You can share a GPT prompt to get an intuitive view of how GPT thinks about responding to your needs while improving code output accuracy.

I need you to write code for a quantifier in JS, and your answer must be thought about and executed in strictly the following order, a total of 3 steps, which are: [Prepared to go] [Thoughts on answering] [Code is implemented] Explained: [Prepared] Fill in here based on my description, you are thinking in multiple angles, thinking about what necessary resources or elements are needed to answer my question, what are the necessary additions that I may have missed but are necessary, including but not limited to what libraries, functions, programming skills, organizational forms or trading strategies are needed, etc. Here you fill in the various preparations based on the description of the question and [preparation], you think about it step by step, listing the elements needed to answer the question and the detailed steps to answer it. [Code Implementation] After filling in your thought process here, follow the detailed steps listed in [Respond to Thoughts] to write readable, highly logical code. Next, your answer must follow the format and order above, and here is my question: [Help me write a binary equity trading strategy, while you add the necessary elements.

The above words are hints. The function is to understand the thinking process of GPT's response requirements, to intuitively judge whether it misunderstands your intent and will run away, while also constraining GPT's focus requirements, increasing its chances of outputting correct code.

The resulting graph: