Well, if that forums is “Rate my Software”, I must be in the right place.
Hello everyone!
I just launched my tool for creating an expert advisors (also known as EA generator, creator, builder, maker…) and wanted to share with you!
It’s something that is free by now, so I suppose that I’m not making unwanted advertisement or spam here.
And it’s free because is yet untested and I dont want to sell something that is not confirmed to be great and really working.
If someone is interested in such a tool…
… well, I need to first make 5 posts in forum, because of spammers control, so I’ll see you later. But until then you can tell me your thoughts and feelings about those tools for automatic creation of tools for automatic trading (I mean EA’s) :5:
You are asking for a world of trouble. There are already commercial products out there and of course MT5 has one built in.
The problem with the existing ones is that they only handle the most common and trivial of cases. I must have coded a dozen different Trailing Stop schemes and probably half of them were standard. Things like 3 level TS where the trailing amount decreases as the profit increases or a time dependent TS. Those kind of things can only be hand coded and the people using an EA generator aren’t capable of that.
Plus I have seen things like compile errors and unhandled errors coming out of these products. So be warned the bar is high. Best of luck on your endeavours. My constructive suggestion - post it on FF as their seems to be much more EA demand on that forum than here at BP.
Ahhhhh… world of trouble, that sounds not so good.
But I agree, because such tools somehow eliminates the control of the code, and what they can do for you - they will do it, but that’s all. There are invisible limits if someone wants to do things that are not included in tool’s possibilities. And as you said that possibilities are often very limited.
But interestingly enough the reason why I decided to code tool like this was quite the same - a world of trouble coding with MQL4 and especially MQL5. I just get tired of managing text information, copying, moving, removing, adding new functions… Trading coding is somehow different from regular coding (of applications, websites…)
And after all I think that all the expert advisors around are still pretty simple, even though they’re created manually. And for me that’s because of invisible limits of hand-coding in trading - changes in code must happens very often, and the resulting expert just cannot be more complex and intelligent from some limit beyond.
Another factor you have to consider is that people always want to tinker or include the latest flash of brilliance in their EAs. Some seem to think that because a “normal” RSI based EA fails, an EA with RSI divergence will work. Others see that doesn’t work, but are sure that RSI plus CCI divergence will be the answer. And of course there are the grid / Martingale types with their dozens of variations. So what looks simple at the outset becomes very complex very quickly.
It sounds like the tool you are building will work well in your hands. Are you intending to use it as a start point to build customized EAs for others?
That’s for sure, at least I know every aspect of the tool and I can control it. I really want to create something that will be easier to use and actually breaks those invisible limits that can exists.
That’s why I’m trying to include a full control ot own codes in it. Something like you work on MQL, but easier to manage codes, libraries, functions. I called that Library Studio and because it needs more upgrading, I really need opinions from others.
just playing around so far I tried to put together an EA to test the strat from this thread. 301 Moved Permanently
basically buy at yesterdays high and sell at yesterdays low. the stop is moved each day.
I was able to get it set up I think (I can code and I have used some very similar EA builders) I exported the MQL when I compiled it I did get an error. Thats no big deal for me. what I am having a hard time with is figuring out what block is creating the code thats giving me the error.
nevermind I didnt scroll up far enough in the code I see that the blocks are commented out.
First of all, I like to see someone really testing it You are the first that connects some blocks.
Because of complexity of the whole system, some bugs may appear… like this one. It will be corrected right now.
Edit:
Done. By the way the “If opened orders” after “No opened orders” does nothing for second block. But both groups can be used separate. There’s no need to have only one beginning block for the whose project. The only thing with blocks is that if some is disconnected (alone) - it will be inactive at all. But will stay there as a reserve.
I will get back to this later tonight. So far I like what I see. I do have a few questions. I cant see how to do a “IF ELSE” on open orders. IF open orders then do this IF no open orders do this? I need an IF Else on my long or short trade. A simple inverse on the logic will not work in this case.
I just edited post above… already on the previous page.
I dont really want to include programming terms directly visible, so nowhere you will see special keywords like “IF”, "OR’, “XOR”…
For your question I think you can find the answer above… Both “If opened” and “No opened” should be used separated. And within their parameters they can be filtered to be used for specific orders group, symbol or type. Is that what you want to acheive?
Yes, you can. All the blocks are represendted as a functions in MQL4. These functions are outside the main Start() function. In Start() only the beginning blocks are called. You may have many independent logic paths and many beginning blocks for each logic path.
On each tick all the blocks that are connected only from their outputs will be executed first, one after another… depending on their ID numbers, but I think that this is not so important.
This was the case with Start(), but in Init() and DeInit() is just like the same story.
Maybe I need to create some mapping tool to visualize the basic structure of outputted code… I’ll do it. And changing their ID’s also… if needed.
I think the way you have it is fine its just different from another builder I have used. I think the way you are connecting the blocks is simpler for the user. I was just stuck in a habit from other software.
That’s great. Because I’m really too busy with making that generator tool, and creating of real expert advisors is not something that I do. I use only small testing pieces of code. And If you or someone else succeeded to create their expert advisors, no matter how complex they are - that will be great because that’s my goal.
I just took a random strat someone posted on BabyPips a very simple strat at that. I even simplified some of the order parameters even more. I am going to go back over it and see if I can get it to do exactly what the poster was describing. I am pretty sure it will not be hard with your software.
Well… as you can see this here is mostly similar to the Forex EA generator, that I personally appreciate the most. At least I think that this is the best builder from all that I am familiar with.
But I’m trying to create something that I believe is the most intuitive to work with, and will continue to make things like this. Because it has to become easier to use from pure code writing and more useful. So if some limits exists… it needs to be broken.
By the way, there are some drawbacks because of web oriented stuff. Exporting is not so easy and some ways of managing downoads are needed. I personally use Firefox with a plugin called “Automatic Save Folder” that automates exporting a little. I still need to push 2 more buttons, but is better than default downloading behaviour.
JDeel - I’ll also have a little play around. I just registered and the front end looks good, but it’s the back end that needs to be tested. Will feedback …
