1 Topic by GD (edited by GD 2015-12-18 19:30:59)

  • GD
  • GD
  • Licensed Member
  • Offline
  • Registered: 2015-05-22
  • Posts: 444

Topic: Comparison of results between FSB, MT4 and MT5

For a long time I disturbed Mr. Popov about this question.

After some time I found that problems I had are because of some custom FSB and MQL indicators I used, which some time do not respond well.

Here is is clear answer for an not perfect EA run M15 GBPUSD but useful to check and compare behavior

http://s1.postimg.org/mlcha2swr/2015_12_19.jpg


I used data of Alpari Demo from MT5 to produce the EAs in MT4 and MT5.

Please correct FXM is FXCM demo data.

As you can see the general behavior of EA in all cases is almost the same (1000 dollars in 100000 is not so big deal). After all who can predict the behavior of metatrader live. I cannot understand why to believe their every tick algorithm to calculate parameters for my EA.

I prefer FSB Bridge in any case.

I would like to listen your comments relative to other timeframes.

2 Reply by GD (edited by GD 2015-12-18 19:41:19)

  • GD
  • GD
  • Licensed Member
  • Offline
  • Registered: 2015-05-22
  • Posts: 444

Re: Comparison of results between FSB, MT4 and MT5

The number of bars for FXCM DEMO in MT4 were less than 50,000 in previous graph.

It is not my fault or fault of FSB. Server of MT4 FXCM DEMO did not give me so much data...

3 Reply by GD (edited by GD 2015-12-19 06:16:11)

  • GD
  • GD
  • Licensed Member
  • Offline
  • Registered: 2015-05-22
  • Posts: 444

Re: Comparison of results between FSB, MT4 and MT5

Another important information is

Be careful what parameters you use in EAs of different servers

i.e. Number of bars and number Lots relative to account

Learn to read Journal results in Strategy tester if there are problems of starting (oninit fail)

Sometimes it is a common reason EAs do not start.

MT5 Strategy Tester is more useful than MT4. So, an FSB-Bridge for MT5 can be very useful in the future.

By including in last change of FSB (MQL version 33), of parameters automatically, optimization can be very useful
and also use of EA Statistics in MT5.

I recommend at first use of FSB bridge during development

I hope Mr. Popov will give a precise guide how to write an indicator in FSB.

It is something important for the development and spread of his product.

Initially Study shows me that even M1 results statistically give acceptable results as in first post.

The use of Multimarkets is important and try to use same  number of data for all Data Servers you use to compare.

This way there is a meaning to compare results and calculate parameters from each of them from same EA

  • footon
  • footon
  • ◄≡≡≡►
  • Offline
  • Registered: 2009-08-17
  • Posts: 2,969

Re: Comparison of results between FSB, MT4 and MT5

Pardon my question, but are you trying to convince to drop FSB's backtest and use MT backtest only, especially MT5? I'm getting an understanding you want to use craptester with FSB-MT bridge? Bridge will not work in that tester.

5 Reply by GD (edited by GD 2015-12-19 14:08:06)

  • GD
  • GD
  • Licensed Member
  • Offline
  • Registered: 2015-05-22
  • Posts: 444

Re: Comparison of results between FSB, MT4 and MT5

Dear Footon

1> I just said WE NEED an FSB BRIDGE for MT5 also
2> I verified that results of FSB and MT4 and MT5 are very close to each other.

Every USER does not care only if FSB Pro is fast but also about compatibility of results and under which point of view.

MY RESULTS PROVE that there is good coincidence. This is IMPORTANT.

I like to use FSB Bridge and I Want a good MANUAL by Miroslav Popov to write by myself Indicators in CSS.

About my opinion on MT4 Tester I cannot be kind enough to explain how bad it is.
I keep this opinion for a long time and before I heard about FSB.
MT5 is better than MT4 but why to use a tool and learn MQL if you can have all in one?

3> Trade is not only indicators. It is creation of functional space using as basis existed functions (indicators, metrics).
For this reason many EAs need also. This is given with help of Generator of FSB
Why you need that? Because you need to increase your possibilities to win with different approaches at the same areas of trade behavior.

