EA N7S_AO_772012 - page 25

 
WitoHOH писал(а) >>

Why not use a construction like:

You can. But universalisation leads to a significant increase in testing time.

MarketInfo RefreshRates etc. will be used in the real version of the EA.

 
gorby777 писал(а) >>
To SHOOTER777. Thank you for your nice EA. I have started testing it too. What about the bird in hand. There is an excellent EA of kimiv designed for closing all positions in my account when it reaches certain profit. Maybe we should not complicate this one?

To close, it will close, but to prohibit opening until the new optimal sets are set, it is necessary to organize a ban, for example, through a global variable and complicate the Expert Advisor again)))

Although, I would like to look at the "well-known kimiv great Expert Advisor to close all positions in an account at a certain profit". Where may I have a look at?

 
SHOOTER777 писал(а) >>

You can. But the universalisation causes the testing time to increase significantly.

MarketInfo RefreshRates, etc. will be used in the real version of the EA.

Wouldn't an EA for real trade have to be optimized?

For example, in my last version when combining both indicators one optimization phase takes approximately 8 hours.

 
SHOOTER777 >> :

Cumulative totals from the past seven weeks.

Seven weeks is something ... ;)

Maybe it makes sense to take the losing pairs out of the game or not enough stats yet?

 
WitoHOH писал(а) >>

Wouldn't an EA for real life have to be optimised?

For example, the last version of my EA will take around 8 hours to optimize when both indicators are combined.

It will be one EA and optimization will be necessary but functions necessary for real trading and those which are not during testing and optimization will be ignored.

if ( IsOptimization( ) ) TrBlnc = false;if ( IsTesting( ) ) TrBlnc = false;

Some of them can be done in other ways, but to minimize testing time I usually make two fast versions - for optimization and for trading. If you've noticed, all working versions have odd numbers, while I work through optimization and try different changes and additions in even numbers. But it is not so easy that the logic does not change yet.

 
SHOOTER777 писал(а) >>

There will be one Expert Advisor and it will be necessary to optimize it, but the functions necessary for real trading and those unnecessary during testing and optimization will be ignored.

if ( IsOptimization( ) ) TrBlnc = false;if ( IsTesting( ) ) TrBlnc = false;

Some of them can be done in other ways, but to minimize testing time I usually make two fast versions - for optimization and for trading. If you've noticed, all working versions have odd numbers, while I work through optimization and try different changes and additions in even numbers. But it is not so easy that the logic does not change yet.

That's good.

And, of course, if it is not difficult, may I post both versions here too?

Otherwise I just do not have time over the weekend to optimize.

Many thanks in advance!

I like this EA very much.

 
goldtrader писал(а) >>

Seven weeks is something ... ;)

Maybe it makes sense to take unprofitable pairs out of the game or there are not enough statistics yet?

Yes, next week I start to look for other optimization periods for the Canadian, Franc and Swiss. I have a feeling that they are optimizing badly and their gap is the worst. I think to decrease first phase up to three or two weeks and maybe G12() to H1 for them.

 
to SHOOTER777. You can't do what? It's on his website www.kimiv.ru. He also runs a branch here. You just need to edit the WinUser22.mgh library a bit
 
Here, I enclose. And why ban it, let it refine with the old ones. Or, according to the logic, will it reopen in the same direction?
Files:
winuser32.mqh  18 kb
 
Yesterday there were problems with optimisation on micro in Alpari. I added zeros to the stops and the optimization went on. Now I'm trying to do the same thing on the demo - it's not working.