Is the advisor suitable for real life? - page 31

New comment
 
Yes, try it.
 
OnGoing:
I don't know, yeah, give it a try, though.

I don't know. Although it may be non-normalized, but the condition will still be fulfilled. You have to normalize the stops after the calculations, if there are any.
 
FOReignEXchange:

I don't know. Although it may be some kind of non-normalized, but the condition will still be met. You have to normalize the stops after calculations, if there are any.
Just when the difference between Ascom and the value being compared is not very big, the number of decimal places matters and the condition often fails.
 

For some reason I think we should do this. 

   if (OrderSelect(ticket_sell,SELECT_BY_TICKET)==true)
      {
      if (OrderType()==OP_SELLSTOP) 
         {
         if (Bid<(OrderOpenPrice()+4*Point)) 
            {

I get the impression that 2 conditions at once sometimes can't read. I'll try putting it in. It looks like some kind of bug in the language.

I will try to do the same with normalization.

Maybe someone has encountered such problems yet?

 

No, they both work. It's more likely that the values are not given in the comparison.

Try tracing the values you get at each step.

 
OnGoing:

No, they both work. It's more likely that the values are not given in the comparison.

Try tracing the values you get at each step.


I've already checked with the comments. I even put a comment before this condition. The signal goes to this line, but it does not go any further. And it rarely happens. 2-4 times in a day. All the rest of the time everything works properly. I checked all the values in the comments when the order was not deleted. It should have been deleted, but it didn't reach the delete function. The signal went before this condition and no further.
 
I'm so fucking stupid. I forgot to say the most important thing. Everything's working fine in the tester. That's the problem. If the order was not deleted in the tester, I would not bother. But it is not deleted in the real account and is deleted in the tester if the visualization is activated after a deal. That's why I think it looks like some kind of bug in the language. I have the impression that it cannot remove much data from the selected order. This is not the first time I have encountered this issue. To be more exact, I actually encounter it all the time. Everything is OK in the tester, but not in real life.
 
Now I've turned on the visualisation and I'm looking at it. There's a huge time margin for removal. Half a minute almost. In the tester it deleted and today on the demo account there it didn't even try to delete and opened a salmon trade :)
 

Output the values not in front of the condition, but exactly those expressions that are inside the condition. Control what is compared to what. Then the next time such a failure occurs, you will be able to catch the cause.

After all, we know that the condition is failing. So we have to find out why. To do this, all compared values must be constantly monitored.

 

I've already done all that. Now I'm back on visualisation for one more place where the order was not deleted. Every tick I looked at the comments on the condition. Everything is correct there and the tester deleted the order. And the time to delete it was 10-15 seconds.

I looked through the logs of the demo account during this time and there were no attempts to delete the order. The tester did, but the demo account did not. I tried to put comments before and after the conditions, at the end of code and at the beginning of code all at once. All the conditions are fulfilled, but I have not gone further than that. We have not even tried to delete the order! We have plenty of time to delete it, but it is not the matter of time since we have not even tried to do it. There were no jumps on 2-3 ticks either. Just the condition does not pass to the delete function and that's all.

Ok I'll try another option - set these conditions with each new line separately. I'll see what happens tomorrow. On the first problem I kind of guessed how to solve. Tomorrow I will see how the updated code will behave.

1...242526272829303132333435363738...52
New comment