I Used ChatGPT To Code Quantamental Value's Screener Backtest Against The S&P 500

And it is addictive!

I never coded before, but the experience was very enjoyable. I do see myself putting some time into learning coding, as long as I find a some niche or useful application for it.

I paid $8.99 for a week’s subscription to Genie’s ChatGPT, I asked it the following: “Write a code that backtests an equal-weighted portfolio of stocks within a specified period, with the stock list taken from an excel sheet using a specified path, and compare it against the S&P 500, using yahoo finance, rebase the results into 100, then plot the results.”

I believe the code was almost correct, but there were some small issues that took me almost three hours to fix (given my limited knowledge).

Here is the result on the backtest (without rebalancing, given the difficulty of the code… at least for me):

Python published plot of Quantamental Value Screener using initial equal-weighting, no rebalancing, screener results based on adjusted closing prices (approximately 208 stocks, excluding different share classes)

Although it does not have a rebalancing factor, you can look at it as if you have a holding period of “forever”, as Warren Buffett highlighted, which is also in line with what we do…

For your reference and indulgence, I share the final code:

---

import yfinance as yf

import pandas as pd

import matplotlib.pyplot as plt

# Specify the stocks

df = pd.read_excel('/Users/naifgsoufi/Downloads/backteststocks.xlsx')

# Create a list of stocks from the Excel file

stocks = df['Stock'].tolist()

# Specify the dates

start_date = '2010-01-01'

end_date = '2022-12-31'

# Get the data

data = yf.download(stocks, start=start_date, end=end_date)

# Calculate the portfolio returns

portfolio_returns = data['Adj Close'].pct_change().mean(axis=1)

# Calculate the S&P500 returns

sp500_returns = yf.download('^GSPC', start=start_date, end=end_date)['Adj Close'].pct_change()

# Rebase the portfolio returns to 100

portfolio_returns_rebased = (portfolio_returns + 1).cumprod() * 100

# Rebase the S&P500 returns to 100

sp500_returns_rebased = (sp500_returns + 1).cumprod() * 100

# Print the portfolio returns to 100

print(portfolio_returns_rebased)

# Print the S&P 500 returns to 100

print(sp500_returns_rebased)

# Plot the returns

plt.plot(portfolio_returns_rebased, label='Portfolio')

plt.plot(sp500_returns_rebased, label='S&P 500')

plt.legend()

plt.show()

---

Hope you enjoyed this short and interesting article!