Please lend me a working profitable Expert Advisor, I'll share $$$ - page 16
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
How much time did you spend on it? Was it better than neuroshell?
It took a little over two hours to learn the grid.
I don't know about neuroshell, as I don't use it, so I can't compare.
can you at least outline what this "hellish code" does?)
my use of this "hellish code" is based on the history on D1. i recently downloaded a program that creates digital filters - indicators and it creates indicators based on the history
so this "grid" works only on history, if you change the period/testing period, then, behold:
i.e. again we need to prepare the data for the history
here's the code of the digital indicator, may Reshetov help to learn his "grid" online, there's a dll for calculating parameters on the "fly" in the program (indicator has a link)
I don't know how to use this "awesome code", but I recently downloaded a program for creating digital filters - indicators, it also creates indicators based on the history
so this "grid" works only on history, if you change the period/testing timeframe, then, behold:
I.e. again we have to prepare data for history
Of course, on H4 compared to D1, the range of bars is obviously smaller, which is why the result is so low.
Theoretically we can add training examples from at least all timeframes and the result will be a universal neuronet. There are only a few problems here:
1. Where can we get this very history to feed it into MT4 and convert it to a training examples file using an Expert Advisor?
2. The computer capacity for all timeframes may not be enough. Approximately 64 cores are required to calculate everything in a reasonable time + all training examples should be placed in memory, not in swap.
Here is the code of the digital indicator, maybe Reshetov can help you learn your "grid" online, there is a dll for calculating parameters on the fly in the program, in the code (there is a link to the indicator)
I'll give it to you in good hands. Not a bit sorry.
Not much profit, but enough for a loaf of bread.
Is this with the forward section, or only on the learning interval?
This is a rough fit to the history obtained on the previous version of R-Net.
The new version includes check for contradictions in the training sample and the grid immediately after the first pass detects that there is rubbish on the inputs - the training examples are highly contradictory and stops, instead of looking for a Confucian black cat in a dark room when there is none, as it is done in usual neuro-packages. There are no miracles.
Mendeleev first dreamed of a station with Shurma, and then he decomposed it into elements.