StockBot

An open-source Python backtesting engine for designing and evaluating daily stock trading algorithms against real historical market data.

Features

• Fetches up to 10 years of historical price data via the Yahoo Finance API
• Computes 10, 50, 100, and 200-day moving averages with derived indicators (slope, concavity, crossover detection)
• Position sizing via the Kelly Criterion
• Compares algorithm performance against a buy-and-hold baseline
• Visualizes results with matplotlib (price chart, moving average overlays, buy/sell markers)

Installation

git clone https://github.com/ryantcullen/stock-bot.git
cd stock-bot
pip install -r requirements.txt

Usage

python stockbot.py

Enter any stock ticker (e.g. AAPL, TSLA, IBM) to run the backtest. Type exit or Ctrl+C to quit.

Example Output

Capital: $1,247.83
Shares: 42
Buy and Hold portfolio value: $15,230.00
Buy and Hold returns: $5,230.00
Algorithm portfolio value: $16,012.49
Algorithm returns: $6,012.49

Designing Your Own Algorithm

All trading logic lives in the Decide() method of the Portfolio class. To implement your own strategy, modify three parameters based on the moving average data:

prob_profit = 0.501         # probability of profit (e.g. 50.1%)
expected_return = 1.2       # expected return multiplier (e.g. 1.2 = 120%)
modifier = 0.2              # aggressiveness (0-1); higher = larger positions

The method receives four MovingAverage objects (f1 through f4), each with attributes you can use to build your logic:

Attribute	Description
percent_difference	How far the current price deviates from the moving average
slope	First derivative — momentum / direction of the trend
avg_slope	Average slope over the window
concavity	Second derivative — whether the trend is accelerating or decelerating
avg_concavity	Average concavity over the window
above	Whether the price is currently above the moving average
flipped	True on the day the price crosses over the moving average

Architecture

Class	Responsibility
Portfolio	Manages capital, share count, order execution, and Kelly Criterion position sizing
MovingAverage	Computes rolling averages and derived technical indicators over a configurable window
Decisions	Enum for buy/sell/hold actions

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.

License

MIT

Author

Ryan Cullen