MT5 - when to switch to new version ?

[ianj]

The main block to getting something off the ground is access to current trading accounts.

• LMAX APi is one possibility - but its limited to LMAX and chargeable so its out
• Integral integration would be interesting - if only the entry cost was not so high
• FIX is a possibility - but NO ONE currently supports its at any sensible price or account size (and its only a protocol - you still need a supporting broker)
• cTrader looks ok for manual trading - but its backtesting capabilities are pants
• Empty4 bridging from a separate environment is clunky - but i can work - and has access to many Empty4 brokers - it could at least open up options - until MQ explicitly block it

Without traction with brokers you are nowhere - so what do you have to offer the broker - not traders . It is chicken and egg - brokers will not offer it without clients.

There are no open server side frameworks that have been widely adopted and the only open protocol to connect to those at the moment is FIX. Creating a new client platform is one thing - creating another server side platform is all together a different ball game - you need fast matching, liquidity integration, trade and client administration systems - the job is at least as difficult if not more difficult that a new client platform - just ask LMAX. Probably the only way to get it up and running would be to make it (at least partially) open and economical - at least to early adopters. Perhaps a new client platform would attract traders - enough to make brokers take a closer look . Perhaps a chat with a few of the current Empty4 bridge providers would provide some options here

On the question of EA programming language - there is no reason at all why an EA programming language cannot be completely agnostic - be it C/C#/Java/Python - as long as it had a common style/comparable feed/trading interface and decent backtesting/debugging capabilities - cTrader would be more popular now if it had decent backtesting.

On the question of platform implementation language - what matters is MODULARITY - if the system is modular and has a performant mechanism for communicating between modules then it can be implemented in multiple languages

Modularity can be coarse - the charting package can be separate from the manual order management, which is separate from the core server - Imagine the core server running without an interface - lean & mean. Imagine multiple charting front ends - on native windows, on linux, in Java and/or C#
Imagine the core server running strategies in C#, Python, Java - at the same time (BTW i can testify to the fact that Python is being adopted in institutional environments - but not on its own - often as a rapid development system and then being fully implemented in a faster language - be it C/C++/C#/Java)

If i were to start right here, right now - i would look at the following components:

• A competent charting package
• A competent backtesting platform
• A flexible, extensible cross platform API for EA development of multi pair strategies
• A performant server side matching engine
• Broker/Liquidity provider back end integration (Empty4 bridge providers would seem a good start)

the list goes on - but without the above you are nowhere - extensiblility/modularity is the key

[AnotherBrian]

I like the idea of cTrader - it looks good and I like the programming language, not that I'm familiar with it but it should be more powerful than MQL.

Here is an email I sent to Amibroker - FYI

Hi Amibroker guys;

I own version 5.4.

I wanted to let you know that there is considerable pushback from Empty4 users now – Metaquotes have done some stupid things, one of which is forced upgrades resulting in some not being able to access there account.
You probably have heard of the rumblings since its your line of business, if not, here is a start....

http://www.stevehopwoodforex.com/phpBB3 ... f=6&t=2267

http://www.stevehopwoodforex.com/phpBB3 ... 1&start=10

What this email is about is just to give you some thoughts....

There could be a business opportunity for you, adjust Amibroker for the Empty4 users who want to leave to another platform.. Some are considering cForex or jForex but those platforms aren’t quite right.

The reason I don’t use Ami any longer to trade is the trading interface. I want data and a trading interface that is not clunky, it needs to be built in. I want to trade, not program interfaces.
I traded with IB but found to much of the actual trade interface to be a pain in the ass. I wrote TipsterTrendlines, I know you have heard of it.

You should give some consideration to connecting traders to brokers, the FXCM’s of the world. I don’t like the order2go plug in, or importing/exporting data. I like to turn on the software, the data comes in, the connection is made with the broker using programming from an expert (like you guys) and I do the trading. You could write a back end or use another platforms, maybe jForex, its open source, but I don't know anything about it (just suggestions)

If you want to get more of a feel and some good feed back from some traders I suggest you sign up at SteveHopwoodForex. Friendly bunch of people who were fed up with ForexFactory so they started their own forum. I hope to see you there.

