How to find good value when using WF optimization?

Viewing 14 posts - 16 through 29 (of 29 total)

← 1 2

Author

Posts
04/25/2020 at 9:35 PM #128090 quote

Moda
Participant

Senior

ok thanks @Francesco, i will double check

04/25/2020 at 9:37 PM #128091 quote

Moda
Participant

Senior

full output attached…

RB-1.png

04/26/2020 at 10:47 AM #128171 quote

Francesco
Participant

Veteran

In hundreds of rts that ive done, i’ve never seen a 20 5 and 21 5 having an higher score then 10 10.
That’s why i told you to check it better and told that it was unusual @nonetheless

Anyway, this could be an ecception, maybe synonymous of a good system 🙂

Moda, Paul and nonetheless thanked this post

04/26/2020 at 1:50 PM #128203 quote

nonetheless
Participant

Master

Francesco wrote:

never seen a 20 5 and 21 5 having an higher score then 10 10

thanks Francesco. I never use 20,5 so wasn’t aware of any comparison between that and 10,10.

05/02/2020 at 6:28 PM #129480 quote

Moda
Participant

Senior

Guys,

Need eyes here, is this too much unbalanced results ?

Cheers

DJ.png

05/02/2020 at 7:02 PM #129484 quote

nonetheless
Participant

Master

Well, that truly is an unusual result. Either you’ve written the perfect code or something has gone wrong in the VRT.

Have you got this Excel sheet? #107211

It automatically graphs the results so it’s easier to read.

Moda thanked this post

05/02/2020 at 7:07 PM #129486 quote

Moda
Participant

Senior

Yes using the last version made by @Paul

Here it is and the code is yours i’m working on MoD Dax/Dow

DJ-1.png

05/02/2020 at 7:14 PM #129489 quote

Francesco
Participant

Veteran

Having no results in the third one means both gain and % curve went negative, am i right?
Also, the 100% gain curve in the 10 10 is nearly impossible, especially when in 20 5 and 21 5 go negative.
There’s something inusual for sure 🙂

Moda thanked this post

05/02/2020 at 7:16 PM #129490 quote

nonetheless
Participant

Master

Did you sort the average gain from low to high? It looks very peculiar. Almost impossible to get a 100% score there.

If it’s my algo, then it’s definitely wrong! These are the results I got for DAX v2.

DAX-5m-Mother-of-Dragons-v2-VRT.jpg

05/02/2020 at 7:22 PM #129494 quote

Moda
Participant

Senior

Obviously, thanks for your time @Francesco

It looks suspicious… for the 21/5 i did not run it

Yes i sorted results … i found out sorry for the noise i made a mistake sorting data

05/02/2020 at 7:26 PM #129496 quote

Francesco
Participant

Veteran

If you want to share your code i can take a look for you

05/02/2020 at 7:32 PM #129497 quote

Moda
Participant

Senior

sure i post it in the right place where it belongs in the MoD post #129503

05/02/2020 at 8:14 PM #129506 quote
Francesco
Participant

Veteran
The problem was in the robustness tester code, replace it with this
```
//Robustness Tester
once j = 0
once flag = 1

if flag = 1 then
j = j + 1
if j > qty then
flag = -1
j = j - 1
endif
endif

if flag = -1 then
j = j - 1
if j = 0 then
j = j + random
flag = 1
endif
endif

tradeon = 0
if opendate >= 20000101 then
if barindex mod qty = 0 or barindex mod qty = j then
tradeon = 1
endif
endif
```
I just did the 10 10 and the results are “normal”, so also 20 5 and 21 5 will surely work 🙂
05/02/2020 at 8:17 PM #129507 quote

Moda
Participant

Senior

Many thanks @Francesco 😉
Author

Posts