10 Cycles are not enough for a statistically relevant result, since the price curves and the performance is random and different on any run. I would use at least 100-200 cycles. You can set the seed it you want, but it is not required. MRC with training will overwrite the parameters, so train again afterwards.

- Version 3.11 is now officially released.

Hello,

I use a separate development folder for coding my strategies, mainly because the projects may become more complex later and I want to keep them under standalone version control. After compilation, a post-build script copies the generated .dll file into Zorro’s Strategy folder, here's how.

This workflow normally works as expected.
However, after I included the Eval Shell (<evars.h> and the other stuff to _optimize), built one script successfully and try to run,
I get the following error:



As I see Zorro’s built-in compiler needs access to the source .cpp file in order to extract the _optimize variables and compile the individual strategy variants. In my setup, however, the source file is not located in Zorro’s Strategy folder. It also does not have the expected strategy name, since the entry file is called main.cpp.

Is there a supported way to use the Evaluation Shell with externally compiled strategies and a separate source directory?
I would prefer to keep the Zorro Strategy folder clean while preserving my Git-based workflow, especially for potentially large, multi-file codebases.

This tutorial video shows how to export continuous futures data from Sierra Chart to T6, using a free ACSIL script that I put up on GitHub page.

You can view the tutorial here:
https://youtu.be/144texonxfU