Portfolio backtest tool for Empty4

[ezy]

My daytime job is software developing (C#).
I came up with the idea of creating tool for backtesting the same bot on multiple currencies.
My original idea is based on Nanningbob's strategies. I would like to develop a bot, and backtest it on all the 24 (or whatever) currencies with just a few clicks.

The first concept is the following:
I create a small utility tool where you can set up the backtest parameters just like you can in Empty4.
You can then check what currencies would you like to include on backtest.
Empty4 can start backtest using commandline parameters, so after gathering the required parameters, my tool simply would fire up an Empty4 for each currency pair. The results are saved in html reports.
For analyzing these results I would use some existing tool like strategyquant ea analyzer which can aggregate all the results.

So that's my idea. I hate to reinvent the wheel and also hate to do things that are known to be useless.
I would like to hear you suggestions, cautions.
Maybe I should drop the idea without thinking because it's flawed or because there are existing tools for this. Maybe you can share some experience about do's and dont's. Maybe you have some great feature suggestions.

Thanks in advance for inputs.

[phil_trade]

Hi Ezy

did you read this ?

http://www.stevehopwoodforex.com/phpBB3 ... 249#p82249

also, if you're a C# ingeneer you should have a look at Forextester http://www.forextester.com/

cheers

Philippe

[ezy]

Phil,
Darwin's WFA only uses one currency - at least after a quick look. The point here is to easily test on multiple pairs. So what's your suggestion? Should I extend Darwins code for portfolio backtests?

Forextester... It's a commercial product. What should I do with that?

One thing I want to mention is that if the goal is to run a bot in Empty4, then I reeaally would not like to develop in other platforms while doing backtests. (Porting is not fun and also easy to do mistakes.)
So do you think that I should not use Empty4 for running bots?

[phil_trade]

Phil,
Darwin's WFA only uses one currency - at least after a quick look. The point here is to easily test on multiple pairs. So what's your suggestion? Should I extend Darwins code for portfolio backtests?

Forextester... It's a commercial product. What should I do with that?

One thing I want to mention is that if the goal is to run a bot in Empty4, then I reeaally would not like to develop in other platforms while doing backtests. (Porting is not fun and also easy to do mistakes.)
So do you think that I should not use Empty4 for running bots?

Forextester... It's a commercial product. What should I do with that?
Develop with a professional powerful language, in multi currency, with debugger etc

So do you think that I should not use Empty4 for running bots -> Yes

Porting is not fun and also easy to do mistakes.) Yes and No...You're a professional developer

Porting is difficult but working with
- correct M1 DATA
- with a debugger so important for complex EA
- with powerful language (Delphi or C#)
is more important IMO

cheers

Philippe

[ezy]

Phil, I really appreciate your advice.
The way I was playing with bots till now was the following:
I used Empty4, write a few lines of code in MQL to import my dll developed in C#. The dll was just a chunk to communicate with the real bot logic service, a seperate executable, so I can debug my code in visual studio and use fancy shit like GPU, machine learning and databases...

Phil, I'm not sure what you are saying actually. Do you suggest that I should completely forget Empty4, maybe delete it? Or do you suggest that I should develop in my own environment and then port the code and run on demo/live accounts with Empty4? What's your practice?

Thanks Phil.

[SpiderX]

Hi ezy,

That looks like a ton of work, probably easier to extend Darwin's WFA.
Good luck on your venture.

Cheers