Access

Repository (GitLab): https://gitlab.com/fsbc/theses/quantbacktest PyPI: https://pypi.org/project/quantbacktest/

Master's thesis: https://ssrn.com/abstract=3620154

Setup

Install the project via the shell: pip install quantbacktest.

Update to a newer version via the shell (do this twice!): pip install quantbacktest --upgrade.

Exemplary usage

from quantbacktest import backtest_visualizer

# Importing modules from this repository
import sys

# For managing dates
from datetime import datetime

# For allowing for flexible time differences (frequencies)
from pandas.tseries.offsets import Timedelta


display_options = {
    'boolean_plot_heatmap': False,
    'boolean_test': False,  # If multi-asset strategy is used, this will cause sampling of the signals to speed up the run for testing during development.
    'warning_no_price_for_last_day': False,
    'warning_no_price_during_execution': False,
    'warning_no_price_for_intermediate_valuation': True,
    'warning_alternative_date': False,
    'warning_calculate_daily_returns_alternative_date': False,
    'warning_no_price_for_calculate_daily_returns': False,
    'warning_buy_order_could_not_be_filled': True,
    'warning_sell_order_could_not_be_filled': True,
    'errors_on_benchmark_gap': True,
    'boolean_plot_equity_curve': False,
    'boolean_save_equity_curve_to_disk': True,
    'string_results_directory': '/home/janspoerer/code/janspoerer/tmp/results'
}

general_settings = {
    'rounding_decimal_places': 4,
    'rounding_decimal_places_for_security_quantities': 0,
}

excel_worksheet_name = 'weights'

strategy_hyperparameters = {
    'maximum_deviation_in_days': 300,
    'prices_table_id_column_name': 'token_itin',
    'excel_worksheet_name': excel_worksheet_name,  # Set this to None if CSV is used!
    # For OpenMetrics: 9.8
    'buy_parameter_space': [9.8],  # [11, 20] # Times 10! Will be divided by 10.
    # For OpenMetrics: 9.7
    'sell_parameter_space': [9.7],  # [5, 9] # Times 10! Will be divided by 10.
    'maximum_relative_exposure_per_buy': 0.34,
    'frequency': Timedelta(days=1),
    'moving_average_window_in_days': 14,
    'id': 'TP3B-248N-Q',
    'boolean_allow_partially_filled_orders': True,
    'string_file_path_with_signal_data': '/home/janspoerer/code/janspoerer/quantbacktest/quantbacktest/assets/strategy_tables/test.csv'
}

constraints = {
    'maximum_individual_asset_exposure_all': 1.0,  # Not yet implemented
    'maximum_individual_asset_exposure_individual': {},  # Not yet implemented
    'maximum_gross_exposure': 1.0,  # Already implemented
    'boolean_allow_shortselling': False,  # Shortselling not yet implemented
    'minimum_cash': 100,
}

comments = {
    'display_options': repr(display_options),
    'strategy_hyperparameters': repr(strategy_hyperparameters)
}

backtest_visualizer(
    file_path_with_price_data='/home/janspoerer/code/janspoerer/quantbacktest/quantbacktest/assets/raw_itsa_data/20190717_itsa_tokenbase_top600_wtd302_token_daily.csv',
    # ONLY LEAVE THIS LINE UNCOMMENTED IF YOU WANT TO USE ETH-ADDRESSES AS ASSET IDENTIFIERS!
    # file_path_with_token_data='raw_itsa_data/20190717_itsa_tokenbase_top600_wtd301_token.csv',  # Only for multi-asset strategies.
    name_of_foreign_key_in_price_data_table='token_itin',
    name_of_foreign_key_in_token_metadata_table='token_itin',
    # 1: execute_strategy_white_noise()
    # 2: Not used anymore, can be reassigned
    # 3: execute_strategy_multi_asset() -> Uses strategy table
    # 4: execute_strategy_ma_crossover()
    int_chosen_strategy=4,
    dict_crypto_options={
        'general': {
            'percentage_buying_fees_and_spread': 0.005,  # 0.26% is the taker fee for low-volume clients at kraken.com https://www.kraken.com/features/fee-schedule
            'percentage_selling_fees_and_spread': 0.005,  # 0.26% is the taker fee for low-volume clients at kraken.com https://www.kraken.com/features/fee-schedule
            # Additional fees may apply for depositing money.
            'absolute_fee_buy_order': 0.0,
            'absolute_fee_sell_order': 0.0,
        }
    },
    float_budget_in_usd=1000000.00,
    strategy_hyperparameters=strategy_hyperparameters,
    margin_loan_rate=0.05,
    list_times_of_split_for_robustness_test=[
        [datetime(2014, 1, 1), datetime(2019, 5, 30)]
    ],
    benchmark_data_specifications={
        'name_of_column_with_benchmark_primary_key': 'id',  # Will be id after processing. Columns will be renamed.
        'benchmark_key': 'TP3B-248N-Q',  # Ether: T22F-QJGB-N, Bitcoin: TP3B-248N-Q
        'file_path_with_benchmark_data': '/home/janspoerer/code/janspoerer/quantbacktest/quantbacktest/assets/raw_itsa_data/20190717_itsa_tokenbase_top600_wtd302_token_daily.csv',
        'risk_free_rate': 0.02
    },
    display_options=display_options,
    constraints=constraints,
    general_settings=general_settings,
    comments=comments,
)

Information for maintainers/contributors

To make changes available in GitLab and as a pip install, please first push your changes to a new branch to GitLab and merge them.

1. Update the version number in VERSION.
2. Build wheel: python setup.py sdist bdist_wheel.
3. Upload to PyPI: twine upload --skip-existing dist/*.*
4. Get the current version on your machine: pip install quantbacktest --upgrade

Maintainers can also refer to this great guide: https://realpython.com/pypi-publish-python-package/#versioning-your-package

Further reference to quant trading in general

Quantopian offers state-of-the art backtesting for quantitative trading strategies for equity markets. Their YouTube channel hosts some excellent, generally applicable talks from renowned experts:

• "When Should You Build Your Own Backtester?" by Dr. Michael Halls-Moore
• "Optimizing Trading Strategies without Overfitting" by Dr. Ernest Chan - QuantCon 2018
• "Beware of Low Frequency Data" by Ernest Chan