Dual time resolution

To help Sebastien manage different time scales within the same algorithm, here's a structured approach using QuantConnect's capabilities:

Step-by-Step Implementation

1. Set Up Data Subscriptions:

   • Subscribe to both hourly and minute data resolutions for the assets you are trading. Use self.add_equity to add securities with the desired resolution.

     self.add_equity("SPY", Resolution.HOUR)
     self.add_equity("SPY", Resolution.MINUTE)
     

2. Use TradeBarConsolidators:

   • Implement TradeBarConsolidators to aggregate minute data into hourly bars for your alpha model. This ensures the consolidator time period matches the desired resolution.

     from datetime import timedelta
     self.hourly_consolidator = TradeBarConsolidator(timedelta(hours=1))
     self.hourly_consolidator.data_consolidated += self.on_hourly_data
     self.subscription_manager.add_consolidator("SPY", self.hourly_consolidator)
     

3. Implement Risk Management:

   • Use the CompositeRiskManagementModel to combine multiple risk management strategies, such as MaximumDrawdownPercentPortfolio and MaximumUnrealizedProfitPercentPerSecurity, which can operate on minute data.

     from Algorithm.framework.risk import MaximumDrawdownPercentPortfolio, MaximumUnrealizedProfitPercentPerSecurity
     self.set_risk_management(CompositeRiskManagementModel(
       MaximumDrawdownPercentPortfolio(0.1),
       MaximumUnrealizedProfitPercentPerSecurity(0.05)
     ))
     

4. Schedule Risk Management Checks:

   • Utilize the Schedule.on method to run risk management checks at a minute resolution, ensuring these checks are independent of the alpha model's hourly resolution.

     self.schedule.on(self.date_rules.every_day("SPY"), self.time_rules.every_minute(), self.risk_management_check)
     

5. Handle Data in OnData:

   • In the OnData method, process both the consolidated hourly data for your alpha model and the minute data for risk management, ensuring each component operates on its respective data. ```python def on_data(self, data):Handle minute data for risk management if data.bars.contains_key("SPY"):

       self.risk_management_logic(data.bars["SPY"])
     

   def on_hourly_data(self, consolidated_bar):

   # Handle hourly data for alpha model
   self.alpha_model_logic(consolidated_bar)
   

   ```

6. Test and Validate:

   • Backtest the algorithm to ensure that the risk management model effectively monitors positions at a minute resolution without interfering with the alpha model's operations.

7. Optimize and Adjust:

   • Fine-tune the parameters of both the alpha and risk management models based on backtest results to achieve the desired performance and risk profile.

By following these steps, Sebastien can effectively manage different time scales within the same algorithm, ensuring that both the alpha model and risk management operate efficiently.

Sources: Time Period Consolidators - QuantConnect.com, Data Consolidation Algorithm, Getting Started - QuantConnect.com, A Risk Parity Approach to Leveraged ETFs, Intraday Application of Hidden Markov Models

I have a universe of 200 stocks, not a manual universe of few tickers. What should be the resolution in the main loop ?