Neuromongers, don't pass by :) need advice - page 12
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
Starting from 15M TF, the history from different sources differs very little, and starting from 1H OHLC prices almost always coincide. Of course, if you do not use a story found on a dumpster somewhere. I don't understand what some people are talking about here either....
I agree with you. Moreover, the data for this TS is clearly filtered by Hodrick-Prescott filter (oh, now they'll scream about redrawing and looking into the future).
But it's not even about the absolute coincidence/difference of quotes from different providers. It's just that some people don't understand how NS works - if history differs at the back, it also differs at the front. Roughly speaking, if NS is trained to approximate a sine function, it will work just as well on cosine.
What some comrades here are talking about I just don't understand either....
It's just that some comrades don't understand how NS work
Judging by the mediocre articles, some comrades themselves don't understand a lot of things...
If the story is different at the back, it is also different at the front.
"Masterpiece", you can go straight into the annals...))
If NS is trained to approximate a sine function, it will work just as well on cosine.
Here it approximates something that didn't and doesn't actually exist. The "polished" story is plus and the real story will be minus.
Judging by your mediocre articles, there's a lot you don't understand...
Here she is approximating something that didn't and doesn't really exist. The "polished" story is a plus and the real story will be a minus.
You've got me confused with someone else.
Ugh, for fuck's sake, what the fuck?
Fuck you!
What an idiot, man!
You are cruelly mistaken.
You got me confused with someone else.
Ugh, what the fuck?
Fuck you!
What an idiot, man!
https://www.mql5.com/ru/articles/55
Apparently inadequate as well))) Spring break, I guess...)
All right, take care.
...
All right, take care.
And don't you cough.
Please to the moderators: please delete the off-topic posts, including those slandering the historical quotation data of MetaQuotes Software Corp.
Let's move on.
I think we should experiment further. Have you already tried different learning/retraining periods? I'd be interested to see how the Expert Advisor behaves during one year of training, the trade period can be shortened as well...
How symmetrical is the result of the NS? I.e. if a pattern gives a buy, does its mirror image necessarily give a sell? Is the trend component excluded?
I'd be interested to see how the EA behaves when trained for a year, the trading period can also be shortened...
Ok, as soon as I adjust the normalisation. This will probably give a good qualitative improvement. Then I will do test runs with different parameters.
How symmetrical is the result of NS? I.e., if some pattern results in buying, its mirror image will surely result in selling? Is the trend component excluded?
I want to devolatilise with this one:
Criticise before it's too late :) hopefully I'll post the devolatiliser itself soon.