Machine learning in trading: theory, models, practice and algo-trading - page 1208
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
HH: Python itself can be plugged into C# (wrapped) and Python scripts can be run from files, there are plenty of examples on the web, and if you can from a file, you can drop commands directly from MT into Python
Not the case. Maxim wants to use Python, not separate functions.
That's good.
It's not about performance...but this version seems to lack consistency in results and are having strange behavior sometimes...
Hence, I am using external filters to filter out trades if the optimisation results are not good.
I don't know how it works...)) But it seems to work in out of sample data...))
I will see live testing next week.
strange behavior can be when you stop optimizing manually, so some of files are not writes (not always, but sometimes it can happen... about 1 time on 50-100 stops)
also, after every optimization must be a single pass in tester, to reset the maximum errors files, from previous optimization
you can't find out.
reset, for example,
Ohhh...NO:)) Do I have to run the tester after for every optimisation?
If I restart the MT5 after optimisation, do I still need to run the single pass test?
2. After every optimisation is over, is it required to apply the EA again to the chart which I used to do in previous versions?
I use NET Sockets in C++ to do this. It's a high-level lib, unlike winapi. It's easy to roll it into C# for MT.
and here's https://www.mql5.com/ru/docs/network/socketcreate
hahahaha
Ohhh...NO:)) I have to run the tester after for every optimisation?
If I restart the MT5 after optimisation, do I still need to run the single pass test?
2.After every optimisation is over, is it required to apply the EA again to the chart which I used to do in previous versions?
1 yes, or you can just delete "lasterrors" file from common mt5 folder, or your next optimization will start from old errors, not from zero
2 sure, for upload new files
but here https://www.mql5.com/ru/docs/network/socketcreate
hahahaha
Creepy.
Creepy.
That's it, no crutches, no need.
I was just using a task scheduler to schedule all these tasks like testing, restarting MT5 automatically everyday...now, the tasks numbers will increase exponentially:))))
Also, thanks for telling these things, otherwise, I would be simply wondering around without knowing what is going on:))
maybe I can fix it in last versions
Creepy.
I remember the horror of trying to implement simple matrices in MQL for SSA porting, everything was ok, but the lack of multidimensional dynamic arrays in MQL was a huge crutch, I don't want to re-invent the wheel anymore, it's easier to use .If the developers have made support for C# then they should take advantage of it
Anyway, I have almost 99% automated the process using MQL5. The last 1% can be done either using python or autoIT. I am using "AutoIT".
But I hope there are no other mysteries in this version of the EA:))
If there are any other changes in this version, you can update me...
Because I have done a series of tests on out of sample data data and I get 90% of the times profit which is a good sign for success in forward testing :)))
I think nothing... that thing with files was described in article, but hard to understand for you on russian