I think that only forward testing on unknown data is viable.
Put your strategies aside for 3 weeks and re-test them on newer data.
You can run them on a demo account for that period or on a real account but with the minimal entry amount possible.

Will let the demo run continue for now as well.

I think that only forward testing on unknown data is viable.
Put your strategies aside for 3 weeks and re-test them on newer data.
You can run them on a demo account for that period or on a real account but with the minimal entry amount possible.

Thanks for the advice Popov.

I have started using Express Generator pretty much daily following this cycle

1. Generate strategies for few pairs, and others collecting 2 years of data. Using Premium data with my main broker Spread, swap short and swap long with Monte Carlo enabled and others criterias.
2. I then simulate demo to live trading. One month demo and one month live after for the top 10 simulated strategies;
3. From the Real simulated data I load it to EA Advisor Validator with no Monte Carlo and Acceptance Criteria;
4. I go to EA Advisor Strategies and use extra filters mainly profit factor which I've also used in Express Generator. Is using filters required there ? I did start sending those in a Portfolio and started demo trading them to see the effectiveness and so far the results if profiting. Am yet to move it to real trading account.
5,. From this point is where I have questions:

5.1. Is it recommended we use Walk Forward or the Normalizer on top after with say 20% Out of Sample as they are not features of Express generator ?
5.2 How robustness is the Express Generator over using the Reactor with Normalizer or Walk Forward? Can this lead to over optimization?

So where I am going with this is are these extra steps after point 4. a necessity ?

Any comments are welcome.