Brian

Hello,

Thank you very much for your e-mail.

What I am planning is to release source codes for existing IBController with a free license to
create derivative works and some docs regarding
how to connect 3rd party solutions to GetTradingInterface function of AmiBroker.
This should provide enough ground for 3rd parties to add their own trading interfaces.

At the moment I can not promise that I will develop anything especially for Empty4 users since
our resources are limited, and trading interfaces require lots of maintenance as brokers change their APIs.
But I think that opening sources for IBController may be good first step.

Best regards,
Tomasz Janeczko
amibroker.com


Edit: Update - the IBcontroller is open now.
http://www.amibroker.com/devlog/2013/06 ... -released/
This is the software that links amibroker to other trading platforms. The last time I used it, it link the trading interface from Amibroker to TWS (Interactive Brokers desktop application).

So 2 things are important for charting apps like Amirboker, one is to get the data to the application, the other is to get the trades to the broker. There is no use interfacing with Empty4 though except that Ami has better language and backtester.

Ami also released the development kit - so you can code in C,and other languages. That makes your code portable I think?
http://www.amibroker.com/devlog/2006/12 ... -everyone/
I assume this means if you write an indi using C for cTrader you can port it to C Amiborker no problem? (Not sure of teh exact languages, I don't really understand the diff between C, C#, C+, C++)

I still like the cTrader, but i haven't got into it far enough to find the gotchas.

[ianj]

I just looked at the spotware.com Open Api (the developers of cTrader/cAlgo) - it looks to have potential - they also look to be inviting alternative front ends as well as addons - and with at 2 decent brokers supporting it already/soon (IC Markets/Pepper) it might be a good starter option.

I have emailed them to see if they will ship the open API details - it will be a good test for how "open" to innovation they are

[ianj]

If Steve will agree, I suggest to open separate thread for this subject.
It would be great to put all Empty4/MT5 substitutes technologies with all features on one table. Then it will be easy to compare.

Who knows maybe this discussion born some new Open Source Project. - I hope

Remember - NeoTrader started something related to this - it seems to have stalled but keep it in mind

http://www.stevehopwoodforex.com/phpBB3 ... m.php?f=53

[scanner]

Heartily, what I read this month in the Empty4's user World is a storm for me.

Everywhere, every one shouted on that.. and myself, almost criing..

I work on an Ea since now one year, 8 to 10 hours each day, include week-end most of the time..

I'd nearly finish it, rest nearly 20 to 30 functions to be implemented.

it is Beautifull, Great like a Cathedrale, actually 7000 lines, 18 includes, using Global variable, all following with file csv on disk, Multi-paire, multiSymbol, multi trade, multiEntry Setup, multiExit Setup, Multi Modify "Trade running" Setup, with the capacity to make arbitrament ( choice) between selected setup, even with a setup that is good and a trade running ( if the setup seems better than the trade, and the risk capacitie is over, it gets the capacity to close the low trade getting free marging ,and openning the new one winner of this setup).
It contains all the checklist necessary to be safe before starting and monitor everything for you, like the monitor checklist on a plane ( account, computer, broker, connection), even there is no more tick, with the help Empty4 Ticker, alarm you that broker's connection is out..), blocking ketboard when you are not present if your Ea is not on a VPS etc...
And it is done to allow the backtesting, if not with multi paire together, all the Ea and its setups, with pair one by one . that permiss to adjust parameters for each Symbol..., then after make running them in live all together with individual crossing parameters by Symbol,( indic, Ut, Set-up, risk)

So, I'm criing... one year working on it...learning MQ4 language, need to only working with Array 4D, File, GlobaleVariable, everything that was possible with it, for nothing...

What I can say about that is : never again !

So, how it is possible ?

