Market phenomena - page 14

 

Slava, there are approximate formulas that prescribe the number of intervals when constructing a histogram.

Of course, the interval should preferably be larger than the sampling step (1 point). Otherwise you get a sieve instead of a histogram...

 
Mathemat:

Slava, there are approximate formulas describing the number of intervals for the histogram.

Of course, it is desirable to take an interval larger than the sampling step (1 point). Otherwise you will get a sieve instead of a histogram...


Well, it's a sieve. Why there aren't troughs up to zero is because of the alps switching to an additional sign. After this transition, the gradients started to fall into intervals that previously could not even theoretically fall into :)
 
Avals:

Well, it's a sieve. Why there are not troughs to zero, it is because of the transition of alps to the additional sign. After this transition the increments began to fall into intervals that previously could not get even theoretically :)


there is no "sieve", it doesn't affect it in any way. It's my personal right, with how many intervals I will build the histogram. I build it for visualisation and nothing more. It just helps to see the phenomenon, that's all. Here are the intervals of the phenomenon:


The phenomenon itself is two processes, one "bulls" and one "bears", in this arrangement. The same effect is obtained in a different way, by not applying the histograms at all. I have written about it. I will post it one day.

 
Farnsworth:


There is no "sieve", it does not affect it in any way. It's my personal right how many intervals I build the histogram with. I build it for visualisation and nothing more. It just helps to see the phenomenon, that's all. Here are the intervals of the phenomenon:


The phenomenon itself is two processes, one "bulls" and one "bears", in this arrangement. The same effect is obtained in a different way, by not applying the histograms at all. I wrote about it.


which is what I'm trying to reproduce. If the spread is less than the number of intervals, then some intervals will not contain the increments. So why do you take more intervals than the value of the spread in points?
 
No, why, the intervals are 27, and the range is at least 46 points (0.002435780 - (-0.002221510)). Taking into account the fatness of the tails it is much more, twice as much at least.
 
Avals:

Well, that's what I'm trying to reproduce. If the range will be less than the number of intervals, then in some intervals increments will not fall. So why do you take the number of intervals greater than the value of the spread in points?


To see the "fine structure" and nothing more. Visually the second "sub-process" becomes visible. This histogram does not take any part in the calculations. It no longer matters who gets where. Next, other algorithms, and obtaining highlighted sub-processes.

Tail phenomenon - more obvious, but gives the same result. I wanted to post them together, but didn't have time. Or rather, almost the same, this algorithm gives a "cleaner" series with less noise.

 
Mathemat:
No, why, the intervals are 27, and the range is at least 46 points (0.002435780 - (-0.002221510)). Taking into account the fatness of the tails, it is much more, at least twice as much.

So there are 27 intervals here and on the histograms on the first page there were up to 700
 
Farnsworth:


To see the "fine structure" and nothing more. Visually the second "sub-process" shows up, it just becomes visible. This histogram takes no part in the calculations. It no longer matters who gets where. Next, other algorithms, and obtaining highlighted sub-processes.

Tail phenomenon - more obvious, but gives the same result. I wanted to post them together, but didn't have time. Or rather, almost the same, this algorithm gives a "cleaner" series with less noise.


But how do you check if the histogram doesn't already count? :)
 
Avals:

but how do you check if the histogram doesn't already count? :)
Explain it to me, country boy. Are there any holes or have they run out?
 
paukas:
Explain to me, country boy. Are there any holes, or have they run out?

There are no more holes - more subtle structures are at work :)