Simple trading backtester
Project description
Overview
Quantitative approach to trading is done via applying mathematical models to various financial instruments. In order to get money for you strategy, mathematical model beneath it should be sound. And to prove that this model worth money one should do proper backtesting. This project aims to provide easy and straitforward backtesting oportunities.
Relevance
There are number of python projects for backtesting: backtrader, pyalgotrade, zipline, rqalpha, etc.. When i was trying out them, i was dissatisfied with one or more of the following: event driven, unnecessary complex architecture, no support for trading multiple instruments in convinient way, no proper performance evaluation, bugs, etc.. This project solves those issues at cost of not so wide functionality compared to mentioned ones above. Project is designed to be easily build on top of it.
Features
Data manipulations are made with pandas
Backtesting operations are vector(no loops, no event driven)
Extensive statistical evaluation of strategies
Number of visualizations embedded
Strategy robustness tests
API to work with OHLC data(download, prepare)
Clean and straitforward project structure
PEP8 compliant code
Installation
Install via setup.py:
git clone git@github.com:bluella/stbt.git
cd stbt
python setup.py install
Install via pip:
pip install stbt
Run tests:
pip install pytest
pytest
Usage
import datetime as dt
import pandas as pd
import matplotlib.pyplot as plt
from Simple_trading_backtest.simulator import Strategy
import Simple_trading_backtest.helpers as hlp
# creating fake trading data
# prepare datetime index
some_date = dt.datetime(2017, 1, 1)
days = pd.date_range(some_date, some_date + dt.timedelta(30), freq='D')
# initialize close prices
data_values = list(range(1, 32))
closes_df = pd.DataFrame({'Date': days, 'inst1': data_values})
closes_df.set_index('Date', inplace=True)
# initialize weights
weights_values = [1 for i in range(31)]
weights_df = pd.DataFrame({'Date': days, 'inst1': weights_values})
weights_df.set_index('Date', inplace=True)
# create strategy
s = Strategy(closes_df, weights_df, cash=100)
# run backtest, robust tests, calculate stats
s.run_all(delay=2, verify_data_integrity=True, instruments_drop=None,
commissions_const=0, capitalization=False)
# check strategy stats
print(s.stats_dict)
# save strategy to futher comparison
s.add_to_pnls_pool()
# plot pool correlation heatmap
heatmap_fig, corr_matrix = s.get_pool_heatmap()
plt.show()
Links
Futher development
Improve test coverage
More API for data download
More technical indicators
Portfolio optimization tools
Releases
See CHANGELOG.
License
This project is licensed under the MIT License - see the LICENSE for details.
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.