Backtest Engine

A modular Python backtesting framework for evaluating trading strategies. Supports customizable entry/exit signals, multiple position sizing methods (fixed fractional, Kelly criterion), commission and slippage modeling, comprehensive performance metrics, and equity curve visualization.

Features

• OHLCV Data Loading — Accepts CSV files or pandas DataFrames
• Modular Strategy System — Plug in any strategy function that returns buy/sell/hold signals
• Performance Metrics:
  • Total Return & Annualized Return
  • Annualized Volatility
  • Sharpe Ratio
  • Maximum Drawdown
  • Win Rate & Profit Factor
  • Trade-level P&L log
• Position Sizing:
  • Fixed fractional (% of capital per trade)
  • Kelly Criterion (dynamic, based on trade history)
• Realistic Modeling:
  • Commission (as % of trade value)
  • Slippage (as % of fill price)
• Equity Curve Visualization — Saves as PNG with buy/sell markers
• 4 Built-in Strategies: SMA crossover, RSI mean reversion, MACD crossover, Bollinger Bands reversal
• Synthetic Data Generator — For quick testing without external data

Installation

cd backtest-engine
pip install -r requirements.txt

Usage

Quick Start (with synthetic data)

python engine.py

This runs all 4 built-in strategies against randomly generated OHLCV data and saves equity curve charts:

[*] Generating sample OHLCV data...

=======================================================
  Strategy 1: SMA Crossover (fast=10, slow=30)
=======================================================
=======================================================
  BACKTEST RESULTS
=======================================================
  Total Return:               12.45%
  Annualized Return:           6.28%
  Annualized Vol:             18.32%
  Sharpe Ratio:                0.23
  Max Drawdown:              -15.67%
  Win Rate:                   42.30%
  Profit Factor:               1.34
  Total Trades:                 26
  Winning Trades:               11
  Losing Trades:                15
=======================================================
[+] Equity curve saved to equity_curve_sma.png
...

Using Your Own CSV Data

from engine import BacktestEngine, load_csv
from strategies import sma_crossover

# Load your data (must have: date, open, high, low, close, volume)
df = load_csv("my_data.csv")

# Create engine with 1% commission and 0.05% slippage
engine = BacktestEngine(
    df,
    strategy_fn=sma_crossover,
    initial_capital=50_000,
    commission=0.001,
    slippage=0.0005,
    position_size="fixed",
    position_size_pct=0.25,
)

# Run with strategy parameters
result = engine.run(fast=10, slow=30)
print(result.summary())
engine.plot_equity_curve("my_strategy.png")

Writing a Custom Strategy

import pandas as pd

def my_strategy(df: pd.DataFrame, threshold: float = 0.02) -> pd.Series:
    """
    Must return a pd.Series with values: 1 = buy, -1 = sell, 0 = hold.
    Index must match df.index.
    """
    close = df["close"]
    signals = pd.Series(0, index=df.index, dtype=int)

    # Example: buy when price drops > threshold, sell when up > threshold
    returns = close.pct_change()
    signals[returns < -threshold] = 1
    signals[returns > threshold] = -1

    return signals

Built-in Strategies

Strategy	Parameters	Logic
sma_crossover	fast=10, slow=30	Buy when fast SMA crosses above slow; sell on cross below
rsi_mean_reversion	period=14, oversold=30, overbought=70	Buy when RSI recovers from oversold; sell when drops from overbought
macd_crossover	fast=12, slow=26, signal_period=9	Buy when MACD line crosses above signal; sell on cross below
bollinger_bands_reversal	period=20, std_dev=2.0	Buy at lower band touch; sell at upper band touch

CSV Data Format

Your CSV must include these columns (lowercase):

date,open,high,low,close,volume
2024-01-01,100.0,102.5,99.0,101.2,1500000
2024-01-02,101.2,103.8,100.5,102.7,1800000
...

Performance Metrics Explained

• Sharpe Ratio: Risk-adjusted return (excess return / volatility). > 1 is good.
• Max Drawdown: Largest peak-to-trough decline in equity.
• Win Rate: Percentage of trades that ended profitably.
• Profit Factor: Gross profit / gross loss. > 1.5 is strong.

Dependencies

• pandas — Data manipulation
• numpy — Numerical computations
• matplotlib — Chart generation

License

MIT

---

Created by Israel Romero Apo | israelromero.xyz