Configuration

Stochastix uses a hierarchical configuration system that allows for sensible global defaults which can be easily overridden for specific backtest runs. The primary configuration file for your project is config/packages/stochastix.yaml.

This document outlines the available parameters within that file. Any of these settings can be overridden at runtime via CLI options or API request parameters.

Global Defaults (stochastix.defaults)

This is the main root key for all global backtesting parameters.

bc_scale

• Description: The default scale (number of decimal places) for all high-precision bcmath operations throughout the framework.
• Type: integer
• Default: 8

stochastix:
    defaults:
        bc_scale: 8

initial_capital

• Description: The starting capital for each backtest.
• Type: float
• Default: 10000.0

stochastix:
    defaults:
        initial_capital: 50000.0

stake_currency

• Description: The currency to use for all calculations, portfolio value, and staking.
• Type: string
• Default: 'USDT'

stochastix:
    defaults:
        stake_currency: 'USDT'

symbols

• Description: A default list of trading symbols (e.g., pairs) to use for a backtest if not otherwise specified.
• Type: array
• Default: ['BTC/USDT']

stochastix:
    defaults:
        symbols: ['BTC/USDT', 'ETH/USDT']

timeframe

• Description: The default candle timeframe for the backtest. Must be a valid timeframe recognized by the framework (e.g., 1m, 5m, 1h, 4h, 1d).
• Type: string
• Default: '1d'

stochastix:
    defaults:
        timeframe: '4h'

start_date / end_date

• Description: The default date range for the backtest, in YYYY-MM-DD format. If null, the backtest will use all available data from the beginning (start_date) or to the very end (end_date).
• Type: string|null
• Default: null

stochastix:
    defaults:
        start_date: '2023-01-01'
        end_date: '2024-01-01'

commission

• Description: Defines the commission (trading fee) model for the simulation.
• Type: array
• Default Model:
  • type: percentage
  • rate: 0.001 (which is 0.1%)

There are three commission types available:

1. percentage: A percentage of the total trade value.
2. fixed_per_trade: A flat fee for every trade.
3. fixed_per_unit: A fee based on the amount of the asset traded.

Examples:

# Example 1: 0.075% fee per trade
stochastix:
    defaults:
        commission:
            type: 'percentage'
            rate: 0.00075

# Example 2: Flat $1.50 fee per trade
stochastix:
    defaults:
        commission:
            type: 'fixed_per_trade'
            amount: 1.50
            asset: 'USDT'

# Example 3: $0.002 fee per unit of the base asset traded
stochastix:
    defaults:
        commission:
            type: 'fixed_per_unit'
            rate: 0.002

data_source

• Description: Specifies where the backtester should find its market data files.
• Type: array
• Default:
  • exchange_id: 'binance'
  • type: 'stchx_binary'

stochastix:
    defaults:
        data_source:
            exchange_id: 'okx'

Metrics Configuration (stochastix.metrics)

This section allows for fine-tuning the parameters of specific performance metrics.

beta

• Description: Configures the Beta calculation.
• Type: array
• Parameters:
  • rolling_window: The number of data points to include in the rolling window for calculating the covariance and variance needed for Beta.
  • Type: integer
  • Default: 30

stochastix:
    metrics:
        beta:
            rolling_window: 60

Full configuration example

This is a complete example of the stochastix.yaml configuration file with all possible keys set to their default values.

stochastix:
    # Global default settings for all backtests.
    # These can be overridden via CLI options or API requests.
    defaults:
        # The default scale (number of decimal places) for all bcmath operations.
        bc_scale: 8

        # Default trading symbols.
        symbols: ['BTC/USDT']

        # Default timeframe (e.g., 1m, 5m, 1h, 1d, 1w, 1M).
        timeframe: '1d'

        # Default start date (YYYY-MM-DD). If null (~), the backtest uses all available historical data.
        start_date: ~

        # Default end date (YYYY-MM-DD). If null (~), the backtest runs to the end of the data.
        end_date: ~

        # Default initial capital.
        initial_capital: 10000.0

        # Default currency for staking and reporting.
        stake_currency: 'USDT'

        # Default stake amount per trade. Can be a fixed value (e.g., 1000) or a percentage ("2.5%").
        # If null, the strategy must handle position sizing itself.
        stake_amount: ~

        # Default commission (trading fee) model.
        commission:
            # Type can be 'percentage', 'fixed_per_trade', or 'fixed_per_unit'.
            type: 'percentage'
            # Rate for 'percentage' (e.g., 0.001 for 0.1%) or 'fixed_per_unit'.
            rate: 0.001
            # Amount for 'fixed_per_trade' (e.g., 1.5 for $1.50).
            amount: ~
            # The asset in which the commission is charged (e.g., USDT).
            asset: ~

        # Default data source configuration.
        data_source:
            # The exchange identifier used for locating data files (e.g., 'binance', 'okx').
            exchange_id: 'binance'
            # The type of data source. Currently, only 'stchx_binary' is supported.
            type: 'stchx_binary'
            # Options specific to the 'stchx_binary' data source type.
            stchx_binary_options: []
            # Placeholders for future data source types.
            csv_options: []
            database_options: []

    # Configuration for specific performance metrics.
    metrics:
        # Configuration for the Beta calculation.
        beta:
            # The number of data points for the rolling window calculation.
            rolling_window: 30