Strategy & Backtest

This page displays all private strategies of the user, including the strategy name, strategy type, data type, message type, tags, run records, creation time, updated time, creator, comment, etc. Users can create, edit, publish, export or delete strategies here.

Create: Create a new strategy here. You need to specify the strategy name, strategy type (stocks, futures, options), data type (snapshot, minute frequency, daily frequency), and message type (TABLE, DICT). Tags and comment can also be set for easier differentiation and explanation of the strategy. Click Create to enter the strategy editing page.

Import: Import a strategy file saved in JSON format from local system and name the strategy.

Edit: Click a strategy name or click OK after the new strategy creation to enter the editing page. You can define the event function, including initialize, beforeTrading, onTick, onBar, onSnapshot, onOrder, onTrade, afterTrading, finalize. More details about each function can be viewed by toggling the Help Docon the right.

• Code History: The last five saved versions of the code are recorded, and any two of them can be compared.
• Edit Description: Edit the strategy name, strategy types, data type, tags, and comment.
• Run: Run a backtest based on the current strategy. Enter the running page.
• Save: Save the code.

Run: Click Run to select either historical backtest or real-time backtest. Currently, this window won’t pop up for cryptocurrency.

• For historical backtest, you need to set configuration, data source, custom parameters, visualization templates, attribution and parameter combination.
• For real-time backtest, users need to set parameters and set data source.

You can quickly set the above parameters using presets. For more details on parameter settings, please refer to the Backtest Presets section.

Export: Export the strategy as a JSON format file to the local system.

Publish: Publish the strategy to the department, allowing other users in the department to use it.

• Exec History: Select the run records for approval reference.
• Approver: Select the department and approver for publishing. If the strategy has been published before, and this is a republish after modification, the department from the initial publish is mandatory.
• Encrypt Code (Users): Other users can only run the published strategy but cannot view its source code.
• Encrypt Code (Approvers): Approvers cannot view the source code during the approval process.
• comment: Add comments about this strategy for other users' reference.

Delete: Delete the current strategy.

Batch Delete: Select multiple strategies for batch deletion.

For the same strategy type and data type, users often need to fill in the same or similar parameters. The backtest presets help users quickly fill in parameters. Additionally, users can configure multiple sets of parameters to optimize parameters.

This page displays all presets of the current user, including name, backtest type, tags, data type, strategy type, comment, creator, creation time, update time, etc.

Create: Create a new preset. You need to specify the strategy type, data type, and backtest type for the preset, which determines the parameters to be filled in.

Click Create to enter the preset editing page.

Import: Import a preset file saved in JSON format from local system and name the preset.

Edit: Edit presets.

• Configuration: Basic parameters are determined by the strategy type, including backtest time, initial cash, commission, etc. Historical backtest allows adding multiple sets of parameters, while real-time backtest only supports one set of parameters.
• Data Source: Basic data source is determined by the strategy type and market data source is determined by data type and benchmark data source. Here, you can set the table information and market data required for backtest. You can set data source by specifying a table or writing custom code. The structure of the basic data source, market data source and benchmark data source has clear requirements, i.e., column names and corresponding data structures are fixed. Therefore, when using an existing table, you need to ensure the mapping relationship between the existing table and the target table; when writing custom code, the result table's column names and data structure must meet the requirements. Specific requirements can be found in the target column names and types on the From Table tab. Additionally, the calculation result must be assigned to the variable `result`.
  Note:

  In each tab of the editingpage, click Save to save the configuration on the current tab.

  
  

  

  
  

  

  

  • For the ​​benchmark data source​​:
    • For ​​historical backtest, choose between ​​existing database tables​​ or ​​custom calculations​​.
    • For ​​real-time backtest​​, specify a ​​stream table​​.

• Custom Parameters: Define any custom parameters used in event functions. For example, you can define the parameter shortWindowNused in the initialize function and assign multiple values (separated by “,”) to this parameter to optimize it.

  
  

  

  
  

• Visualization Template: Select a visualization template. If none is selected, the default template will be used for rendering.

  
  

  

  
  

• Attribution: Specify whether to perform performance attribution for the backtest. If yes, you can select an existing attribution template, fill in the corresponding parameters, and set the visual template. When Control type is Strategy Run Record Selector, you can select the current_job as Value to perform attribution for the current results.

Export: Export the preset as a JSON format file to the local system.

Delete: Delete the current preset.

Batch Delete: Select multiple presets for batch deletion.