Hey guys, I'm making this idea around old XMT Scalper system, as the developer has been inactive since early 2017 I'd like to continue from the base they left.
I saw ASSAR EA based on the exact same code, they claim that it has some variations but I'm not sure about that. They've based Assar from the 2.4.X version of XMT, I do have that version and the last known (2.522).
The 2.4.3 is a fixed version, I have no idea how the original code looks like, but on backtesting on M15 with 90% modeling it gives pretty neat results, the issue with this version is that it sends a lot of Invalid TP/SL errors and not quite sure why (I started a demo test today, I'll update as it goes)On the 2.522 version theoretically its the original last known version of the EA but at least the version I have it has what I think its a bug that sends a ton of orders on every candle, I'm not quite sure why and that's what I want to fix. I think this issue might relate to one of the following:
The order send statement
The indicators variable used
Failure on checking open orders at the moment
I also include a fixed (by me) version of the 2.522 that had a minor code error when importing two mqh files. All the EA's has their own source code on the attachment, as I saw, the Asssar V9 is almost (I can't say exact just because I haven't made a line by line comparison) the same code as the XMT 2.4.3
If we get these EA to work it would bring big results.
All testings that I made have been on ICMarkets on True ECN account.
Note: I can't backtest the 2.522 on every tick because of the ammount of orders it sends, it takes like 15-20 on my laptop (Alienware R4 17 i7-7700) for each pass and even applying genetic algorithm it takes like 10K passes.
I post it on spanish and english forum:) if you mean that someone else has made this thread could you please post here the link?:)
Hai I am new in EA. I have friend that also familiar with EA making.
My I know from the beginning?
What is the method of this entry open position? Any indicators?
Thanks by the way.
I have a bad news. The XMT Scalper doesnt profitable on real accounts despite your backtesting results. I spent a lot of time with this project in 2014-15.
Unfortunatelly MT4 can't produce realible backtests. 90% is not enough. I have benn developing FX strategies since 2010. I have been developing under MT4, MT5 and cTrader platforms.
MT4 doesn't store real tick data (ask, bid and spread). Only the M1 OHLC data and Ask prices were stored and a simple math calculation produce calculated (false) ticks for backtests. See this article: https://www.mql5.com/en/articles/2612
The TickStory's 99,90% backtest is better but not enough too! On other hand MT4 is too slow. So you are wasting your time with MT4.
Metatrader5 runs 5-8x faster than MT4 and can produce 100% quality backtests with Every tick with Real tick mode!
So if you want more realible backtests you should forget MT4 forever. I recommend you: change from MT4 to MT5.
But the best backtest is the Forward test.
Could you explain why tickstory 99.9% backtest is still not enough ?
And why MT4 being too slow is a problem for reliable backtests ?
I wouldn't swear by the "based on real tick mode" ... I don't claim that's an unachieved feature, but I cleary let it ear.
EDIT : It was an old screenshot, here's the last release (1755) :
Obviously, still in dev'.
Can someone please post any article regarding how to change mql4 to mql5:)? I'll give it a try since I'm still learning.
As far as I know, the difference is that MQL5 is Object Oriented and MQL4 doesn't (Still not quite sure about OOP and "normal" one) so if someone could share a good course to learn MQL5 as well I would be grateful.
Regarding XMT its basically a either a channel trading based or channel breakout based, so I think it could worth trying to find some good parameters since the volatility filters are quite useful to filter noise.