Unexplained parameters

AI & Machine learning
1

I bought the course Mean Reverting and



During che video "Mean Reversion Strategy" no one explain why we heve to set-up bb at mm5 and 0.5 STD



"

Here N is the lookback period to compute
mean and standard deviation and X is the
entry point which are free parameters to
be optimized using a training data set.
For mean reversion strategy on AUDCAD
in next unit, we will take N as 5
and X as 0.5
"



I would like to know according to which criteria were chosen


 
2

To find the best strategy parameter, you need to divide the dataset into two parts namely train and test dataset. Then, find the strategy parameters (using trial and error method) which give you the best results in train dataset and validate those strategy parameters in the test dataset. 



Further discussion on this topic can be found at below link: http://epchan.blogspot.com/2017/11/optimizing-trading-strategies-without.html



Thanks!

3

I do not understand why you give an out of context answer I know very well the difference between IS and OOS (learning / test). I asked why I have to use 5 and 0.5 … I would like to see your backtest and the reason why you chose those two parameters.

4

As the focus in the course was to illustrate a simple bollinger band mean reversion strategy, these two strategy parameters (5 and 0.5) were arbitrarily chosen. However, it is advisable not to use these parameters values as it is. In the previous answer, I was trying to highlight a scientific method to find the best strategy parameters and to cross validate it. I hope this helps. Thank you!