STOX — Algorithmic Stock Trading Bot

A Python-based algorithmic trading bot for US stocks (NYSE/NASDAQ) via the Alpaca brokerage API.

Strategy

EMA + RSI + MACD + Bollinger Bands — multi-indicator confluence:

Indicator	Role
EMA 9 / 21	Fast trend direction & crossover entry/exit signals
EMA 50	Long-term trend filter (only trade in uptrends)
RSI 14	Momentum filter — avoid overbought entries (40–65 zone)
MACD (12/26/9)	Momentum confirmation — histogram turning positive
Bollinger Bands (20, 2σ)	Volatility context — price position within bands
ATR 14	Volatility-adjusted position sizing & stop-loss distance

Entry (BUY)

All conditions are scored (0–100). A score ≥ 60 triggers a buy:

• Price above 50-EMA (macro uptrend)
• Fast EMA above or crossing above slow EMA
• RSI between 40 and 65
• MACD histogram positive or turning positive
• Price below upper Bollinger Band

Exit

• Stop-loss: 1× ATR below entry price (broker bracket order)
• Take-profit: 3× ATR above entry price (3:1 reward/risk)
• Signal exit: SELL score ≥ 60 (EMA death cross, overbought RSI, etc.)

Risk Management (Conservative Compounding)

Parameter	Default	Description
Max position size	5% of equity	Per stock
Stop-loss	1× ATR (≈ 2%)	Per trade
Take-profit	3× ATR (≈ 6%)	3:1 R:R ratio
Max open positions	10	Concurrent
Daily loss limit	5%	Halts trading for the day
Profit reinvestment	100%	Full compounding

Project Structure

stox/
├── main.py                    # Bot entry point / scheduler
├── config.py                  # All settings (reads from .env)
├── requirements.txt
├── .env.example               # Copy to .env and add your keys
│
├── data/
│   └── fetcher.py             # Alpaca market data (OHLCV bars)
│
├── analysis/
│   ├── indicators.py          # EMA, RSI, MACD, Bollinger Bands, ATR
│   └── signals.py             # BUY/SELL/HOLD signal scoring engine
│
├── strategy/
│   ├── base_strategy.py       # Abstract base class
│   └── ema_rsi_macd.py        # Main strategy implementation
│
├── trading/
│   ├── alpaca_client.py       # Alpaca broker API wrapper
│   ├── risk_manager.py        # Position sizing, daily loss limits
│   └── portfolio.py           # Trade tracking & performance metrics
│
├── backtest/
│   ├── engine.py              # Bar-by-bar backtesting engine
│   └── run_backtest.py        # CLI backtest runner
│
└── utils/
    └── logger.py              # Logging (console + daily file)

Quickstart

1. Install dependencies

pip install -r requirements.txt

2. Configure API keys

cp .env.example .env
# Edit .env — add your Alpaca API key and secret

Get free API keys at alpaca.markets (paper trading is free).

3. Run a backtest first (no API keys needed for data)

# Backtest 10 stocks with 500 days of history
python backtest/run_backtest.py AAPL MSFT NVDA GOOGL AMZN --days 500

# Backtest entire watchlist
python backtest/run_backtest.py

4. Paper trade (safe — no real money)

Ensure ALPACA_MODE=paper in your .env, then:

python main.py

5. Dry-run mode (scan signals only, no orders)

python main.py --dry-run

6. Go live (when ready)

Set ALPACA_MODE=live in .env and run:

python main.py

Warning: Only switch to live mode after validating the strategy with paper trading for at least 30 days.

Capital Growth Model

With conservative compounding:

• All profits are reinvested — position sizes grow as equity grows
• 3:1 reward/risk means you can lose 3 out of 4 trades and still break even
• Daily loss limit protects capital during adverse market conditions
• ATR-based sizing automatically reduces position sizes in volatile markets

Watchlist

The default watchlist contains 45 large-cap S&P 500 stocks. Customise via Config.WATCHLIST in config.py or set in .env.

Logs

• Console: real-time output
• File: logs/YYYY-MM-DD.log (daily rotation)
• Portfolio: logs/portfolio.json (persisted trade history)