An automated trading system places and manages trades without a person clicking. It watches the market, applies a decision process, and sends orders when its conditions are met. The person’s role is to build it, monitor it, and switch it off.
The components
A data feed — prices, volumes, order book, anything else the system uses. A decision layer — the rules or model that turns data into a trade. An execution layer — the connection to the venue that places orders and handles fills. A risk layer — position limits, loss limits, and the logic that stops everything if something breaks.
Systems that skip the last one are the ones you read about.
The spectrum
At one end, a rule: if price crosses this average, buy. Fully automated, trivially simple, no judgement.
At the other, a quantitative system: statistical models predicting returns, regime detection deciding which models to trust, portfolio construction sizing positions, execution algorithms working orders into the book. Also fully automated. Entirely different in what it does.
Both are “automated trading systems.” The label describes how orders are placed, not how decisions are made.
What automation does and doesn’t fix
It removes hesitation, fatigue and inconsistency. It doesn’t add edge. A rule that loses money manually loses money faster automated. The question to ask of any automated system is what the decision layer knows that the market doesn’t — and if the answer is nothing, automation is just speed.