However, your idea of running 3-day generator sessions and across varying timeframes is great. I like the idea of the 3 days because we know that any strategies across that collection will have the correlation filter applied. Also I agree the multi timeframe is a good plan to minimize correlation among those collections.

I am planning to run multiple generators nonstop. I'd like to get the criteria strict enough that I don't really have to do much work once it finds and validates great strategies. But the 3 day cycle of review could make good sense I think.

I usually run 3 sessions in parallel with a 3-day cycle.

To reduce the correlation risk you mentioned, I separate sessions by timeframe, for example:

Session #1 → EURUSD M15

Session #2 → EURUSD M30

Session #3 → EURUSD H1

This helps diversify both entry timing and trade frequency, limiting the chance that all sessions generate highly correlated strategies.

Which settings do you use in the correlation box?
I’ve never really tested those parameters in detail, so I’d be interested in your feedback.

Thanks,
Vincenzo

But in a workflow where multiple Reactors are running, or Express Generator, then I would have multiple strategies and/or EAs that haven't been processed through that correlation filter.

I do not see a way currently to invoke the correlation filter manually.

What process is recommended to get this filtering done? Should I merge all the collections into one? Will this apply the correlation filter?

Presumably the correlation filter means that two strategies are performing "nearly identical" so therefore it would be less diversified logic to run both at the same time.

Thank you