Errors, bugs, questions - page 505
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Same function returns zero. I triedSERIES_TERMINAL_FIRSTDATE too. The conclusion is that theSeriesInfoInteger function doesn't work in the tester. It works only for the instrument on which the test is running. Maybe you have some ideas how to run the test for instruments with different dates of history start when you select the "All history" interval.
I decided to check your conclusions. I ran the test Expert Advisor on Alpari-Demo and the same in the tester in visualisation mode since 2000. The results of init execution are in the attachment.
If the history for the tested period is on the server, SeriesInfoInteger() returns in the tester the same data as it was on the demo. If there is no history (like on EURAUD, from 2004.06.16) at the date of testing, tester returns error: 2011.09.05 09:40:53 EURAUD: contains 0 M1 records.
So, SeriesInfoInteger() still works in the tester, if the server has the corresponding history and the tester can load it. If it hits the "empty" history, the test simply hangs, although the progress bar is moving.
I decided to check your conclusions. I ran the test Expert Advisor on Alpari-Demo and the same in the tester in visualisation mode since 2000. The results of init execution are in the attachment.
If the history for the tested period is on the server, SeriesInfoInteger() returns in the tester the same data as it was on the demo. If there is no history (like on EURAUD, from 2004.06.16) at the moment of test start date, the tester returns error: 2011.09.05 09:40:53 EURAUD: contains 0 M1 records.
So, SeriesInfoInteger() still works in tester, if server has appropriate history and tester can load it. If it hits "empty" history, the test simply hangs, although the progress bar is moving.
strangely I have returned zeros in the tester, what I do wrong?
And you set the test interval not to 'All history' but so that the history is guaranteed for all characters, e.g. 'Last year'. Then there should not be zeros, but the beginning of the history uploaded by the tester. I get it from 4.01.2010.
P.S. But actually strange, if the tester downloaded the existing history with 2004g. , it should give out and date of the beginning of history 2004 for this symbol, and not zeros, even if the start date of testing earlier than the existing history. I think this is a bug.
Apart from the fact that the time in MT5 terminal suddenly jumped from GMT+1 to GMT+0, I also found a mistake - after the end of the just passed one-hour candle the volume values in previous candles have changed, and significantly. I can only say for certain about today's candles - I saved all of them successively from the beginning of the session: 0:00 1841 - 2269, 1:00 2556 - 3235, 2:00 2436 - 2975, 3:00 1700 - 3096, 4:00 1391 - 1690, terminal time is "wrong". That is incorrectness with MT5.
And you set the test interval not to 'All history' but so that the history is guaranteed for all characters, e.g. 'Last year'. Then there should not be zeros, but the beginning of the history uploaded by the tester. I got it from 4.01.2010.
P.S. But actually strange, if the tester downloaded the existing history from 2004. , then it should issue a start date of history in 2004 for this symbol, and not zeros, even if the start date before the existing history. I think this is a bug.
Discovered while normalising the lot. Check yours:
Print(NormalizeDouble(0.07,2);
Contrary to expectation, prints: 0.070000000000000000001
Discovered while normalising the lot. Check your own:
Print(NormalizeDouble(0.07,2);
Contrary to expectation, it prints: 0.07000000000000001
https://www.mql5.com/ru/forum/4033
I'll try to ask the question a third time.) MetaDriver showed an example here. This is where I gave my example.
The chart shrinks when the number of trades starts to exceed 3000. Is this problem considered by the developers?
This problem does not allow us to analyze trade results in the terminal at once when testing the system on a large strip of history data when the amount of deals may be about 10000 and more.
I remember I had something similar in Excel. But the program was overloaded with complex formulas and simply hung up if the number of lines exceeded 5000. What could be the problem here?