You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Suppose I am able to model a synthetic that has a distribution of increments, the shape of the ACF, the shape of the ACF of the squares of the increments, and the behaviour of the variance over time quite similar to what is actually observed. How can this help us to model the price?
Now it remains a little bit, to find a mat model, which has the same shape of ACF and has the same behaviour of variance. Determine the parameters of this model. And according to the results of the model's work, check its forecasting accuracy and time horizon, on the forward area.
So you should start with this, without a model, all the arguments above are just shaking of air and nothing more. I am generally surprised that an almost empty article has led to two and a half pages of discussion.... (or am I wrong, and the author has planned a series of lectures here?)
Trolls:
теперь осталось чуть )), найти мат мадель, которая имеет такую же форму АКФ и имеет такоеже поведение дисперсии. Определить параметры этой модели. И по результатам работы модели проверить её точность прогнозирования и временной горизонт, на форвардном участке
The model is there, but the matapparatus I know is not enough to identify both the parameters and one of the important functions.
strange with your knowledge of matrix algebra ((. You can also communicate via Skype, sometimes it is easier to explain by voice than by wiping the keyboard. leave me your coordinates in a private message and I will contact you.
Suppose I am able to model a synthetic that has a distribution of increments, the shape of the ACF, the shape of the ACF of the squares of the increments, and the behaviour of the variance over time quite similar to what is actually observed. How can this help us to model price?
To be honest, I'm completely "off topic" here.
What is "synthetic", "ACF"? Why can't we calculate the price if we have a glass model? I tried to read the articles, but it didn't get any lighter.
And the main thing is that anything can be modelled. But what do we want from the model - to predict the rate, the behaviour of the glass, to draw rate-like curves?
Where do you want the horse to go?
Why are you trying to model the glass and not directly the rate and volumes? Maybe it would be better not to model a glass, but to model the behaviour of agents as in econophysics?
In short, there is no clear statement of the problem.
To be honest, I've gone completely "off topic" here.
By "synthetic" I meant that I know how to generate time series, which have some properties similar to those observed in time series of quotes. I do not model stakes.
The point is that the properties I mentioned are not enough for price forecasting. Something more is needed. It is the same with random volumes and stack models.
strange with your knowledge of matrix algebra ((. You can also communicate via skype, sometimes it is easier to explain by voice than by wiping the keyboard. leave me your coordinates in a private message, I will contact you for sure
Thanks for the offer, but for now I have other directions for research.
So it is necessary to start with it, without the model all the arguments above are just a shaking of the air and nothing more. I am generally surprised that an essentially almost empty article has led to two pages of discussion... (or am I right and the author has planned a series of lectures here? (or am I wrong and the author has planned a series of lectures here?).
Thanks for the positive assessment of my modest work ("almost empty" is still more than zero).
Special thanks for the idea of "lecture series". I will write it.
What would you like to read in the next lectures?
Interesting article. Without unnecessary maths, loaded with mind-boggling formulas for nerds (of which there are only a few, but they piously believe that it is FOR THEIR LIGHT EDITION that articles are written), everything is clear and understandable. I don't quite agree, but I won't argue, everyone has their own vision. However, I have learnt some points from the article....
Tothe author: there are few nerds, they rub their fingers in the blood, sitting on the forum and leaving piles of unnecessary messages. You can't please everyone. The article is written for us - mere mortal adequate people, devoid of pseudo-intellectual posturing, of which 99%. For my part I will say thank you for the article.
The topic is very interesting, there are questions
1- Does the (short term) market have a memory?
2. How are the values of the trendiness indicator obtained - there are some values there, how are they obtained? Didn't see the formula.
3. When you write about trendiness - does the new tick for forex depend on the previous tick? Or two ticks or how many more? But in general it can be written in. And in general it makes no sense to talk about trendiness, saying that your coin is trendless - but it is true.
4. But from point 3 we can draw a very important conclusion.
- you can try to find a trend if you remove the trendlessness from the real picture.
i.e. z(x)=x(x)-y(x)
Why a coin, exactly? It has two sides - what do they reflect? Only an ideal random wandering on a straight line (analogue - up, down), i.e. one-dimensional. The price can have another state - flat, i.e. it is already a coin with three sides, i.e. we have a 2-dimensional random walk. The above charts show that such a market state is practically not modelled - a hard flat is not seen once.
If we take the market chart by ticks and not by time, there is no flat, there are only fewer ticks per unit of time.