Topic: In sample - what is it

Can you explain to me what that option and what it is used for?

Post's attachments

1.jpg 910 kb, file has never been downloaded. 

You don't have the permssions to download the attachments of this post.

Re: In sample - what is it

Out of sample testing (OOS) reduces the risk of curve fitting. For more information:

Re: In sample - what is it

OOS stands for Out of Sample.
The idea of OOS is to test and to see the performance of the strategy against unknown data. If the strategy performs well on this unknown (OOS) data, it means that the risk of over-optimization is lower.
If you set "In sample", FSB will use the full data set for backtesting. It means that no OOS is performed.
If you set 25%, FSB will EXCLUDE the newest 25% of the data bars and will calculate the test on the first 75% of the data. When the program finishes generating or optimizing, it will reload the full data set and will calculate again.

The green part of the chart is the OOS data zone. It is the most important part of the chart because the program was not used it for optimizing. It is also called future testing on historical data. It also simulates the "trading" part.

Google for "forex Out of sample" for more info.

Re: In sample - what is it

Very important to read this article to minimize over-optimization:

Re: In sample - what is it

Dear Popov
Can you please elucidate if this is also valid for Pro?

"OOS option does not influence on the “Limitations”. Even if OOS is in use, the Generator evaluates the limitations criteria for the full data series."

I believe this shouldn't happen. It's not statistically sound.


Re: In sample - what is it

FSB Pro performs "real" out of sample. It actually cuts the data set and generates the strategy on this subset. All Acceptance Criteria are real. Generator loads the full set only after finishing the generating process.
You can see this on the Input Parameters.

Re: In sample - what is it

That is the best to use for example 25% OOS - and after optimization graph on a green field must climb to the top. Yes?