Writing Clear Prompts for Trading Analysis and Indicator Code
Summary
This tutorial offers practical advice for getting more useful answers from the SuperMind assistant when asking trading and coding questions. It recommends stating the task clearly, providing relevant data formats and context, using precise terminology, and checking generated answers for errors. Examples cover converting natural-language stock filters into Python, translating a technical indicator formula, and asking for a risk assessment of a MACD crossover idea.
The article’s main lesson is that detailed inputs can reduce ambiguity: specifying columns, sorting, exclusions, and desired outputs gives the assistant less room to invent structures or assumptions. It also stresses validating generated code and correcting errors. One example includes a translation of a complex indicator formula, but the displayed code is malformed and should not be treated as a reliable implementation. The tutorial provides no systematic evaluation of assistant accuracy, so its guidance is procedural rather than evidence that generated trading code will be correct.
Key ideas
- State the requested task and expected output precisely when asking an assistant for trading code.
- Provide data structure details and relevant context to reduce ambiguity in generated answers.
- Use consistent terminology for indicators and the operation being requested.
- Check generated explanations and code for factual and implementation errors.
- The example indicator translation is malformed, illustrating why generated code needs verification.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.