Same strategy different back test results

[mlprice12]

I copied my custom strategy and back tested both the original and copy, exact same inputs and parameters. They resulted in different outcomes. Looking at the charted analysis I determined that the indicator that corresponds to the strategy is being drawn differently for the two strategies. Trading hours are both set the same "instrument hours". Every 2 hours the indicator draws a static average price from the previous 2 hour period. In strategy 1 the indicator starts drawing at 2300 and in strategy 2 the indicator starts at 2345.

If the strategy and indicator code is exactly the same why are they drawing at different times?

Please let me know what more information I can provide.

[mlprice12]

This can be closed...I figured it out.

[NinjaTrader_ChrisL]

Hello, thanks for your post.

Can you try the same thing on the SampleMACrossover strategy and see if the same difference happens? Make a copy of the SampleMACrossover and ensure to test it on the exact same data set with the exact same parameters. When I do this on my end I get matching strategy results. There must be a different parameter between your two script copies. It would be recommended then to double-check the parameters, test on a completely different data set and bar type (try a simple 1 minute chart on one day of data), or even rebuild the script from scratch to ensure each script is exactly the same.

Please let me know if I can assist any further.

[mlprice12]

Thanks Chris, It was an issue with EOD being checked.

However I found a secondary issue. When I run the strategy through back test the indicator starts drawing at 2400 but when I add the indicator to a chart it starts at 2300.

I have my time zone set to GMT -5 but live in GMT-6 could this be the root of the issue?

[NinjaTrader_ChrisL]

Hello, thanks for your reply.

The strategy's start and end time is determined by the session template that is selected in the "Trading Hours" setting in the strategy analyzer. That will define exactly what the start and end of the session are. Test in the SampleMACrossover strategy for the same behavior. If you do not see the same issue on that strategy, there is something mis matched between your two scripts. Are your "BarsRequiredToTrade" parameters the same? If you Print out the timestamp data from OnBarUpdate do you see the same amount of timestamps?

[mlprice12]

Thanks Chris,

Testing the Sample MA Crossover strategy doesn't help because the issue is with my custom indicator that draws a static avg. price every 2 hours. I figured out the root of the issue but don't know why it happened or how to fix it.

After the close of trading on November 27 2020 the indicator draw setting changed. On and before Nov. 27 the indicator was drawn at 2400, 2200, 2000... At the start of the next trading day (1800 Nov. 29) the indicator started to be drawn at 1845, 2045, 2245...

Did some sort of internal change occur within Ninjatrader between Nov.27 1200 to Nov.29 1800?

[mlprice12]

Further Note: If I change the start date of the strategy to October 29, 2020 or earlier the change doesn't occur. The change only occurs when the start date is between October 30 and November 27 2020 and the end date is after Nov. 27 2020

[NinjaTrader_ChrisL]

Hello, thanks for your reply.

On Kinetick I have hourly time stamps all throughout November:

https://clauber.tinytake.com/tt/NTExMzg0Ml8xNjA0OTI0NQ

I'm sorry but I can not reproduce what you are reporting. Try reducing the indicator down to see if there is a bug in there, If not, you can always delete your historical data and load new data:

See the guide here on editing historical data in the platform:

Removing Historical Data

• Once the data has been deleted close NinjaTrader. Open the Documents\NinjaTrader 8\db\cache folder. Select all files then right mouse click and select “delete.”
• Open NinjaTrader and connect to your data provider. Open a new chart.

Please let me know if I can assist any further.