back testing wtih the same strategy and different results

[memonic]

Dear all.
Attached you can find two plots of two backtesting .
The strategy is the same for the two backtesting .
The parameters are the same for the two backtesting .
First backtesting is FDXM 06-20 from 01/01/2019 up to 12/12/2019
Second backtesting is FDXM 12-19 from 01/01/2019 up to 12/12/2019
In all cases the contract is merged back adjusted.
The results are highly differents , the number of trades and the results.
The strategy uses tick replay.

[NinjaTrader_Kate]

Hello mnemonic,

Thank you for your note.

This would be expected as you're testing two different contract expiries.

With MergeBackAdjusted (default setting) the chart will be continuous, stitched together by expiry month from the rollover date of the expiry month to the rollover date of the previous expiry month, and adjusted by an offset to prevent gaps between contract months (which trade at different prices) for a smooth chart. When you're using the 12-19 contract, that's going to have the original non-offset adjusted prices for that contract, while when you use the 06-20 contract it will be using the adjusted prices.

As far as why the two data series are different, these two publicly available wikipedia links will go into more detail.

https://en.wikipedia.org/wiki/Delivery_month

https://en.wikipedia.org/wiki/Mark-to-market

Please let us know if we may be of further assistance to you.

[memonic]

Hi Kate.

Using Merged non back adjusted the results are dissimilar too.

[NinjaTrader_Kate]

Hello memonic,

Thank you for your reply.

After making changes to the merge policy, you have to right click a chart of the instrument and select "Reload all historical data" - have you done so before testing?

Thanks in advance; I look forward to assisting you further.