Managing Multi Strategie on a Single EA - page 2

[Deleted]  
Comments that do not relate to this topic, have been moved to "Discussion of article "How to Buy a Trading Robot on MQL5 Market?"".
 
meherbt37 #:

it's the idea it's not possible to test such EA and leave no choice to test live; and i searched and there is no platform

that can backtest strategie with market correlation for exemple or EA with multiple pairs entries.

if there is a tool or a platform that can do that it will be very usefull.

search the site. dont you know how to search? There is a feature of this website where, at the top of this site, the desktop site specificly, where you type out words like "multi symbol strategy" and you will see at least 3 articles on the first page of search results. And "to top it off", they have an ea already typed out and coded for you to add your own strategies into each section of the code -- you will find these towards the bottom of each article marked as "MQL5.zip", OR, "Download Zip".

And as already mentioned by the moderators, you can backtest them in mt5 and they will open and close trades on multiple symbols "at the same time". You will have to search the site again for instructions for that. We wont do everything for you.

 
Yu Pang Chan #:
Your idea is good. But EA can only backtest on a single symbol at a time.
So technically you might need to use other platform or tools.
see comment above.
 
I've gone quite far in this direction and personally I don't think it's overkill at all.

I actually prefer having multiple independent strategies inside one larger system, as long as each strategy remains logically separated.

Each strategy can have its own:

- entry and exit logic
- filters
- timeframe
- instruments
- trade management
- magic number
- risk allocation

Then another layer can manage portfolio-level exposure and drawdown.

The biggest mistake, in my opinion, is assuming that simply adding more strategies automatically creates diversification.

If several strategies respond to the same underlying market behaviour, you can have 10 strategies and still effectively have one large correlated bet.

That's why I think the interesting part isn't just building a multi-strategy EA. It's measuring how those strategies interact with each other.
 

It is possible to set a global variable and a delay/sleep so that a new position is opened only if the variable has not been set by another EA. link regarding global variables: https://www.mql5.com/en/articles/2744

 
The technical side is only part of the problem.

MT5 can handle multi-symbol and multi-strategy testing, but in my opinion the real challenge starts when you try to validate the portfolio as a whole.

Two strategies can look robust independently and still become fragile when combined because of:

• hidden correlation between entries
• simultaneous exposure during the same market regime
• concentration on one currency or underlying risk factor
• drawdown clustering
• one strategy dominating most of the portfolio profit

For me, the important question is not only “can MT5 backtest it?”, but whether the combined system remains robust when correlations and execution conditions change.

A multi-strategy EA can be very powerful, but only if the portfolio-level risk is treated as seriously as the individual strategies.

Has anyone here tested portfolio-level drawdown or correlation stress rather than evaluating each strategy separately?