Libraries: MultiTester - page 14
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
For an unknown reason, running a test multiple times unloads the script. It is not always repeated, sometimes earlier, sometimes later. It just happened at 413 iterations, for example. IsStopped is set to true, and the script is unloaded, locking all the work in the middle of the case. There are no error messages. Debugging messages showed that this happens after Sleep (100) after SendMessageW (Handle, WM_COMMAND, ID_EDIT_PASTE, 0 );
in a call to static bool SetSettings (const string Str)
Below is the minimum set to play. But I suspect that the matter is still in MetaTrader. Build 2245, but on the last 2253 the same. Build 2190 also suffers from this, perhaps earlier too.
Hi. Any solution for this problem?
I'm having the same problem using the MT5 build 2265.
Thanks.
Now EAs running in the Tester can receive their data.
Great.
Do you plan to consider managing the number of activated agents? I haven't found any approaches. And when switching to the postictic test, it makes sense to deactivate some agents.
No urgency, just a desirable development.
I have also today updated win_10 to build .535: Multitester stopped working altogether, but differently than here. It gets confused when going through the list of pairs and starts chasing one in circles. This happened before, but rarely and after long calculations. Here is a piece of the optimisation list:
Terminal 2280, multitester is the latest. Please don't ignore win_10, MQ will not support win_7 forever....
Do you plan to consider managing the number of activated agents?
I don't plan to, as there are no tasks where it is needed. I may use multitester primitively, but I do it only on custom symbols on real ticks. Therefore, based on the need to leave some resources free, I set up an appropriate number of Agents. As a rule, one pass lasts about a second, i.e. not much resources are needed.
MTTester.mqh + TesterCache probably can do everything except tst-format.
In plans (something implemented long ago) - it is no longer libraries, but tasks themselves. For example, you introduce a new input parameter into an Expert Advisor. You need to evaluate its influence on the results of the TS. Walk-forward is full-fledged. Portfolio selection from several TS, etc.
And also in the plans is the task to run a multi-tester for thousands of optimisations and from these thousands to find single interesting places. These will be researches where, indeed, no human foot has ever set foot. The computational complexity is very high. That's why I'm thinking of ways to quickly find a needle in a haystack.
As soon as the tst-format is known, the possibilities will expand very much.
win_10, MQ will not support win_7 forever.
I'm sitting on a custom minimalist Windows 7 Embedded build. Everything works, no reason to upgrade.
If MultiTester was finished in the middle of a case, on the next start it will say it didn't finish last time and ask if it should continue from the old location. If you continue, it will go to TesterSettings.Deinit(Pos - 1); where it will have an if (this.iDeinit[Pos]) inside after the first run is finished. But the iDeinit array is empty, which will essentially crash it. Maybe something else may fall off somewhere else, I haven't tested it further.
I'm sitting on a custom minimalist Windows 7 Embedded build. Everything works, no reason to upgrade.
Last night, a windup update arrived (at least I installed it yesterday).
I too have win_10 updated today to build .535
Tester logs need to be looked at.
If MultiTester was finished in the middle of a case, on the next start it will say it didn't finish last time and ask if it should continue from the old location. If you continue, it will go to TesterSettings.Deinit(Pos - 1); where it will have an if (this.iDeinit[Pos]) inside after the first run is finished. But the iDeinit array is empty, which will essentially crash it. Maybe something else may fall off somewhere else, I haven't tested it further.
Thanks, updated.
Thanks, updated.
Is it possible to zipped already or not? (It's like - what if it works ))))