Topic: Inconsistent test results FSB-MT4

Its strange. I already make two robots on eurusd and gbpusd. Their backtests on FSB and MT4 the almost identically.

Now I'm working with usdjpy and tests on FSB and MT4 tester do not mach. I try different experts and always the backtest results a seriously different on FSB and MT4.
Even at the different data sources balance graph not so bad. But on the tester MT4 very different result.
I adding files and screenshorts.

Post's attachments

files.zip 229.06 kb, 3 downloads since 2016-12-27 

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

Re: Inconsistent test results FSB-MT4

Probably different data, you will have to verify that the data in the two programs is the same, I suggest that you redownload FSB data and see if that fixes the problem

There should be no reason to waste time running MT4, I can assure you that FSBPro is much more accurate and the diagnostics in FSBPro are far superior.

Once you have established that your data in FSBPro is accurate, there is no need to be running MT4 except as a bridge to the market

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....

Re: Inconsistent test results FSB-MT4

Your tester report says the following - Control points (a very crude method, the results must not be considered)

4 (edited by AnarchoTrader 2016-12-27 20:42:17)

Re: Inconsistent test results FSB-MT4

footon wrote:

Your tester report says the following - Control points (a very crude method, the results must not be considered)

On the other tickers this test method worked normal. Tick test in MT4 is still not correct, tics a artificial, and it terribly slow.
But I will try.

Re: Inconsistent test results FSB-MT4

The other question, is why System quality number is zero?
http://savepic.net/8734765m.png

Re: Inconsistent test results FSB-MT4

System Quality Number is a ratio calculated by a formula..... I suggest that you establish a minimum on the Control panel so that you do not record inferior strategies as they will turn out to be a waste of time.
Read about System Quality Number on the internet and then apply it in the Acceptance Criteria of Control Panel.

If you accept strategies that are a low quality you will have tremendous frustration when you start to use them.

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....

Re: Inconsistent test results FSB-MT4

AnarchoTrader, it must be zero for loosing strategies only. Please send me the strategy for a test. It must be a bug or at least it looks like one.

Re: Inconsistent test results FSB-MT4

Hmm.. very strange... Perhaps after adjusting usdjpy symbol properties. I changed the method for calculating the Commission to the more correct, from money to % (by the way, Miroslav, please do more digits after the decimal point in the field for the Commission - can not fit 0,002% -  20$ for 1,000,000).
Now all that founded strategy in the repository are displayed with zero SQN and do not satisfy the Acceptance criteria. While previously it was not. Now even with a marked checkbox - not show it - it showed.

And now Generator strategies can no longer find any normal strategies. All new generated have SQN = 0 and not fulfil Acceptance criteria.

http://savepic.net/8750113m.png

http://savepic.net/8772640m.png

http://savepic.net/8759328m.png

Re: Inconsistent test results FSB-MT4

Popov wrote:

AnarchoTrader, it must be zero for loosing strategies only. Please send me the strategy for a test. It must be a bug or at least it looks like one.

I add the file.

Post's attachments

usdjpy_m5_001.xml 10.55 kb, 3 downloads since 2016-12-27 

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

Re: Inconsistent test results FSB-MT4

Mhm, It works with my settings. Probably the other settings make the difference.

http://s23.postimg.org/bfu0fkmmv/screenshot_2016_12_27_at_22_51_01.jpg

Re: Inconsistent test results FSB-MT4

AnarchoTrader wrote:
footon wrote:

Your tester report says the following - Control points (a very crude method, the results must not be considered)

On the other tickers this test method worked normal. Tick test in MT4 is still not correct, tics a artificial, and it terribly slow.
But I will try.

Hmm. On tick-test in MT4 and MT5 the results are correct!

Re: Inconsistent test results FSB-MT4

Popov wrote:

Mhm, It works with my settings. Probably the other settings make the difference.

http://s23.postimg.org/bfu0fkmmv/screenshot_2016_12_27_at_22_51_01.jpg

What settings I must look? On one screen is a clearly bug - checked filter not worked out.
http://savepic.net/8750113m.png

Re: Inconsistent test results FSB-MT4

you can use the sliders on the left side....

put a value of, say, 1 in system quality.

BUT, first. attach the strategy so Popov can examine for a bug because he  thinks there is a little bug that is affecting the output.

Let him examine that strategy to detect whatever might be a problem before you try to do much more.

I have had the same problem. and thought it was because of a bad strategy, I did not realize was a bug.

My 'secret' goal is to push EA Studio until I can net 3000 pips per day....