Algorithmic trading, or algo trading, means using computer programs to decide when, how much and at what price to trade, following rules set in advance. The rules can be as simple as slicing a large order or as complex as a statistical strategy.
Two different uses
| Use | Purpose | Example |
|---|---|---|
| Execution algorithms | Trade a decision already made at a better price | Splitting a large buy into small orders across the day to limit market impact |
| Strategy algorithms | Generate the trading decisions themselves | Buying when a signal crosses a threshold. See statistical arbitrage |
An illustrative execution example
A fund wants to buy 5,00,000 shares of a stock that trades 50,00,000 shares a day. Buying it all at once would be 10% of the day's volume and would push the price up. A volume-weighted algorithm spreads the order over the day in proportion to expected volume, keeping each slice small. The figures are made up.
Common building blocks
- Signal: the rule that says what to trade.
- Risk checks: limits on position size, price bands and loss before an order goes out.
- Execution logic: order slicing, timing and venue choice. See slippage modeling.
- Monitoring: alerts and a way to stop the system.
Risks
- Bugs and bad data can send wrong orders very quickly, so pre-trade controls are essential.
- Overfitted rules that worked in a test fail live. See overfitting and backtesting.
- Costs from fees and spreads add up with frequent trading. See transaction costs.
- Regulation: exchanges and SEBI set rules for algorithmic orders, and brokers must meet them.
Algo trading is about how trades are placed. Quantitative investing is about how investments are chosen, and a quant portfolio can be traded by hand or by algorithm.
Frequently asked questions
What is algorithmic trading?
It is trading in which a computer program places orders following rules set in advance, instead of a person deciding each order.
Is algo trading the same as high-frequency trading?
No. High-frequency trading is one type, characterised by extremely fast, short-lived trades. Many algorithms trade slowly, such as slicing an order over a day.
What is the difference between algo trading and quant investing?
Algo trading concerns how orders are executed or generated by programs. Quant investing is a way of choosing investments with models, which may or may not use algorithmic execution.
Is algo trading risky?
It can be. Bugs, bad data and overfitted rules can cause losses quickly, which is why risk checks and monitoring are used.