QUANTUM AI • GUIDES & RESEARCH
How automated trading works
Separate data, rules, execution and oversight when evaluating a system.
Four parts to identify
An automated process normally has a source of data, a method for making decisions, a way to send orders and controls for monitoring its behaviour. Evaluating one part does not establish the quality of the others. An attractive interface tells you little about order execution or risk limits.
Rules and models
A rules-based strategy follows conditions specified in advance. A trained model derives a relationship from examples. Either approach can fail when assumptions are wrong or conditions change. Ask which inputs the system uses and how missing, delayed or erroneous data is handled.
Execution is a separate problem
A signal is not a completed trade. Order types, availability, fees and changing prices affect the result. Check what happens when an order is rejected, partially filled or duplicated. A simulation should explain which of these conditions it represents.
Oversight remains necessary
Identify who can pause the system, change its limits or access the account. Write down the conditions under which you would stop it. Automation does not remove the need to understand the product or the authority granted to an application.