Regardless the language ,
I think that the only way is to build something independant from any console, that will be only able to send orders to a console. ( easy to rebuild in case)


of course the problem is double :

1) DATA Source:
- does exit a data source that would be avalaible for years ? or do we have to prepar a multiple possible source of data field to be collected before treatment?

2) Interface's order sent

- does existed a common denominator between all the console ? ( Csv file, text , Dll..)


You know, leaving in Tahiti, I'm going to bath myself all this week.. not going to the sea during this year and my tears will not be visible in the sea.

Another example of site where peope shouted with some Options ideas:

http://mechanicalforex.com/2013/05/an-o ... uotes.html

and a ling for an open source Plate Forme ( I'm only specialised on Empty4 language so for who are able to judge it) :
http://tradable.com/

[ianj]

http://tradable.com/

Trouble with tradable is they do not seem to want to talk to anyone - well at least to me - despite several attempts - is it real ?

[jaguar1637]

Well
Check this website https://www.quantopian.com/

[scanner]

Just answer " to tradable" ,2 mn ago and ask to get an access, will see... but you're right, looking at the staff.. my feeling is not so good, seems to be same kind of Empty4 Guys...

[idempotent]

Well
Check this website https://www.quantopian.com/

Came across this the other day. It's heavily linked to Pandas (the Pandas author now works there) and Python.

AFAICT, it's geared more for stocks, and only for backtesting. Which is not to discount it - some big brains there.

[aSushiRoll]

Heartily, what I read this month in the Empty4's user World is a storm for me.

Everywhere, every one shouted on that.. and myself, almost criing..

I work on an Ea since now one year, 8 to 10 hours each day, include week-end most of the time..

I'd nearly finish it, rest nearly 20 to 30 functions to be implemented.

it is Beautifull, Great like a Cathedrale, actually 7000 lines, 18 includes, using Global variable, all following with file csv on disk, Multi-paire, multiSymbol, multi trade, multiEntry Setup, multiExit Setup, Multi Modify "Trade running" Setup, with the capacity to make arbitrament ( choice) between selected setup, even with a setup that is good and a trade running ( if the setup seems better than the trade, and the risk capacitie is over, it gets the capacity to close the low trade getting free marging ,and openning the new one winner of this setup).
It contains all the checklist necessary to be safe before starting and monitor everything for you, like the monitor checklist on a plane ( account, computer, broker, connection), even there is no more tick, with the help Empty4 Ticker, alarm you that broker's connection is out..), blocking ketboard when you are not present if your Ea is not on a VPS etc...
And it is done to allow the backtesting, if not with multi paire together, all the Ea and its setups, with pair one by one . that permiss to adjust parameters for each Symbol..., then after make running them in live all together with individual crossing parameters by Symbol,( indic, Ut, Set-up, risk)

So, I'm criing... one year working on it...learning MQ4 language, need to only working with Array 4D, File, GlobaleVariable, everything that was possible with it, for nothing...

What I can say about that is : never again !

So, how it is possible ?

Regardless the language ,
I think that the only way is to build something independant from any console, that will be only able to send orders to a console. ( easy to rebuild in case)


of course the problem is double :

1) DATA Source:
- does exit a data source that would be avalaible for years ? or do we have to prepar a multiple possible source of data field to be collected before treatment?

2) Interface's order sent

- does existed a common denominator between all the console ? ( Csv file, text , Dll..)


You know, leaving in Tahiti, I'm going to bath myself all this week.. not going to the sea during this year and my tears will not be visible in the sea.

Another example of site where peope shouted with some Options ideas:

http://mechanicalforex.com/2013/05/an-o ... uotes.html

and a ling for an open source Plate Forme ( I'm only specialised on Empty4 language so for who are able to judge it) :
http://tradable.com/

Well one consolidation is once you have all the logic figured out yes it's a lot of work to reprogram it into another language but it'll take considerable less time..

In fact if it's that good I'll help you with it.. if you let me examine it and decide to use it