Can You Predict Stock Bottoms? A Python Scraper Analyzes Fund Success Rates

"Bottom‑buying" (抄底) refers to purchasing a stock when its price has sharply dropped, hoping for a quick rebound. The article investigates how often this strategy succeeds by constructing a data model and using Python to scrape fund data.

Step 1: Build the Data Model

A simple model flags a day as a bottom‑buying moment when the fund’s intraday decline exceeds 3%; if the next day’s rise exceeds 2%, the trade is considered successful.

Step 2: Scrape Fund Data with Python

The author accessed a fund website, identified the API endpoint via browser developer tools, and noted the request parameters (callback, funCode, pageIndex, pageSize, startDate, endDate). To speed up collection, pageSize was increased to 10,000, and a Referer header was added to bypass anti‑scraping measures.

The retrieved data were plotted to visualize performance trends.

Step 3: Analyze a Single Fund

Applying the model to one fund, the script counted bottom‑buying instances, classified each as success or failure, and output the results.

A pie chart visualized the outcomes: 22 bottom‑buying moments, 3 successes, 19 failures, yielding a 13.6% success rate.

Step 4: Analyze Ten Funds

The same process was repeated for ten randomly selected funds over the same period, aggregating successes and failures.

The combined pie chart shows a 24.6% success probability and a 75.4% failure probability, indicating roughly one successful bottom‑buying out of four attempts.

Conclusion: The analysis demonstrates how Python can be used to scrape financial data and evaluate investment strategies, but the results are for learning purposes only and should not be taken as investment advice.