Errors, bugs, questions - page 2372
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
And it is also very suspicious that each successive approach of genetics gives better and better results. It's as if it has a memory (between individual optimizations).
Bruteforce: best result out of 15223 passes is 158.38 (castum criterion).
Genetic run #1 - 126.24, #2 - 126.24, #3 - 137.06, #4 - 137.06, #5 - 137.06, #6 - 137.06, #7 - 137.06, #8 - 137.06
Coincidence?
It makes sense if it runs its own algorithm, but does not recalculate the finished data, but simply grinds the numbers from the cache.
This makes sense if the complete enumeration has not been completed before. If it was, the GA must not be executed. Otherwise there will be a mishap - all the data is there, but each GA run produces different results.
Changed a parameter that does not affect the result (mejic), ran another series of genetic optimisations.
Results: 19.08, 105.07, 112.47, 112.47, 112.47, 112.57, 112.57, 132.87, 132.87
There is an error on the face of it.
This makes sense if a complete overshoot has not been completed before. If it was, then the GA should not be executed. Otherwise there will be a mishap - all the data is there, but each GA run produces different results.
There is no mishap, GA does not always have to find the maximum. It does not know and should not know that the whole cache is there.
There is no mishap, the GA does not always have to find the maximum. That the whole cache is there, he does not know and should not know.
The tester does.
The tester knows.
Well, that's not an argument. GAs are not supposed to break down.
Kindergarten.
Thanks for the constructive :)
Where is this information coming from? What you are doing - UB, the Expert Advisor can display an obscene word on the screen and will be right, any assumptions about the logic of this algorithm are groundless.
Thanks for the comment.
I agree, I was just worried about 100% repeatability and code result: in one class's instance it was incremented by 0.1, got 0.1. In an instance of another class it was incremented by 0.2 and got 0.3.
As I wrote myself, it was an oversight. But the result is interesting, not random rubbish, but exactly such. That's what I share with platform developers in the first place. Maybe it's 'zzzzzz' for a reason.
According to MQL4.
The RVI handbook has
MODE_MAIN
0
Main line
MODE_SIGNAL
1
Signal line
Signal line is "Signal style" which is red ?
According to MQL4.
The RVI handbook has
MODE_MAIN
0
Main line
MODE_SIGNAL
1
Signal line
Signal line is "Signal style" which is red ?