on the same platform. If the instrument has low liquidity, most likely, only one major player dominates, which interacts with all the others. In such a situation, it is quite easy to identify the behavior of all participants, rather than having absolutely equal traders in terms of capital. To develop a system that will be able to profit from the actions of competitors, it is enough to understand what rules they follow and in which cases they deviate from them.
Scalping.
Execution of transactions in the narrowest time interval to obtain the minimum profit and loss. The price fluctuation by a thousandth of a fraction seems insignificant, but it is this difference that forms the basis of high-frequency scalping. A robot program comes into play, which is designed to catch absolutely every oscillation. In fact, it is not possible to conduct such transactions manually. This algorithm is the fastest possible.
Predictive algorithm.
Such an algorithm involves collecting and analyzing data on price behavior in the past. The more tick data, the more accurate the forecast will be. After that, I add fresh ticks and use any other secondary information. They actively use fundamental analysis and technical analysis, which consists of price patterns, a return to the average price value in a given period, trend compliance on high timeframes, and portfolio rebalancing.
The return to the average price is the price to which the market constantly returns in a given time range. The algorithm that uses this technique tries to create a normalized price pattern based on its behavior in the past.
A price pattern is a variety of conditions repeated over and over again, always based on the price behavior in the past. Such patterns can be seen on the chart even visually, without using the program.
The algorithm that follows the long-term trend should identify who is currently opening positions.
The whole point of rebalancing investments boils down to just one idea — to reduce the risk of getting a losing position. Here, the entire amount of capital is divided into several parts and transactions are made on completely different trading instruments.
This classification is not complete and does not include, for example, machine learning algorithms. Although, they are the future of all algorithms. But, the types presented are enough for a novice trader to be able to evaluate all the variety of price forecasting methods for working on the exchange.
