1 (edited by dedok3000 2015-06-17 13:53:21)

Topic: Big difference between FSB backtest and MT backtest

Hi,

When I generate strategies I see very good results but when I export any strategy as an EA and do a backtest I see complete money losing... I'm using the same data from my MT.
See an example please, maybe I do something wrong.

Daniel.

Post's attachments

Backtests.zip 13.81 kb, 13 downloads since 2015-06-17 

You don't have the permssions to download the attachments of this post.

Re: Big difference between FSB backtest and MT backtest

Anyone? sad

Re: Big difference between FSB backtest and MT backtest

Don't you trust FSB's backtest? How many ambig. bars do you have? If you put it on a demo for couple of hours and then run that data in FSB, do you have huge discrepancies in trades?

Re: Big difference between FSB backtest and MT backtest

Popov will be interested to solve this....... he will need  to see detailed pics of the trade from the tester and MT4 both, he con not do much until you give him exact pictures.

Please do not expect him to run your strategy and attempt to find your errors

Document the problem completely, do not take shortcuts or you are wasting people's time.

Time after time Popov invests his time to discover errors made by the guy submitting the error.

5 (edited by GD 2015-06-18 06:59:46)

Re: Big difference between FSB backtest and MT backtest

I have done my own research on the problem.

Only if there is mistake in some indicator can be happened  a difference or if EA is very sensitive in very small changes of brokers data. Use Monte Carlo

So please check carefully.

I checked your strategy and it does not look to be in good "shape".

Re: Big difference between FSB backtest and MT backtest

If there is an error, it must be fixed immediately. Please update your FSB Pro data directly from a chart. Run the expert on that chart and compare the latest indicator values with the values in the FSB Pro's Indicator Chart. If there is differences in an indicator please report. Use Ctrl+S to save data from a chart.

The second check is to run the expert on MT's backtester on a predefined time period and to compare the deals with the FSB Pro chart. Please note that if you have used "Download data" function in MT, the backtesting data of MT will be different than the data in the MT chart.

Re: Big difference between FSB backtest and MT backtest

I just compared the values:

http://s28.postimg.org/erdqblm89/screenshot_983.jpg

It looks like the Round Number II Plus indicator shows different values in FSB and MT. I hope Ahmed will fix it. Higher Round number = 1.16667 doesn't look rounded at all.

8 (edited by dedok3000 2015-06-18 14:34:51)

Re: Big difference between FSB backtest and MT backtest

Who is Ahmed?

Same for almost all other strategies sad

Noticed one weird thing... in the collections these strategies look awesome:

http://s9.postimg.org/k466115l7/image.jpg

but when I switch to a strategy tab I see the 'real' picture:

http://s16.postimg.org/dd9yo2ymp/image.jpg

What should I do? sad

Re: Big difference between FSB backtest and MT backtest

how much data is OOS

Re: Big difference between FSB backtest and MT backtest

"Round Number II Plus" is a custom indicator you have probably downloaded from our repository. We are not the authors. This one is created by Ahmed - a member of our community. You can read who created the indicator in the page in the repository.

Re: Big difference between FSB backtest and MT backtest

do u know that u can click in the repository on 'open' to open the relevant strategy, that wioll probably give you better results

Shade the one u want, then click open and it will be as in the repository.

I think that is the error

Re: Big difference between FSB backtest and MT backtest

Noticed one weird thing... in the collections these strategies look awesome:
but when I switch to a strategy tab I see the 'real' picture:

Both charts are real, but represent different states.

The strategy collection shows stats and charts from the data on which the strategies were generated on. When you later open a strategy the program loads the latest data and recalculates the backtest.

If you have used OOS, the collection was generated on partial data.

Re: Big difference between FSB backtest and MT backtest

Popov wrote:

If you have used OOS, the collection was generated on partial data.

OOS = All data.
Also when I was generating these strategies (few days ago) I saw the same chart on the strategy page as the one from repository page. Now when I reopened FSB I see this bad chart sad

Re: Big difference between FSB backtest and MT backtest

Well "Ross Hook" is a custom indicator as well?

Re: Big difference between FSB backtest and MT backtest

Also when I was generating these strategies (few days ago) I saw the same chart on the strategy page as the one from repository page. Now when I reopened FSB I see this bad chart

You have probably changed something  - Data Horizon settings, data files, account settings...

Well "Ross Hook" is a custom indicator as well?

Yes, it is.

16 (edited by dedok3000 2015-06-18 15:30:07)

Re: Big difference between FSB backtest and MT backtest

Popov wrote:

You have probably changed something  - Data Horizon settings, data files, account settings...

Hmm.. no sad did everything just like on video lessons.

17

Re: Big difference between FSB backtest and MT backtest

Try to  use slot 0.01

Re: Big difference between FSB backtest and MT backtest

When you open the strategy from the repository, you save it with its unique name...... and then do whatever.... optimize, then save it again....

Of course if you restart FSB without having saved the thing you will have a mess.

I have done a lot of this sort of work and not yet experienced what you are describing, I do this all day long...... I think your technique is faulty, you have to be sure to save the one you want if you are going to close FSB and restart.

Any changes to FSB are going to change results.

Re: Big difference between FSB backtest and MT backtest

hey Dedok and all

When i downloaded the strategy posted here it gave me a bad results in FSB !!!! 
and when i checked the two indicators and how they excute trades within FSB i found there is no problem as well

THE ONLY THING I FOUND WRONG in the code and i donot know if this is the reason  of the problem or not :

Current code ;

// The NumericUpDown parameters
            IndParam.NumParam[1].Caption = "Digits";
            IndParam.NumParam[1].Value   = 2;
           [b] IndParam.NumParam[1].Min     = 1;[/b]
            IndParam.NumParam[1].Max     = 4;
            IndParam.NumParam[1].Enabled = true;
            IndParam.NumParam[1].ToolTip = "Number of digits to be rounded.";

You need to make the minimum = 2 not 1

Re: Big difference between FSB backtest and MT backtest

also i am running the same strategy''in post 1'' for one hour  linking all slots but i see only a down curve!!!

Post's attachments

b.png
b.png 6.35 kb, file has never been downloaded. 

You don't have the permssions to download the attachments of this post.