a trading strategy based on Elliott Wave Theory - page 59

 
<br / translate="no"> There are a few questions:
1. What to take for the inflow? The full price, the modulo difference, only the positive difference? In other words, does the concept of "influx" in the method in question have any effect on the pre-preparation of the data? Or should the data to be investigated be taken as the influx. I intuitively, for example, took the closing price in my calculations.


Inflow is also inflow in Egypt. That is, in the classical interpretation it is necessary to take the difference Close[i]-Close[i+1]. From my reading of Peters.
 

Есть несколько вопросов:
1. Что брать за приток? Полную цену, разность по модулю, только положительную разность? Другими словами, имеет ли понятие «приток» в рассматриваемом методе влияние на предварительную подготовку данных? Или следует за приток принимать данные, которые надо исследовать. Я интуитивно, к примеру, взял в расчетах цену закрытия.


An inflow is also an inflow in Egypt. So the classical interpretation is to take the difference Close[i]-Close[i+1]. From my reading of Peters.


Thanks. But the difference Close[i]-Close[i+1] is often negative (it may be OK in Egypt).
Is the difference modulo or as is? And where can I read works of Mr. Peters?
 
It was here - http://stock01.narod.ru/ And actually solandr gave a link in this thread to some astronomy department, and it's 3 pages long.
 
It was here - http://stock01.narod.ru/ And actually solandr gave a link in this thread to some astronomy department, and it's 3 pages long.


Probably missed it when reading the forum materials.
 
I have read several chapters of "Chaos and Order in Capital Markets" by E. Peters on the calculation of the Hearst index. I did not find anything about "what is an inflow".

From my engineering view Close[i]-Close[i+1] is very different from Close[i] series. In its essence it is a very different series. If you take it modulo, it probably resembles a graph of potential profits, and it seems questionable to make assumptions for Close[i] based on its difference. But what if, for example, I want to analyze profits? Should I take the difference from the difference? It seems to me that I should simply take Close[i] for inflow, if I want to analyze Hearst for it and not its difference.

In my calculations I'm confused by the average inflow. Or should I take one number calculated for N for all n observations or for each n on a segment from 1 to N I must calculate its inflow? Who would answer?
 
The level of the reservoir... it changes... in some random way. More water comes in, less water comes out. So there's a difference in inflow minus outflow. That's the difference that causes the level to fluctuate. We need to understand if the changes in level are random or have a trend, we need to know if it's drying up or overflowing. We measure water level every year and get graph. We also need to understand from the graph whether it is an accident or a tendency. Maximum water level minus minimum is our spread. Changes between successive years are random variables. Measure the standard deviation for N years and compare it with the spread. If the ratio is too big - it is not a chance, if it is small - it means that the level will be broken through neither upward nor downward. It's the same with the price - we should compare the price swing with the random increments of this price.
 
grasn, page 12 shows the algorithm to calculate the Hearst index according to Vladislav's recommendations. Read posts
solandr 15.05.06 19:09
Vladislav 15.05.06 21:18
 
On the same website there is E. Peters' Fractal Analysis.
There, somewhere on page 69, is a recipe for counting. 69 there is a recipe for calculation.
If I understand it correctly, log(Close[i]/Close[i+1]) is used, and all partitions into equal segments of lengths from 1 to N are used.
 
On the same website there is "Fractal analysis" by E. Peters. <br/ translate="no"> There's a recipe for counting somewhere on p. 69 there is a recipe for calculation.
If I understand it correctly, it uses log(Close[i]/Close[i+1]), and also uses all partitions into equal segments of lengths from 1 to N.


Log-normalization is mainly relevant for stocks on a long time horizon.
 
The level of the reservoir... it changes... in some random way. More water comes in, less water comes out. So there's a difference in inflow minus outflow. That's the difference that causes the level to fluctuate. We need to understand if the changes in level are random or have a trend, we need to know if it's drying up or overflowing. We measure water level every year and get graph. We also need to understand from the graph whether it is an accident or a tendency. Maximum water level minus minimum is our spread. Changes between successive years are random variables. Measure the standard deviation for N years and compare it with the spread. If the ratio is too big - it is not a chance, if it is small - it means that the level will be broken through neither upward nor downward. It's the same with the price - we should compare the price spread with the random increments of this price. <br / translate="no">.


Do I understand correctly that in our case we take Close[i] "as if" for the level in the reservoir? If so, the inflow will be the modulus of the difference Close[i]-Close[i+1]?