Is backtest in MT4 strategy tester using variable spread from recorded tick data possible?

 

Hi traders/coders,

I want to ask whether is it possible that we could use the tick data recorded, with variable spread and volume in MT4 strategy tester?

If its possible :
1. can I use any tick recorder?
2. i need to convert the data saved in CSV to FXT file to be able to load in strategy tester, right?
3. could the tester read variable spread?

I have been thought of this for a long time, asked a programmer once and he said that it is possible, but now he is not into trading anymore.
This came back to me when I saw Birt Tick Data Suite ver 2 able to use random variable spread, but from few videos I watched, it can be made variable but its not real spread.

So, I would love to learn and read responses from this community, and I hope I will get the answer.

I dont mind paying extra bucks if there are things need to be coded etc2 before running a backtest like I wanted.

Thanks and Regards.

 
Ghadaffi Khalid:

Hi traders/coders,

I want to ask whether is it possible that we could use the tick data recorded, with variable spread and volume in MT4 strategy tester?

If its possible :
1. can I use any tick recorder?
2. i need to convert the data saved in CSV to FXT file to be able to load in strategy tester, right?
3. could the tester read variable spread?

I have been thought of this for a long time, asked a programmer once and he said that it is possible, but now he is not into trading anymore.
This came back to me when I saw Birt Tick Data Suite ver 2 able to use random variable spread, but from few videos I watched, it can be made variable but its not real spread.

So, I would love to learn and read responses from this community, and I hope I will get the answer.

I dont mind paying extra bucks if there are things need to be coded etc2 before running a backtest like I wanted.

Thanks and Regards.

You have it all with MT5.

Download bit + ask ticks from your broker and test with variable spread.

 
Bernhard Schweigert:

You have it all with MT5.

Download bit + ask ticks from your broker and test with variable spread.

Thanks for the reply man. Yeah, I am aware of that as well. I will definitely use MT5 soon to run my EA, early next year.

But at this current moment, I really need to know whether it is possible to do it with MT4.
I don't mind the hassle using some third party EA to record ticks, then use a converter to convert CSV to FXT.
I want to know, if I did that, will I able to test it with variable spread from the tick data that I recorded?

 
Ghadaffi Khalid: But at this current moment, I really need to know whether it is possible to do it with MT4.

I don't mind the hassle using some third party EA to record ticks, then use a converter to convert CSV to FXT.
I want to know, if I did that, will I able to test it with variable spread from the tick data that I recorded?

It is possible using 3rd party application and data - such as with Tick Data Suite and Dukascopy data (and yes it is real spread).

To do it with your own brokers data, requires you to collect it and convert it into a binary LZMA compressed format (not CSV) compatible with Tick Data Suite, that would require advanced C++ coding skills which I suspect you don't have.

 
Fernando Carreiro:

It is possible using 3rd party application and data - such as with Tick Data Suite and Dukascopy data (and yes it is real spread).

To do it with your own brokers data, requires you to collect it and convert it into a binary LZMA compressed format (not CSV) compatible with Tick Data Suite, that would require advanced C++ coding skills which I suspect you don't have.

Thanks a lot! If Tick Data Suite uses real Dukascopy spread, then it should be good already. 

As for the conversion part, I suppose I might need to hire some expert to do it then.

 

Ghadaffi Khalid: Thanks a lot! If Tick Data Suite uses real Dukascopy spread, then it should be good already. 

As for the conversion part, I suppose I might need to hire some expert to do it then. 

The skills required to code such an application for this will not come cheap but it will be a one time payment, because once you have such a conversion application on hand, you will be able to run it on your PC as many times as you want.

My suggestion however, is to just use the Duskascopy data and just pad the spread in order to simulate different broker scenarios (this is what I do). There is no need to use your own broker data, because your EA should be written to adjust to different broker and market situations anyway.

 
Fernando Carreiro:

The skills required to code such an application for this will not come cheap but it will be a one time payment, because once you have such a conversion application on hand, you will be able to run it on your PC as many times as you want.

My suggestion however, is to just use the Duskascopy data and just pad the spread in order to simulate different broker scenarios (this is what I do). There is no need to use your own broker data, because your EA should be written to adjust to different broker and market situations anyway.

Great! At least I know that it will be developed to be some kind of application which I can use anytime.

I think for now, I will buy Tick Data Suite, download Dukascopy data, and adjust the slippage/variable spread setting in the test.

Thanks a lot for your suggestion and your input. You have been helpful.  

 
Fernando Carreiro:

The skills required to code such an application for this will not come cheap but it will be a one time payment, because once you have such a conversion application on hand, you will be able to run it on your PC as many times as you want.

My suggestion however, is to just use the Duskascopy data and just pad the spread in order to simulate different broker scenarios (this is what I do). There is no need to use your own broker data, because your EA should be written to adjust to different broker and market situations anyway.

You mean start from low spread run the test and then keep changing and doing the test again and again on duckascopy data? 
Is this the best we can get for backtesting?
Reason: