Econometrics: one step ahead forecast - page 111

 

As I understand it, the reversal point is predicted, then the reversals up and down are strictly alternating. So, we can hold an open position from reversal to reversal and then open in the other direction. If we simulate this and calculate the profit in pips, will it work?

 
faa1947:
You have the same thing. NS in packages (EViews does not have it, but others do) takes the place of smoothing, and this is only a small part of the problem and not the most important one that has to be solved. In the case of NS, it's an art. If you take splines and wavelets, it's mathematics

that's what I'm saying... they read all kinds of crap ...that ns is smoothing and then they brainwash people with it)))) faa1947 you don't realize that you're just being a jerk... it doesn't really matter what model you use .... if there is no information for a prediction you won't get it... it takes time and tests... lots of tests... but you have no time... you write articles and post...))
 
alexeymosc:

As I understand it, the reversal point is predicted, then the reversals up and down are strictly alternating. So, we can hold an open position from reversal to reversal and then open in the other direction. If we simulate this and calculate the profit in pips, will it work?

It's understandable, there is no prediction of reversals. What is posted raises very, very big doubts.
 
Vizard:

That's what I'm saying... they read all kinds of crap ...that it's smoothing and then they brainwash people with it))) faa1947 you don't realize that you're just being weird ... it doesn't really matter what model you use .... if there is no information for a prediction you won't get it... it takes time and tests... lots of tests... but you have no time... you write articles and post...))
More specifically. Words. The whole topic is one question: predictability by model, and the answer is again words.
 
Vizard:

That's my point... They read all this crap ... that NS is smoothing and then they brainwash people with it.
I'm talking about the place of NS in the modelling process - everything else is theory, irrelevant to the process. Could be right, could be wrong - I don't care. I'm sticking rigidly to the topic of the thread. There are problems that I am trying to solve publicly. NS does not solve any problems at all and in its place can be replaced by a simpler and clearer tool.
 

I don't know if anyone has told you, but the zigzag is a version of fractals. That is, the essence is the same only the lines.

The essence of fractals is that a new one is not drawn until the necessary number of "up" is reached... 5-6-7 and so on. The upwards concept is relative. To have a start point, the range (in the zigzag) is set in fractals and this is a timeframe.

Here's a tip of the research direction, look at the median, not the scale. This is the most common value in the range.


http://www.automated-trading-system.com/moving-median-better-indicator-than-moving-average/

 
faa1947:
More specifically. Words. The whole topic is one question: predictability by model, and the answer is words again.


You think you' ve written an article, opened a thread and people will come in droves to tell you what you've done... but why this or that won't work has been said before on the thread ... and not a few have said it ... this is already a big plus...

If you're a bit more specific ... for the future, I can tell you ... if you get a model you don't understand at the first step or so ... just take the formula and apply it to other data (sample) ... saves time and takes the euphoria out of it... good luck...

 
good luck...
Files:
median.mq4  2 kb
 
SProgrammer:

Here's a tip of the research direction, look at the median, not the scale. This is the most common value in the range.


I didn't see anything interesting in the link. Robustness is a kind of antipode of an unsteady market. You have to compare it with a market, not two displays of that market. You could probably compare the dummy to the median by comparing the residuals between the cotier and those two averages. The one is more robust whose residual is stationary. I think so.

There is one more objection to the median. It can be considered correct if your trading system is single-currency. But in median, unlike in the waving, the time point to which it is attached is always different. In multicurrency, we will compare price values referring to different time points. Initially, we put an element of non-comparability into such a TS.

 
Vizard:


you won't be specific... don't you get it... you think you'll write an article and start a thread and people will come in droves to tell you what you've done... but why this or that won't work has been said before on the thread and not in small quantities ... this is already a big plus...

if a little more specific ...for the future, I can give you a hint...if you come across a model that you don't understand at the first stage or so...just take the resulting formula and apply it to other data (sample)... saves time and takes the euphoria out of it... good luck...

What I see so far is different: people sitting there with their TA shoes on and puffing up their cheeks.