Let's talk about something. - page 3

 
A little bit of concrete information appears.
 
sayfuji >> :

That's the way it is! Don't you understand about predicting when the price reaches a certain level? Together with the prediction of that level, even allowing for the margin of error, you have a complete prediction of price behaviour. Doesn't that sound too sweet?

Hence the rawness and the discussion - the topic is interesting, but there are no serious expositions in the public domain on this topic.

It turns out that you have already tried it? Your results, as well as mine, are not very good. I have also tried to predict significant points using Fibo's price/time ratio. I also used indices. I got nothing.

 

Maybe it works out differently for someone else. Maybe this issue has not been fully researched. I think it is possible to make a profitable TS based on Fibo, applying the price ratio over time. It just takes desire and donkey's stubbornness. Something at least not draining, I think, can be devised.

 
this came to mind... about fractals, take H1 and look for a fractal on m5 (for example), let the neural network do the fractal search.
 
SProgrammer писал(а) >> let's talk about something.

Why don't we drink remotely but together? ))))

 
LeoV писал(а) >>

Why don't we drink remotely but together? ))))

Are you drinking?

What do you mean, already? :)

 
SProgrammer писал(а) >>

Are you drinking?

What do you mean, already? :)

I mean drinking, but not drinking and not already....))))

 

The results were not very encouraging, but that is understandable. It's too big of a task. I am attaching an example of why it does not work. Of course the waves are outlined in my opinion, but I think it's clear that nothing is clear. The standard time zones, don't show much.

As for alternatives, it is the price+time achievement that has been well "beaten" by Gunn, we all know his results. But the English literature is not much explanatory and hard to read, and the Russian one all the more so. Hjerjik is the only one.

So it is also a mystery covered in darkness. Not for nothing, but because hardly anyone wants to share - and quite rightly so.

 
sayfuji >> :

The results were not very encouraging, but that is understandable. The task is too big. I am attaching an example of why it does not work. Of course the waves are outlined in my opinion, but I think it's clear that nothing is clear. The standard time zones, don't show much.

As for alternatives, it is the price+time achievement that has been well "beaten" by Gunn, we all know his results. But the English literature is not much explanatory and hard to read, and the Russian one all the more so. Hjerjik is one.

So, it is also a mystery. Not for nothing, but because I don't think anyone wants to share - and quite rightly so.

I, for one, don't even know if we are taking the right time? What time parameters should we take in general? I don't mean tamframes. Maybe the number of real ticks of this or that movement plays a role here. Maybe the time intervals depend on the timeframes we consider. I do not understand what time is to begin with. About price changes everything is clear and specific. The number of pips is the price change. But the second parameter is time? What is this parameter? I have not studied the theories about it. I confessed in the beginning.

So I don't understand what this parameter is - time. Where is this parameter even explored in the literature. I haven't come across it.

 

Oooh, that's a long talk. Physicists - theorists love to talk about time. Many works have been written on this topic. As for the market, the constructive thread is lost here - too little research has been done on this topic. You can try to take some time intervals, by activity bursts only, try to work on your own noise filter and in general, the article was about drawing the candlesticks thickness by their volume, you can try to develop it in this way.

It is difficult to say something in a nutshell, because we have complete annihilation (just kidding), it's just that the topic of time prediction is not very perspective because of its absolute difficulty. I would like to see price forecasting in the event of normal drawdowns.

Reason: