MT5 Optimization Results default

New comment
 

Hello everyone

During genetic optimization (Every tick based on real ticks), some passes show extremely high profits (e.g. 2,000,000+), which is clearly unrealistic.

However, when I run a single backtest using the exact same parameters:

  • The result is normal

  • No huge profit

  • Sometimes even losing

  • Lot sizes and margin look correct

So the optimization result does not match the single backtest result.

Settings:

  • RiskPercentMode (1% of equity)

  • SL and TP enabled

  • No martingale or grid

  • Margin control implemented

Can you help me to delete this ? 

Thanks in advance for your help and have a good day

 
Michael Myers:


"when in doubt" i close mt4/5 and then i delete my mt4/5 cache and tester folder(s). This may include the Agent folders too.

This seems to fix all things mt4/5 related -- for me.

But note that this will mean that you have to do the optimisations "from scratch".
 
Michael Charles Schefe #:

"when in doubt" i close mt4/5 and then i delete my mt4/5 cache and tester folder(s). This may include the Agent folders too.

This seems to fix all things mt4/5 related -- for me.

But note that this will mean that you have to do the optimisations "from scratch".

Okay, thanks, it worked. 

I wonder if this is recognized as a problem in the MT5 Optimization system. 

Good day to you

MM

 
Michael Myers #:

Okay, thanks, it worked. 

I wonder if this is recognized as a problem in the MT5 Optimization system. 

The problem you described usually means that you have initilization problems in your EA and it can demonstrate sporadic effects. There is no evidence that there is such a bug in MT5 tester - single tester pass should give the same results as optimization.

There are nuances about visual testing and indicator usage (calculation on every tick or not), but you did not provide these details. The matter has been discussed on the forums many times and covered in the docs.

If you think there is a bug, please, provide a minimal test case to reproduce the problem.

 
Michael Myers #:


Stanislav Korotky #:

The problem you described usually means that you have initilization problems in your EA and it can demonstrate sporadic effects. There is no evidence that there is such a bug in MT5 tester - single tester pass should give the same results as optimization.

There are nuances about visual testing and indicator usage (calculation on every tick or not), but you did not provide these details. The matter has been discussed on the forums many times and covered in the docs.

If you think there is a bug, please, provide a minimal test case to reproduce the problem.

If issue does not happen again, then, the op should assume that his issue was some sort of data corruption or too many file fragments.

However Agreed. If the issue happens again and again, then, it is a coding problem.

 
Michael Charles Schefe #:
If issue does not happen again, then, the op should assume that his issue was some sort of data corruption or too many file fragments.

However Agreed. If the issue happens again and again, then, it is a coding problem.

Hello, 


Thanks for your time. the problem didn't happened again. I will take care of your precious informations.


Have a good day


MM

New comment