A complete guide to algorithmic trading

Learn about algorithmic trading, how it works, and how to connect your MetaTrader 4 (MT4) account with Capital.com.
What is algorithmic trading?
Algorithmic trading, often referred to as ‘algo trading’, relies on computer algorithms to automatically execute trades. Algorithms may reduce human error by removing emotional bias from trading.
Rather than manually placing buy or sell orders, algorithmic trading software makes decisions based on predefined conditions. These conditions could involve market indicators such as price, volume or time. Once these conditions are met, the algorithm will execute the trade instantly, provided there is sufficient liquidity.
With Capital.com, you can connect your MetaTrader 4 (MT4) account to algo trade CFDs.
Types of algorithmic trading
There are several types of algorithmic trading, and each operates differently. They range from breaking down large trades to minimise market impact, to exploiting market inefficiencies.
Execution algorithms
Execution algorithms include VWAP (Volume Weighted Average Price) and TWAP (Time Weighted Average Price), which are designed to carry out large orders with minimal market impact. They achieve this by breaking down large trades into smaller ones, executed over time, reducing costs such as slippage and helping to secure the best possible price.
Profit-seeking algorithms
Profit-seeking algorithms aim to maximise returns by identifying inefficiencies, patterns or statistical arbitrage opportunities in the markets. Often used in high-frequency trading (HFT) strategies, these algorithms can be less transparent than execution algorithms, as traders or firms may keep their proprietary strategies secret.
Black-box algorithms
Black-box algorithms refer to algorithms where the internal logic, code or rules aren’t transparent or easily understood by users. They’re often built using complex statistical models like machine learning or neural networks, where the relationships between inputs and outputs are not always clear.
Open-source algorithms
Open-source algorithms are those whose code and logic are fully accessible, available and modifiable by the public, unlike black-box algorithms. Users can inspect, modify or improve them as they see fit.
What is an algo trader?
Algo traders are market participants who use algorithms to automate their trades. Historically, algorithmic trading was exclusive to large financial institutions with access to high-powered systems and technical expertise.
Today, however, platforms like MT4 make algo trading more accessible with advanced strategy building, automation, and backtesting tools that don’t require advanced coding skills.
Many algo trading platforms offer pre-built trading algorithms – often called Expert Advisors on MT4 – with parameters that you can customise based on your trading strategy and risk tolerance. Alternatively, traders with programming knowledge can develop algorithms using languages such as Python or MetaQuotes Language 4 (MQL4).
How does algorithmic trading work?
Algorithmic trading works by using predefined rules and computer algorithms to automatically execute trades based on market data. These characteristics are shared with qualitative trading methods, and as a result, the two can overlap.
Trading algorithms scan real-time market information to identify potential trading opportunities and place orders almost instantly. By removing human emotion and minimising error, algorithmic trading allows trades to be executed with greater precision and speed.
For example, a trader might create an algorithm based on technical analysis, such as moving averages or price patterns. When the market conditions match the rules set by the algorithm, it triggers a buy or sell trade without the need for manual input.
Algorithms can be applied across various markets and asset classes, including stocks, forex and commodities. Some traders use algorithmic trading as part of a high-frequency trading (HFT) strategy to execute numerous trades within milliseconds and respond swiftly to price fluctuations in fast-moving markets.
Algorithmic trading – pros and cons
Benefits – algo trading is faster and potentially more efficient than traditional trading methods, removing delays and emotional bias. Algorithms can execute trades at optimal moments based on set criteria, responding rapidly to shifts in market conditions. You can backtest your algo strategies against historical and live data.
Disadvantages – algorithms are created by people, so the risk of human error remains. A small mistake in coding or strategy could lead to substantial losses, even after thorough testing, because past performance does not guarantee future results.
What’s the difference between algorithmic trading and automated trading?
Algorithmic trading and automated trading are often used interchangeably, but they have distinct meanings.
Algorithmic trading involves automatically executing trades based on defined rules and criteria – such as asset price, volume, or relationships between correlated markets. These systems use technical analysis and statistical models for trading decisions.
Automated trading is broader, covering any system where trades are executed without human intervention, regardless of whether algorithms or predefined strategies are used. This includes basic functions such as limit orders and stop-losses, which trigger automatically when certain conditions are met.
Quantitative trading also uses algorithms and statistical models to identify market opportunities. Here’s more on the focus, tools and usage for each approach.
Algorithmic trading strategies
Algorithmic trading strategies involve computer algorithms designed to automatically execute trades based on predefined rules. These techniques provide a disciplined, data-driven approach that can be customised to your trading preferences and risk tolerance.
Common algorithmic trading strategies include:
Statistical arbitrage strategy
Statistical arbitrage uses statistical models to execute trades based on temporary deviations in the historical price relationship of two or more correlated assets.
When a price divergence occurs, and the algorithm determines it is unlikely to persist, it opens trades based on mean reversion theory – assuming the assets will eventually return to their historical price relationship.
Example: let’s say you create a statistical arbitrage algorithm that monitors the CFD prices of two highly correlated commodities. When the price of Commodity A rises while Commodity B remains static, the algorithm takes a long position on Commodity B and a short position on Commodity A, anticipating their prices will reconverge.
Volume Weighted Average Price (VWAP)
Volume Weighted Average Price (VWAP) strategies aim to execute large orders over time with minimal market impact. VWAP is calculated as the average price of an asset during a period, weighted by volume. The algorithm seeks to execute trades in line with this average price.
This strategy can be useful in conditions where placing a large trade could significantly move the market price. The algorithm splits the order into smaller trades, and executes them over time to reduce impact, helping to fill the order at a price close to the VWAP.
Example: imagine that you want to buy 10,000 Apple share CFDs but you’d prefer to avoid pushing the price higher by placing a large single order. The algorithm uses VWAP to break the trade into smaller orders over several hours, executing each one at a price that reflects the stock’s volume-weighted average throughout the trading day.
Time Weighted Average Price (TWAP)
The Time Weighted Average Price (TWAP) strategy is similar to the VWAP strategy but focuses solely on time rather than volume. yThe algorithm splits an order into equal-sized trades that are executed at regular intervals over a specific time period, aiming to achieve an average price while minimising market impact.
TWAP is often used in situations where traders want to minimise market impact and avoid influencing sentiment by placing a large order all at once.
Example: A trader wants to exit a large position in a low-liquidity forex pair CFD. To minimise market disruption, they set up a TWAP algorithm to split the trade evenly into smaller parts and execute them at regular intervals, helping secure the best possible average exit price.
Steps to start algo trading CFDs
MetaTrader 4 (MT4) is one of the more user-friendly and popular platforms to use, due to its flexibility and extensive tools. You can connect your MT4 account to Capital.com and begin algorithmic trading.
You could build your own algorithms using MT4’s built-in coding language or choose from numerous customisable Expert Advisors (EAs)* – pre-programmed trading bots that automate your strategies.
Steps to start algo trading:
1. Open a trading account:
Sign up with an online CFD trading platform that supports MetaTrader 4, including ours, and download the MT4 platform.
2. Connect your trading account:
Once your trading account is live, link it to MT4.
3. Choose or build an algorithm:
Pick from a variety of pre-built algorithmic trading bots that you can tailor to your strategy, or develop your own using the MQL4 programming language.
4. Backtest your strategy:
Backtest your algorithm using real market data with MT4’s StrategyTester tool to refine your trading strategy before trading CFDs in real-time.
5. Monitor performance:
Adjust your strategy and algorithm based on performance or changing market conditions.