Econometrics: one step ahead forecast - page 133

 
faa1947:
My students do not confuse a random variable with its realization.

What on earth makes you think I'm confused? There is no noise in the market, kind sir. Write it so that when you go to the mirror you will be reminded of it and will not waste your time with nonsense. As for how much you have managed to mess up in your posts, I will tactfully keep silent.

Keep your barbs to yourself and speak out if you have something to say.

The substance? You're going back to the bottle? And then I have to patiently explain that your astrolabe does not work and can not work? faa, it seems to me that you are the realization of a random variable, with no memory.

 
Farnsworth:

No noise in the market kindly.

No no so no. No statistics, no probability, no stochastic diffurs and Markov switching - nothing. All determinism.

I agree in advance with all your posts.

 
faa1947:

No noise in the market kindly.

No no so no. No statistics, no probability, no stochastic diffurs and Markov switching - nothing. All determinism.

I agree in advance, with all your posts.

Then make trades on your trend. And explain to TC that the price should be here and not there. I do not mind, there is noise, and the quote is:

quote = trend + noise + periodicity + seasonality.

Let it be so. I don't mind. But your formula may be rewritten in the following way:

cotier = trend + noise.

There is no periodicity and seasonality - you kind of proved that already, if you remember. You have deftly solved the noise problem, the trend is left.

 
Farnsworth:

then make trades on your trend. And explain to the DC that the price should be here and not there. I don't mind, there is noise and the quote is this:

So be it. I don't mind. But your formula can be rewritten already so:

There is no periodicity and seasonality, you kind of already proved that, if you remember. You have deftly dealt with noise and the trend is left.

About periodicity it is not so. Let's see:


This is bar 236 from 02.01.12 18:00 to 16.01.12 18:00 - exactly two weeks of watchdogs. Here we see 4 large, strong trends and a bunch of small ones - all with different periodicity. The periodicity is there!

 
faa1947:

About the periodicity, this is not the case. Let's see:


This is 236 bars from 02.01.12 18:00 to 16.01.12 18:00 - exactly two weeks of watchdogs. Here we see 4 large, strong trends and a bunch of small ones - all with different periodicity. The periodicity is there!


Take a random walk and count the spectra in the sliding window. You will be very much surprised.
 
faa1947:

About the periodicity, this is not the case. Let's see:


This is 236 bar from 02.01.12 18:00 to 16.01.12 18:00 - exactly two weeks of watchdogs. Here we see 4 large, strong trends and a bunch of small ones - all with different periodicity. The periodicity is there!

Make up your mind, you're wiggling around like a marquee boat (C). You wrote on a previous page https://www.mql5.com/ru/forum/136555/page132 "It is safe to say that there is no periodicity ... " Why are you wiping your posts??????? Scientist, BL****MY Why isn't there any periodicity in that picture??????

You wrote about spades :o)))

What in the ZHJZHOOOPPUUU periodicity? Where did you find it? On a model that doesn't fit the market at all???? Onwards and onwards.

 
Farnsworth:

You've got to make up your mind, you're wiggling around like a marketeer's boat (C). You wrote on a previous page https://www.mql5.com/ru/forum/136555/page132 "It is safe to say that there is no periodicity ... " Why are you wiping your posts??????? Scientist, BL****MY Why isn't there any periodicity in that picture??????

You wrote about peaks :o)))

What in the ZHZHOOOPPPUUU periodicity? Where did you find it? On a model that doesn't fit the market at all???? Onwards and onwards.

Let's look, most importantly calmly, at the window size. in the last one it's 236 bar.

Where did you find it?

We see that there has been grouping around certain periods. There is no market model at all.

on a model that doesn't fit the market at all

I would like to point out that I am using a recognised model which has a very clear macroeconomic explanation. But that is not what we are talking about now. We have an obvious grouping, although it was not present in the large sample.

 
anonymous:

Take a random walk and count the spectra in the sliding window. You will be very much surprised.
I don't know how to take the SB, Then what SB? I see where you are going with this. At 80000 bars it's SB, there's no trend, it's all sideways.
 
faa1947:
I don't know how to take the SB, then what SB? I see where you're going with this. At 80000 bars, that's SB, there's no trend, it's all sideways.

Unfortunately, everything is there - trends, flotsam, all the shapes of graphical TA.
 

Смотрим, главное спокойно, размер окна. в последнем - это 236 бар. Мы видим, что произошла группировка около определенных периодов. Тут вообще нет модели рынка.

No, you deal with your own illusions.

I would like to point out that I am using an accepted model, which has a perfectly clear macroeconomic explanation. But that's not what we're talking about now.

What??????? This software(http://fx.qrz.ru/) uses MESA, which has nothing to do with the generally accepted econometric model. None. Nothing at all. When are you going to stop making things up?

We have an obvious grouping, even though it wasn't in the large sample.

What "obvious" clustering. You think that's how you prove cyclicality? "You see obvious clustering" is it? Science-intensive :o)