In the afternoon, I shared my thoughts on using GPT for quantitative trading with my friends. Then my WeChat exploded 😂. I'll share it with everyone here too, hoping to gain some fans?
Let's take the latest YouTube video from @LabSpeculation (https://twitter.com/LabSpeculation), the master of speculative experiments, as an example to explain how GPT helps a novice achieve quantitative trading. Video link: https://www.youtube.com/watch?v=6DwXUADKLl8
In the video content of the speculative lab, it directly tells us which parameters are used and how to set them. After watching the video, find the most core parameters for opening positions.
After obtaining the source code, copy and paste it to GPT.
Let it rewrite into a version for backtesting.
Then gradually add opening conditions.
Don't be afraid of trouble. Confirm key information with GPT repeatedly to avoid missing any content you input and gradually improve various logics mentioned in the video.
【Pay attention】Don't make too many requests at once. GPT may be like me, single-threaded, and can only do one thing at a time, otherwise it will make mistakes. The advantage is that when you find adjustments in the middle, you can tell GPT to return to the state of version XXX, which is convenient for version control.
As GPT outputs the quantification script to us, it updates in real-time in the Pine editor of TV. Don't be afraid of errors, just paste them to GPT. It's okay, GPT will help!
If there are major issues, just paste the entire code to GPT and let it check your work... This way, we can obtain a relatively decent trading account, with a 40% win rate, so-so, and a maximum drawdown greater than the final profit.
Congratulations, everyone! At this step, even if you're a beginner, you can use GPT to write a so-so quantitative trading strategy based on steps 1 to 8! The first step is completed, and the next step is the fine-tuning part. Adjusting parameters or adjusting logic based on losing orders belongs to advanced content.
Thank you all for watching, and thanks again to @LabSpeculation (https://twitter.com/LabSpeculation) for providing good trading strategies for a long time. Thanks to @coolish (https://twitter.com/coolish) for creating a great content sharing community, and thanks to @pionexcn_com (https://twitter.com/pionexcn_com) for actively and optimistically embracing new things and being willing to take a step forward to explore. If the feedback is good, let's talk about how to optimize a quantitative strategy next time. I believe it's full of practical knowledge, and I hope it didn't choke you all...