Build your own AI-powered trading system with Claude Code's specialized agents!
This is a live trading experiment with real money! Follow along on YouTube: AI Business Lab
Starting with $1,000 CAD in a TFSA account, we're using AI-powered analysis to trade small-cap stocks. Every trade, decision, and lesson learned is documented here and on the YouTube channel.
- Starting Balance: $1,000 CAD
- Current Value: $1,043.32 (+4.33%) ๐
- Benchmark Performance: Beating SPY by 4%, IWM by 2.4%
- Win Rate: 100% (4/4 positions green)
- Current Progress: 4.33% / 15% (28.9% complete)
- Days Elapsed: 2 / 30
- Target Portfolio Value: $1,200
- On Track: โ YES (Ahead of schedule!)
| Symbol | Shares | Entry | Current | P&L | Status |
|---|---|---|---|---|---|
| CHPT | 26 | $10.78 | $12.05 | +11.7% | ๐ Strong |
| EVGO | 82 | $3.63 | $3.94 | +8.6% | โญ Steady |
| TDUP | 19 | $10.45 | $10.51 | +0.6% | ๐ New |
| FUBO | 55 | $3.67 | $3.76 | +2.5% | ๐ Momentum |
- FCEL: Sold at $4.26 (+3.4%) - Profit: $13.42
- Focus: High-growth momentum plays + aggressive position management
- New Goal: 10x portfolio in 6 months ($1,000 โ $10,000)
- Risk Management: Trailing stops on winners, -5% fixed stops on new entries
- Next Targets: IONQ (quantum computing), ASTS (satellite)
This project leverages Claude Code AI agents for:
- Market Analysis - Screening small-cap opportunities
- Risk Management - Position sizing and stop-loss management
- Technical Analysis - Entry/exit signals
- Order Generation - Daily trade planning
- Alpha Vantage MCP - Live stock quotes, technical indicators, and market data
- Use
scripts/alpha_vantage_client.pyfor programmatic access - Real-time price updates for portfolio tracking
This project is a riff on ChatGPT Micro Cap Experiment by LuckyOne7777, adapted for:
- Canadian TFSA trading rules
- Claude Code AI agents (instead of ChatGPT)
- CIBC Investor's Edge platform
- Risk-managed position sizing with trailing stops
open dashboard.html- Live position tracking with P&L
- Market comparison vs Russell 2000 (IWM) and S&P 500 (SPY)
- Dark mode for any-time monitoring
- Position management tracker with action triggers
- Daily trading checklist
- Automatic benchmark tracking shows if we're beating the market
small-cap-screener- Finds momentum opportunitiesrisk-manager- Enforces position limitsquant-analyst- Technical indicators & signalsmarket-researcher- Fundamental analysis
# Record trades from broker (automatically tracks benchmarks!)
python PASTE_TRADES_HERE.py
# View benchmark comparison
python scripts/benchmark_summary.py
# Generate daily orders
python scripts/order_generator.py
# Screen for opportunities
python scripts/stock_screener.pyEvery trade automatically captures IWM and SPY prices for real-time performance comparison:
- Records benchmark prices at exact trade time
- Calculates equivalent benchmark shares
- Shows if you're beating the market (alpha)
- Updates dashboard with live comparison
# Clone and setup
git clone https://github.com/SteppieD/tradinggame.git
cd tradinggame
python setup.py # Interactive setup wizard- โก QUICK_START.md - Get trading in 5 minutes
- ๐ SETUP.md - Complete setup guide
- ๐ค CLAUDE_AGENTS.md - AI agent documentation
tradinggame/
โโโ dashboard.html # Trading dashboard with position tracker
โโโ PASTE_TRADES_HERE.py # Easy trade entry system
โโโ scripts/
โ โโโ stock_screener.py # Find momentum plays
โ โโโ order_generator.py # Create daily orders
โ โโโ portfolio_tracker.py # Track performance
โโโ agents/ # Claude Code AI agents
โโโ data/ # Portfolio & market data
- Maximum 10% of portfolio per position
- No more than 3-5 concurrent positions
- $6.95 commission per trade (CIBC)
- 10% trailing stop-loss on all positions
- Move stop to break-even at +5%
- Trail stop 5% below high at +10%
- Consider profits at +15%
- Clean energy sector (EV charging, hydrogen)
- Momentum plays with volume spikes
- Small-caps with institutional interest
New videos posted regularly on AI Business Lab:
- Trade recaps and analysis
- AI agent development
- Strategy discussions
- Win/loss breakdowns
This is an educational experiment with real money at risk. This is not financial advice. Trading involves substantial risk of loss. Past performance does not guarantee future results. Always do your own research.
Have ideas for improving the trading system? Feel free to:
- Open an issue
- Submit a pull request
- Comment on the YouTube videos
- Share your own trading experiments
MIT License - feel free to fork and adapt for your own trading experiments!
- YouTube: AI Business Lab
- Original Inspiration: ChatGPT Micro Cap Experiment
- Claude Code: claude.ai/code
- This Repo: github.com/SteppieD/tradinggame
Building AI trading systems, one experiment at a time. Subscribe to follow the journey! ๐