Machine Learning: Training

Hi all,

I am building a machine learning algo and am following the step by step process described in "Better Strategies 5: A Short-Term Machine Learning System" from Financial-Hacker.com

My question is: At what point should I introduce parameters of Stop, Trail, TP, etc.?
Is it in step 5 "Generate a test data set" or in step 7 "The Strategy" after R calibration process? Or in both?

Thanks

2nd question: When R script finishes report on TestOOS(), how does Zorro find the right file to train afterwards?

Thanks

After the calibration. The steps are identical to the procedure described under "The development process". The train files are generated by training, not by TestOOS.