Contract rollover offsets in NT8 are different than NT7

[Brillo]

I have been converting my NT7 indicators and strategies to NT8. I’ve found a data discrepancy between the two systems. This discrepancy is the root cause of why backtest results for the same strategy on the two systems does not match.

I traced the cause off the data discrepancy to the offset tables used to construct the back adjusted contiguous contract for the E-mini S&P 500 futures contract ( ES ).

Find attached word document with the specifics. For almost every single rollover data between 2009 and 2015, the offset is different. Also, the date for the rollover this coming Dec. 20015 is wrong.

[NinjaTrader_ChristopherJ]

Hello,

Thank you for your post.

As mentioned in my previous email I am currently looking into the Offsets further and will follow up as soon as I gather additional information.

[Brillo]

Any update Chris?

[NinjaTrader_ChristopherJ]

I currently do not have any additional updates at this time. I have forwarded the information to our development team and will follow up as soon as I get additional feedback.

Thank you for your patience.

[NinjaTrader_ChristopherJ]

After speaking with development futher the Offset difference is expected when comparing NinjaTrader 7 to NinjaTrader 8.

Within NinjaTrader 7 Offsets are calculated based off the RTH (Regular Trading Hours) Close and within NinjaTrader 8 Offsets are calculated based off the ETH (Electronic Trading Hours) Close.

[Brillo]

Hello Christopher,
Given the consequence that this decision makes I'm wondering if it was made without due diligence. Any strategy ported from NT7 to NT8 will not match trades. As a vendor it becomes a challenge to explain to customers the difference in back test performance. Given that customers and user will not immediately understand the cause of the difference, they may tend to resolve their confusion about it by substituting some theory that either the trading system or NT8 has some flaw.
Before NT8 is released I would ask that this decision be revisited at a higher level in your organization to make sure that it's consistent with what you want to do.

[NinjaTrader_Brett]

Brillo,

Thanks for bringing this up.

At a high level we've defined custom rollover dates for instruments depending on when historically most traders actually rollover. This has always been an art not a science and we in some cases can expect to have different rollover dates from other data providers. We've had large demand to switch our daily bars provided from our historical data service from RTH to ETH bars and as a consequence of that change the rollover calculation time had to change as well.

At face value we on purpose made the change to ETH bars, however I will review the impact you found below to insure that is consist with our design goals for NT8 and also review our documentation to see if there is an area it could improve to make this difference more clear.

Thank You!