What is a good time frame for backtesting your strategy

[vpatanka]

Hi Everyone,

May be this has been answered already. I did a quick search and didn't see anything come up at least in the first few pages. I was wondering for an intraday strategy that produces on an average between 1-3 trades a day what's a good time frame for back testing the strategy?

In Strategy Analyzer I can select the instrument say ES 12-22 and can backtest is going back many years but I don't know if that makes any sense or valid as the futures contracts are qtrly. I have been trading mainly manually up until now and looking at some strategies to automated and wanted to check with the pros here.

Please let me know when you can. Any pointers on this would be much appreciated. Thanks in advance

[NinjaTrader_Jesse]

Hello vpatanka,

You can use any time you would like to test the strategy, that is up to you and your goals. The platform can merge future contract data so you can select a long timespan if you would like to.

[vpatanka]

Hello vpatanka,

You can use any time you would like to test the strategy, that is up to you and your goals. The platform can merge future contract data so you can select a long timespan if you would like to.

"The platform can merge future contract data so you can select a long timespan if you would like to" answered my question. Thanks Jesse

[vpatanka]

"The platform can merge future contract data so you can select a long timespan if you would like to" answered my question. Thanks Jesse

Hi Jesse,

Is there a way to use "OnAccountItemUpdate" with strategy analyzer if I want to run my strategy to backtest much longer duration?

I am using the code below
protected override void OnAccountItemUpdate(Cbi.Account account, Cbi.AccountItem accountItem, double value)
{
// Updated Account P&L
AccountRealizedPL = account.Get(AccountItem.RealizedProfitLoss, Currency.UsDollar);
AccountUnrealizedPL = account.Get(AccountItem.UnrealizedProfitLoss, Currency.UsDollar);
}

if (TotalPL >= <amount of profit per day>)
CloseStrategy("MaxProfitClose");

which I can see working in "Replay" mode but when I backtest my strategy but not in Strategy Analyzer with much longer duration backtets. Is there a way to do that?

Thanks in advance

[NinjaTrader_Jesse]

Hello vpatanka,

That would not be the correct way for a strategy to observe its PnL data, strategies are virtualized so using the account values directly would not be suggested. You can see the following sample which works in backtesting to reset daily and has a max profit/loss.

https://ninjatrader.com/support/helpGuides/nt8/using_trade_performance_statis.htm

Using CloseStrategy also would not be suggested in a backtest/playback, this terminates the strategy and disabled it from working further. If you wanted to test multiple days you can use the same type of logic shown in the above sample. That simply prevents the strategy from doing anything logically but keeps it running so it can reset for other days.

[vpatanka]

Hello vpatanka,

That would not be the correct way for a strategy to observe its PnL data, strategies are virtualized so using the account values directly would not be suggested. You can see the following sample which works in backtesting to reset daily and has a max profit/loss.

https://ninjatrader.com/support/helpGuides/nt8/using_trade_performance_statis.htm

Using CloseStrategy also would not be suggested in a backtest/playback, this terminates the strategy and disabled it from working further. If you wanted to test multiple days you can use the same type of logic shown in the above sample. That simply prevents the strategy from doing anything logically but keeps it running so it can reset for other days.

I tried it and that works. Thanks.
Is there a way to backtest a portfolio? I am running 10 different instruments with different timeframes and I wanted to see and say I have a daily profit target of 1k across the portfolio and if achieved it'll close all strategies and be done for the day. Is there a way do that at the portfolio level? Thanks in advance

[NinjaTrader_Jesse]

Hello vpatanka,

Yes that is called a basket test: https://ninjatrader.com/support/help...asket_test.htm

You won't be able to test a daily profit target in a basket test across all instruments. Strategies are run individually and don't communicate with each other. To do a daily target for all instruments would require using a multi instrument strategy. You can see an example of a multi insturment strategy which comes with the platform, its called SampleMultiInstrument.

You can read more on this topic here: https://ninjatrader.com/support/help...lightsub=multi

[vpatanka]

Hello vpatanka,

Yes that is called a basket test: https://ninjatrader.com/support/help...asket_test.htm

You won't be able to test a daily profit target in a basket test across all instruments. Strategies are run individually and don't communicate with each other. To do a daily target for all instruments would require using a multi instrument strategy. You can see an example of a multi insturment strategy which comes with the platform, its called SampleMultiInstrument.

You can read more on this topic here: https://ninjatrader.com/support/help...lightsub=multi

Thanks Jesse. I just looked at SampleMultiInstrument and looks like its the same strategy with same parameters but I can run it against multiple instruments with different timeframes which is good to know but in my case I have different strategies (each one has its own set of parameters) for different instruments in a basket. Is there a way to backtest that basket

e.g.
Strategy ATMLiveTradingCL10Min has variables ThresholdForEntry as 2 and EmaGap as 0.75
where as
strategy ATMLiveTradingES4Min has variables thresholdForEntry as 13 with the same EmaGap

Would it be possible to backtest these two in the strategy analyzer? I have 8 others on 8 different instruments but once I know how to backtest it for 2 I can do the same for others included in the basket. Thanks

[NinjaTrader_Jesse]

Hello vpatanka,

No you can only backtest 1 strategy at a time so you would need to combine those multiple strategies into one which can handle that amount of instruments and parameters for its logic.

[vpatanka]

Hello vpatanka,

No you can only backtest 1 strategy at a time so you would need to combine those multiple strategies into one which can handle that amount of instruments and parameters for its logic.

Got it. Thanks. I will look into that.
One last question - I have some skills with programming a strategy which I have done currently but combining my 10 strategies on 10 different instruments (some of them use different timeframes) into one that I can backtest might be out my capabilities.
Does NT recommend any third party company(ies) that I can get in touch with who can help with this? I am ready to pay a reasonable fee for the project. Please let me know. Thanks

[NinjaTrader_Jesse]

Hello vpatanka,

Out support cannot make recommendations but you can reach out to [email protected] to get infomration on third party developers.