Algorithm trading can help make gains in stock markets

Technology's raison d'tre is to make businesses more efficient. One of the best examples of technology entwined with business is algorithmic trading. This article attempts to demystify algorithmic trading and explain some of the concepts.

Every trade in the market is triggered by a decision, followed by the execution. The execution of the trade has been automated by traders. Algorithmic trading deals with this decision making process. A decision to trade is based on a multitude of parameters, most of which are digitised and hence are programmable.

Also, these decisions are based on mathematical models that are built by subject matter experts. Algorithmic trading is the automation of these mathematical models.

Algorithmic trading, in simple words, uses a 'decision support tool', which based on pre-defined parameters, analyses market data, takes decisions and executes them. It is synonymous with programmed trading or automated trading.

It helps in reducing the time taken to react to market events. This also helps in increasing the number of parallel processes that can be executed by a trader and removes the possibility of human errors and influence of emotions in decision making.

An algorithm is a list of defined instructions for calculation, data processing or automated reasoning.

Due to the automation of trading strategies, decisions are taken at a very rapid pace. The speed of execution has moved from milliseconds to microseconds and is expected to move to nanoseconds. Technology architects have been working at reducing the latency by increasing efficiency of the software, network and hardware. So the execution of the algorithms, though not directly impacting decision making, plays a key role in identifying opportunities. A significant advantage of algorithmic trading is that opportunities can be identified for all types of market trends, such as uptrend, downtrend or even a range-bound market.

Algorithms can be created to calculate and identify arbitrage opportunities across multiple segments, multiple expiries (near vs far) or multiple instruments (futures vs options).

Time-tested strategies can be programmed and set up to execute automatically. With the processors and laptop configurations available today, these strategies can be run from laptops too. One must ensure that adequate risk handling measures are available to monitor exposure. The investment in the software can be as low as Rs 10,000 per terminal or as high as Rs 1 lakh.

Algorithmic trading can also be used for high frequency trading (HFT) or quant-based trading. In HFT, the objective is to enter and exit frequently and take advantage of daily and intra-day changes.

Typically, HFT does not lead to any delivery positions and all transactions are reversed in the same day.

In case of quant-based trading, the performance over a period of time is analysed with the use of historical data and back testing. Based on this, strategies are developed, tested and optimised. Typically, these strategies would not involve a very high frequency of entry and exits and transactions would be with long-term investment objectives.

Ramesh Padamanabhan is Chief Executive Officer, NSE.IT