From theory to practice - page 354

 
Yuriy Asaulenko:

I already gave you an example with random walks. At the incremental level there is a normal distribution. And where's the prediction there? except, the best prediction is the current value.

There's a normal distribution there too, not at the level of increments. if you take 100 outcomes at a time and make a sliding window.

 
igrok333:

there is a normal distribution and not at the incremental level. if you take 100 outcomes at a time and make a sliding window.

It doesn't matter at all. Already in the notorious article by Kolmogorov, 1940 issue, there are no requirements for the shape of the distribution at all. There is only one requirement for the series - stationarity. And the article considers the possibility and forecasting conditions only and exclusively for the stationary time series.

Where did the normality requirement come from in the topic? - I do not know)).

HZ the possibility of predicting non-stationary series in the article simply is not considered, ie, the article in relation to such series simply says nothing. It absolutely does not mean impossibility of forecasting nonstationary series, but, simply, in relation to non-stationary series, means nothing at all.

 

Here is a comparison of the forex distribution (in blue) with the normal distribution (in red).



What are the differences we see?

The forex distribution is more elongated upwards.



And it is more narrowed on the sides.



What does this tell us?

 
igrok333:

Here is a comparison of the forex distribution (in blue) with the normal distribution (in red).



What differences do we see?

The forex distribution is more elongated upwards.



And it is more narrowed on the sides.



What does this tell us?

You didn't see the most important thing, think about it.

 
Maxim Dmitrievsky:

You haven't seen the most important thing, think about it.

And then there are the tails. longer and thicker tails.
 
igrok333:
And then there are the tails. longer and thicker tails.

Well, that's what this whole topic is about :)

 
Maxim Dmitrievsky:

Well, that's what this whole topic is about :)

And who and why did they interfere? What's the point?

 
Yuriy Asaulenko:

And who and what did they interfere with? What's the big deal?

Too much dog talk

 
Maxim Dmitrievsky:

well, that's what this whole topic is about :)

well having thick tails doesn't mean it has to go backwards after such big increments on the chart.
 
Maxim Dmitrievsky:

too much dog talk.

That's good.) If they weren't, then really, ouch.

ZZY the Nightingale Razboinik: Ilya Muromets is a good man, but when he gets drunk, everything goes wrong for him - you fly low, you whistle softly.