In this case FSB pro is in advantage.

  • footon
  • footon
  • ◄≡≡≡►
  • Offline
  • Registered: 2009-08-17
  • Posts: 2,969

Re: Comparison of results between FSB, MT4 and MT5

I see, now I get your point! I agree with you. What I want to add in your case is the similarity between real trades and FSB backtest, from the start of my journey in this forum I've seen occasions were trades do not match. Sometimes the errors have been discovered, but not always. When you say you have had trouble with custom indis, specify those in more detail, I think it is quite important to have them fixed.

Regarding the manual - some kind of introduction is already available, the main thing is to know C#.

  • Popov
  • Popov
  • Lead Developer
  • Offline
  • From: Bulgaria
  • Registered: 2006-11-30
  • Posts: 8,024

Re: Comparison of results between FSB, MT4 and MT5

I want to add some details on the topic Bridge vs EA.

The exported Expert Advisor contains the code of the Bridge EA, Bridge DLL and the Auto Trader in FSB. The code is almost 100% equal. There should not be difference between the Bridge and the exported EAs.
The exported EAs are more robust than the Bridge.
If there is difference, it must be a result of a bug and can be fixed completely. There is no limitations of the exported EA comparing with the Bridge.

Popov's Website

8 Reply by GD (edited by GD 2015-12-19 15:50:43)

  • GD
  • GD
  • Licensed Member
  • Offline
  • Registered: 2015-05-22
  • Posts: 444

Re: Comparison of results between FSB, MT4 and MT5

"Regarding the manual - some kind of introduction is already available, the main thing is to know C#."

Dear Footon, This is not enough.

We need a Bridge for MT5 and a NORMAL manual to create indicators in C#.
It is more convenient.

Also, difference in few trades IS NOT IMPORTANT. THE ALL IS IMPORTANT. Trading is Statistics (possibility, prediction).

The MAIN reason asking for an MT5 bridge is THAT the number of data IN MT5 servers is MUCH LARGER THAN MT4 servers

So the tests and comparisons can be much easier.


"The exported EAs are more robust than the Bridge. "

Maybe. I believe you but I prefer FSB bridge. There is no perfect algorithm to describe the inside bar behavior. This creates differences in EAs made by FSB and MT4, MT5.

  • Popov
  • Popov
  • Lead Developer
  • Offline
  • From: Bulgaria
  • Registered: 2006-11-30
  • Posts: 8,024

Re: Comparison of results between FSB, MT4 and MT5

The MAIN reason asking for an MT5 bridge is THAT the number of data IN MT5 servers is MUCH LARGER THAN MT4 servers

The bridge doesn't load full data history. It loads as much bars as needed to calculate the signals for the last 2 bars only.

There is no perfect algorithm to describe the inside bar behavior.

There is no bar interpolation neither in the Bridge nor in the EAs.

There is no backtesting at all. Both the Bridge and the EAs calculate only signals. They do not calculate trades result and don't need interpolation.

Popov's Website

  • Blaiserboy
  • Blaiserboy
  • Junior Part-Time Aspiring Space Cadet
  • Offline
  • From: Toronto
  • Registered: 2010-09-06
  • Posts: 2,849

Re: Comparison of results between FSB, MT4 and MT5

I am not so sure re MT5, I use it a lot and have had a lot of trouble with it, similar to Mt4.

I feel that Popov has researched MT4 and Mt5 to a great extent and designed the software to have greater accuracy than either of those.

I suspect the testers in MT4 and Mt5

Just my opinion

Blaiserboy's Website

11 Reply by GD (edited by GD 2015-12-19 20:07:58)

  • GD
  • GD
  • Licensed Member
  • Offline
  • Registered: 2015-05-22
  • Posts: 444

Re: Comparison of results between FSB, MT4 and MT5

Dear Mr. Popov

Ι just prefer to calculate with FSB the EAs and then run them with FSB Bridge.
It is more convenient and time efficient than to run EAs in MQL.
Also there are more indicators for FSB in css.

It is just my opinion.