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
If you mean how to look at the optimisation results, I was talking about reading the results from the cache programmatically.
That's not what I meant. But cache reading is possible. You just have to write it.
Cool stuff!!!
Screwed Cycles to it as well and it was exactly what I was missing )))))
Thanks!
I added Cycles to it and it was exactly what I needed ))))
I need button-push functions in the Tester. Then automation would be of a different level.
I added some functions for the controls I need (Depo, Currency, Leverage, OptimisationType, BarsType, Criteria). But I haven't tested it yet. I can publish it afterwards.
I wanted to implement agent management. For example, you can click on the list of agents, press HOME, and select each agent with VK_DOWN. But I don't know how to call the context menu for the selected row. And there would be to press 'd'/'e' for disable/enable.
Can anyone give me a hint?
It's even harder to enable/disable the cloud (disable before tick or cross optimisation). If you even get to the context menu, you can't explicitly enable/disable, you can only flip it. You can't find out the current state programmatically. You can read from the terminal log "2019.07.31 19:27:35.664 Tester Cloud servers using switch off". It seems you can't read the log programmatically either, you have to look in the log file.
There is still a chance for undocumented WinAPI commands. Perhaps some C++ programmers know how to find them?
Faced with the problem of power outage. Now I don't know how much optimisation has been done ((((
When I have time, I'll try to add a check for % of total optimisation and worked characters. I hope the author does not mind? ))) Or will he want to do it himself?
Faced with the problem of power outage. Now I don't know how much optimisation has been done (((
MultiTester reflects its steps in logs. If the terminal has saved the logs, you will see everything there.
Moreover, the optimisation caches are 100% saved. So you can see what has been executed.
I added some functions for the controls I need (Depo, Currency, Leverage, OptimisationType, BarsType, Criteria). But I haven't tested it yet. I can publish it afterwards.
Before solving private problems, it would be good to describe scenarios of using their solutions.
Before solving private problems, it is a good idea to prescribe scenarios for using their solutions.
I described my scenario on the previous page.
Clicker is not a substitute for everything, it is one of several tools needed to automate optimisation. It allows you to do everything within one terminal without launching an additional terminal. I would add frames for analysing results (your new library for reading the optimisation cache is good for the same purpose), functions for managing input variables, SQLITE for storing states and statistics. And you can write intelligent optimisation scripts. Universality cannot be achieved, each Expert Advisor has its own script. And you have to create it manually and only then apply it repeatedly.
Frames have already been added. I had experience with SQLITE.
I just extended your lib for my needs. I'm sure you were planning it yourself, but I didn't wait. This is the only technology of the above that I didn't have a ready solution for. By the way, I checked my add-ons, they work. The agent management issues are still relevant. Spy++ can help in catching messages for an experienced systems engineer, but I had little help, I need to know what to filter. I hope people will pull up. The topic has been popping up since 2008.
Here is how I would like to improve it. My typical optimisation scenario:
Run N genetic runs for OHLC. Take the best result from each of them according to a customised criterion.
Perform slow optimisation on each group of parameters (each group has 2-3 interdependent parameters).
Repeat iteratively slow optimisations for each group until the optimum is reached (parameters stop floating).
Switch to real ticks, perform the same slow optimisations.
Switch to another symbol, repeat all.
I do not have such a task in front of me. Of course, it will be good if automation will allow this.
I myself am striving to develop a criterion for the robustness of the TS through auto-optimisation indicators. It is very much lacking for more objectivity in research.
The scheme of the auto-optimiser has been completely thought out, there are no technical obstacles for its implementation.
Bibla definitely needs to be supplemented with functionality. So please suggest your realisations.