Algorithmic Trading
Quick Answer
Algorithmic trading is the use of computer programs to generate, route or execute trading orders according to predefined rules or calculations. An algorithm may consider factors such as price, timing, quantity or market conditions and act automatically when its programmed criteria are met, reducing the need for manual order-by-order decisions.
How does algorithmic trading work?
An algorithm is a set of instructions that tells a computer how to perform a task. In trading, those instructions can be designed to analyse market information and determine when, where or how an order should be submitted.
The Bank for International Settlements defines algorithmic trading as automated transactions in which a computer algorithm decides order submission and execution. FINRA-related regulatory material similarly describes an algorithmic trading strategy as an automated system that generates or routes orders or order-related messages.
A trading algorithm might, for example, monitor a market price and submit an order only when a predefined condition is reached. More sophisticated systems may divide a large order into smaller orders, choose between trading venues, respond to changes in liquidity or apply a quantitative trading strategy.
Some algorithms are designed primarily for execution, meaning they seek to carry out an existing trading instruction according to specified parameters. Others form part of a broader trading strategy that determines when to buy or sell.
Algorithmic trading should not automatically be equated with high-frequency trading (HFT). HFT is a type of algorithmic trading characterised by particularly high speed and other specialised features, while algorithmic trading covers a broader range of automated approaches.
Key features of algorithmic trading
Several characteristics help explain how algorithmic trading differs from manual trading:
Rule-based operation: Decisions are made according to programmed instructions rather than requiring a person to manually submit every order.
Automated order handling: Algorithms may generate, route, modify or cancel orders based on specified conditions.
Rapid processing: Computer systems can analyse inputs and respond electronically without the delays associated with manual order entry.
Repeatability: The same programmed logic can be applied consistently when the same conditions occur.
Different levels of complexity: An algorithm may perform a simple execution task or form part of a much more complex quantitative strategy.
Human oversight remains relevant: Algorithms require design, testing, monitoring and appropriate controls because software or market conditions can behave differently from expectations.
Simple algorithmic trading example
Suppose a trader creates a simple rule for a hypothetical market:
Buy 10 units if the price falls to $50.00 or below.
The algorithm continuously monitors the relevant market data.
If the price reaches $50.00, the programmed condition becomes true and the system may automatically submit the predefined order.
If the price remains above $50.00, no order is submitted under that rule.
This is a simplified illustrative example only. It does not account for factors such as available liquidity, execution price, transaction costs, slippage, connectivity or whether an order is ultimately filled. It does not represent an FxGrow trading feature, trading strategy or expected result.
Potential benefits and uses
Algorithmic trading can be used for a variety of purposes.
It may help market participants:
automate repetitive order-execution tasks;
apply predefined rules consistently;
process market information faster than manual order entry;
divide large orders into smaller transactions;
control parameters such as timing, price or quantity;
implement quantitative strategies that would be difficult to manage manually.
Automated order routing has also contributed to the capacity and electronic processing of modern securities markets. SEC materials describe automated systems as allowing electronic routing and execution and note the central role of algorithms and smart order routers in contemporary market structure.
These potential operational advantages do not mean an algorithm will produce profitable trades.
Risks, limitations and common misconceptions
A major misconception is that automated trading removes human error or market risk. It does not. Programming mistakes, incorrect assumptions, poor data, system failures, connectivity problems and unexpected market conditions can all affect an algorithm's behaviour.
Automation can also allow errors to occur rapidly or repeatedly before they are detected. Modern regulators therefore place significant emphasis on controls around automated systems and electronic market access.
Another misconception is that algorithmic trading and artificial intelligence are the same thing. Many trading algorithms simply follow predefined rules and do not use AI or machine learning.
Even when AI is used, it cannot reliably predict future market movements. The U.S. Commodity Futures Trading Commission warns investors against claims that trading bots or AI-driven algorithms can guarantee high returns or 100% winning results.
Algorithmic trading therefore remains exposed to ordinary trading risks as well as technology-specific risks. Historical testing or successful past operation also does not guarantee future performance.