This is a bot for trading on Tinkoff broker. It uses Tinkoff investments API

App name is qwertyo1

To run the bot, you need to have a Tinkoff account.

• Generate token for your account at the settings
• Create a file .env with required env variables. You can find an example in .env.example
• Create a file instuments_config.json with configurations. You can find an example in instruments_config.json.example
• [Optional] Create virtual environment and activate it:

  pip install virtualenv
  virtualenv --python=python3.9 venv
  source venv/bin/activate

• Install dependencies with

  pip install -r requirements.txt

• Run the bot with

  make start

• TOKEN: Your Tinkoff token. You can generate it in the settings Can be a token for sandbox or for real account.
• ACCOUNT_ID: Your Tinkoff account id. You can get it using get accounts tool. If not specified, the first account used.
• SANDBOX: Set to false if you want to use real account. Default is true.

List of instruments you want to trade along with their settings. Each list element is a dictionary with the following keys:

• figi: Tinkoff instrument id
• strategy: The strategy configuration
  • name: The name of the strategy to use
  • parameters: Parameters of the strategy. More details can be found in the documentation of the strategy

• interval_size: The percent of the prices to include into interval
• days_back_to_consider: The number of days back to consider in interval calculation
• check_interval: The interval in seconds to check for a new prices and for interval recalculation
• stop_loss_percent: The percent from the price to trigger a stop loss
• quantity_limit: The maximum quantity of the instrument to have in the portfolio

Main strategy logic is to buy at the lowest price and sell at the highest price of the calculated interval.

Interval is calculated by taking interval_size percents of the last prices for the last days_back_to_consider days. By default, it's set to 80 percents which means that the interval is from 10th to 90th percentile.

This is the tool to get your Tinkoff accounts. Useful when you don't know your account id. To run use this command:

make get_accounts

In test/strategies/interval/backtest/conftest.py you can find the test configuration. Set up figi, comission, strategy config object, and from_date offset. To run backtest use this command:

make backtest

The result is saved in test/strategies/interval/backtest/test_on_historical_data.txt

Use this command to display stats:

make display_stats

It will display the list of executed trades