Multi-time frame strategies

[jflaggs]

I have a strategy that needs to add 10 data series using AddDataSeries. If I deploy the same strategy on many different instruments, how will that affect performance? I'm hoping that NT8 is smart enough to not duplicate the same 10 data series each time the strategy is deployed on a new instrument, but I need to confirm this. For example, if my strategy logic uses AddDataSeries to add AAPL 5min bars and MSFT 5min bars and then I deploy this strategy on 50 different stocks, will NT8 internally add 50 AAPL and 50 MSFT 5min bars or will all strategy instances share the same underlying 5min bars objects for AAPL and MSFT?

[NinjaTrader_Jesse]

Hello jflaggs,

The platform will intelligently handle each of the instruments you add. Each strategy will need to do a historical data request before starting however if there is cached data it will use that. Each strategy will subscribe to live market data for each of the instruments that you are using.

[jflaggs]

Are there any performance considerations to make when scaling this across multiple strategies and instruments? For example, if each of my strategies adds 10 data series and I have 30 strategies deployed on 30 instruments that is 30 * 30 * 10 = 9000 data series all running simultaneously.

[NinjaTrader_Jesse]

Hello jflaggs,

The only consideration to make would be to remove the series for strategies that do not need to have those instruments data. If your logic requires those instruments then you would need to add the series.

[jflaggs]

I believe this is the only way to get around the hard coding limitation with the add data series method. Is there a way to dynamically trade an instrument that is not the main series? For example, if the strategy is deployed on EURJPY and I get a buy signal, I may want to buy 1 EURUSD and buy 2 USDJPY.

[NinjaTrader_Jesse]

Hello jflaggs,

All secondary series would need to be hard coded so there would not be a way to do that without making that the primary instrument.