Mechanisation of optimal parameter selection. Finding a common denominator. - page 7

 
Europa:
I have also heard that if two systems have FS for example 4 and 6, then working together (simultaneously) they will give FS=4+6=10. What do you think about this?
Gonewo.
 
OnGoing:
Mantra)
Would love to hear your alternatives. Just not on the level of philosophy, but practice
 
Avals:
I would love to hear your alternatives. Not on a philosophical level, but on a practical level
I plan to present the new concept in a separate thread shortly. However, the absence of alternatives does not mean that the chosen path is the right one.
 
OnGoing:
I plan to present the new concept in a separate thread shortly. However, a lack of alternatives does not mean that the path is the right one.
ok. we'll wait))
 
OnGoing:
I plan to present the new concept in a separate thread shortly. However, the absence of alternatives does not mean that the path is the right one.
Avals are easy to check:))
 
paukas:
Avals path is easy to check:))
Yes, I have. Desperate act) Only it's not someone's personal path and success stories that are discussed here.
 
OnGoing:
Yes, I have. Desperate act) Only it's not someone's personal journey and success stories that are discussed here.
Clearly, clichés and mantras are discussed ))
 
OnGoing:
Yes, I have. Desperate act) Only it's not someone's personal journey that's being discussed here.
No, there was no desperation - there was reinvestment)) But I agree that it's proof of personal character. Just any way you look at it, it still comes down to a history test. Even a real test is already a story)) And you'll encounter statistics anyway if you want a statistical advantage rather than a guessing game :)
 
paukas:

The system is a fit with historical data. If the fit is good, it will be profitable for some time. If it doesn't, it needs to be adjusted again. Trying to come up with a system that works on "all available history" is doomed.

Good fitting differs from bad fitting in that changing the adjusted parameter within a wide range leaves the system profitable.

For example: we adjust and optimize the MA. We obtained the optimum and determined a period of 100. If the system stays in the 50-200 range, it is a good fitting.

That is great. This can be considered a starting point for the problem we want to solve.

Now about the doom "on all available history":

-- If you know (and prompt us) how to select sets that meet the "Good Fit" requirement among the optimization results, then we only need to write a self-optimizing Expert Advisor and teach it to select these "good sets" and feed them into itself.

As a result, we'll get an Expert Advisor that will work on "all available history", and all available history for it will be OOS.

Get it?

----------------------------

Constructive:

let's post here the results of optimization of an EA on MA, included to MT4, (or any other, no problem)

and you try to select a "good set".

If it turns out that the selection methods are correct, let's say for 10 tries, then we'll code the case. The "auto-optimiser" is already there.

It won't take much of your time.

Are you ready, Vladimir? (Terms to be negotiated...)

Maybe someone else is ready to show a master class? Feel free to come along...

 
lasso:

Would anyone else be willing to show a workshop?

Well, let's try it together. Don't be surprised if you have to throw it away)))