Machine learning in trading: theory, models, practice and algo-trading - page 2544
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
This is what I got. It seems to work quite well.... All "Rrrr..." go to the woods. Screw these opaque crutches. Only MT, only hardcore. Polynom, so polynom. But it's hardly anything new. Generally speaking, we got the well-known Expert Advisor of Yury Reshetov, only the selectable weights are several orders greater. Either everything new is something well forgotten old, or I do not understand something, even after reading almost the entire branch.
The number of deals on the forward is not enough.
The trader should make 10-20 moves, each time shifting all areas forward by forward length. You will get a rolling forward test. The sum of forwards will be more reliable estimate of the model.
Ten trades on the forward is not enough.
Do 10-20 training, each time shifting all areas forward by the length of the forward. You will get a rolling forward test. The sum of the forwards will be a more reliable estimate of the model.
would actually be a special case of YoH (https://lurkmore.to/Нёх).
actually a special case of YoH (https://lurkmore.to/Нёх).
Why not? Seems like a good clue to me. I recently came across an article here on how to automate this. But I'll have to automate training for it...) Anyway I'll have more confidence in ten forwarders stacked in a pile than in one. Probably there will be something bad at the forwarders junction, but it will do for stress. I don't know, I can try to use all symbols, I think there is no loss anywhere (currency pairs and spot gold and silver), there will be enough trades and everything will become clear in no time.
Why not? Seems like a good tip to me. I recently came across an article here on how to automate it. I should only automatize training for that) Anyway, I trust more ten forwards clumped together than one. Probably there will be something bad at the forwards junction, but it will be enough for stress. And maybe even more likely I will use a demo, all instruments at once, I think there are no plums (currency pairs and spot gold-silver) trades will be enough, everything will become clear quickly.
as soon as there are more than one forward.... your pants turn, turn
This is what I got. It seems to work quite well.... All "Rrrr..." go to the woods. Screw these opaque crutches. Only MT, only hardcore. Polynom, so polynom. But it's hardly anything new. Generally speaking, we got a well-known advisor to Yury Reshetov, only the selected weights are several orders of magnitude greater. Either everything new is something well forgotten old, or I do not understand something, even after reading almost the entire thread.
Rule number 1: as soon as you smell the grail, look for where the peepers
Rule #1: As soon as you smell a grail, look where the peepers are
From 0 bar you can only feed the Open price, HLC are not known yet. Or do not do it at all, Open 0 bar is almost equal to Close 1 bar.
Sceptorist - I hope you do not do so))
Why not? Seems like a good tip to me. I recently came across an article here on how to automate it. I should only automatize training for that) Anyway, there is more confidence in ten piled forwards than in one. Probably there will be something bad at the forwards junction, but it will be enough for stress. May be and even more likely I will use a demo, all the instruments at once, I think there are no plums (currency pairs and spot gold and silver) trades will be enough, everything will become clear quickly.
And also it's necessary to make a gap between the training section and the forward one. Last bars of the tray will have a similar result as the first bars of the forward. This is also a peek.
Automation is not difficult - in my Expert Advisor, the model retraining on Saturdays is implemented. As a result, in the tester we obtain forward gluing. In real trading it will be the same, i.e. it is trained on Saturdays and trades for a week.
to study on Saturdays and trade for a week.
And how?
And how?
Made based on https://www.mql5.com/ru/articles/2279 there is an example of regular retraining in the OnTimer() function.
Or one could think of fast waves as panic and slow waves as trends.
Not denying the waves... Waves are secondary, impulses of news / events are primary. Which turn into waves. But it is a level problem, what will be the waves from a pebble flying along such-and-such a trajectory, weighing such-and-such a configuration. Waves depend on all three and not only three parameters of the pebble, but also on external influences not only on the pebble...