10points 3.mq4 - page 295

New comment
 
I have not looked at the MQL for this EA, but MT4 does not remove spreads in the backtests.... at least not any version I've come across.

The only way I've been able to get accurate backtests is to store all the spread costs in a variable, and print them to a log. Then I subtract the spread from the profits generated.

And if you want to get REALLY accurate, you use another variable called balance and set:

balance = AccountBalance() - spread;

Then you use the balance variable to set your fixed lots instead of AccountBalance().

Rather complex, might require a reread or two.

I don't have the expertise to question what you are saying and I wouldn't be surprised if you were right. It just seems that if it were entirely true I wouldn't have gotten the results I did. Can you explain why the backtests showed -$4 for virtually every entry until the balance was depleted using SL of 0?

Thanks for any light you shed on this issue,

saintmo

 

hello

Santimo, describe all problem which you met and i try to find solutions

master001

 

master001,

I am not sure I can describe my specific issues so much as describe my status in testing. I can only tell you that I have most recently taken your new version of the program and tried it using the 4HR time frame using 2007 data. (Both for FXDD and IBFX which is where I would trade live). 2007 represents not only the most recent period but the most challenging period for this EA in my tests.

As a basis for comparison the original T-JMA version (for gbp/usd) produced a loss of $10,055 with a 23.21% draw-down ($50k account trading .1 lots). The new version with the default parameters produced a loss of $9,255 also with a 23% draw-down. A slight improvement. After reviewing charts and comparing the ATR indicator results manually I selected some different parameters which produced a loss of $10,400 with a draw-down of 24%. Results in wrong direction.

I have also made several optimization runs for the 4HR time frame and none of them have produced positive results for the 2007 period.

I really don't know what to do at this point. Is there anyone else who has tested and gotten positive results? IF so what settings, pairs, time frames are you using?

Thanks,

saintmo

 

Rmi

saintmo:
Thanks for the info, but since there are so many versions could you post the one you are using? Or reference the post number where it can be found.

Thanks,

saintmo

You can find it there :

https://www.mql5.com/en/forum/174975/page160

Thierry

 
thierrybl:
You can find it there :

https://www.mql5.com/en/forum/174975/page160

Thierry

Which TF uses?

 

Rmi

Kamick:
Which TF uses?

I'm running it on H1

 

hello

Saintmo i use neuimex for 2007.

go download install file from here:

Users Area -> NEUIMEX Trading Terminal :: NEXTT Trading Systems

login: asteroida5@wp.pl

pass: asteraster

i dont know why this server refuses me to upload file.

master001

 

master001,

I will do that when some of the memory-heavy tests I am running now are finished.

I didn't do that originally because I thought why load and test on a platform in which I am unlikely to trade. But I will take a look at what you have.

Thanks,

saintmo

 

Attached is an update covering this week for the original T-JMA version using FXDD for gbp/usd and eur/usd.

I thought it would recover a couple to times this week, but only to turn the other way. It was probably a mistake to include the gbp/usd.

Files:
t-jma-fxdd-2.gif  6 kb
 

Foward test after 2 weeks..

1...288289290291292293294295296297298299300301302...489
New comment