Do you know how to make canals? - page 10
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
Ooh, excellent, substantive, all the more so if with articles, we're waiting :)
I, by the way, am fascinated by your intelligence. Articles on neural networks, fuzzy logic, interests in philosophy, religion, AI. It is the broad-minded person who is capable of solving complex problems.
A channel is good in that you can see the direction of the market, but the borders of the channel may not match expectations.
But it all hinges on the principle by which the channel is constructed.
Maybe Alexey can show us something interesting?
There are a lot of interesting issues here. Even the market direction itself (especially on small timeframes) may be apparent.
I'm fascinated by your intelligence, by the way. Articles on neural networks, fuzzy logic, interests in philosophy, religion, AI. It is the broad-minded person who is capable of solving complex problems.
Wow, thank you )
I calculate a moving probability density and set the probability level - let's say 0.9 - and build a band in which the price falls with that probability, which is the channel.
The problem is that prices that have less than 10% probability on history according to your algorithm start appearing much more frequently in the future. Moreover, after they occur, the market direction changes and does not follow from the previous one. Every 10 years this happens in the financial markets and events occur that, according to your theory, cannot happen in the next millennium. This is what is called NESTATION.
The problem is that prices whose probability on history according to your algorithm was less than 10% in the future begin to appear much more often than 10%. Moreover, after they appear, the direction of the market changes and does not follow from the previous one in any way. Every 10 years this happens in the financial markets and events occur that, according to your theory, cannot happen in the next millennium. This is what is called NESTATION.
I have not written here any particular theory; maybe we may create one together and solve all the problems. I am simply studying it experimentally. These probability distributions (based on sliding windows) show such big jumps as you mention and many smaller ones very well. Channels change not smoothly, but as energy levels between quantum states, i.e. by jumps. This non-stationarity has a kind of discrete genesis and is generated by strongly acting events, which unpredictably (for the characteristics of the time series itself) occur.
I may have phrased something inaccurately, but I agree with you.
I haven't written much theory yet; maybe we can create some more together and solve all the problems. I'm just experimentally studying. These probability distributions (based on sliding windows) show such large jumps, about which you write, and many smaller ones, very well. Channels change not smoothly, but as energy levels between quantum states, i.e. by jumps. This non-stationarity has a kind of discrete genesis and is generated by strongly acting events, which unpredictably (for the characteristics of the time series itself) occur.
I may have phrased something inaccurately, but I agree with you.
On the surface, the point of GARCH models is to combat non-stationarity. In reality, the point of these models is to construct robust models, which (robustness) refers tothe rate at which the model returns to a stationary state after a shock generated by the news.
this is why it is necessary to have an effective and shock-sensitive switch.
For example, baskets of strategies (not among forex traders, but among quants at exchanges, who are more experienced), such as trend counter-trend + some other, which are switched on and off, or strategy weight in the basket is changed. Of course, initial yields are not sky-high and baskets occasionally "tweaked", but they are good for decent capital management.
There is nothing special there, they just diversify as they can, and even large funds with a lot of professionals fall apart from time to time.
I've never seen strategies based on pure garbage models, neural networks or something else. Mostly some 1-2 inefficiencies are traded, randomly found, or a portfolio of strategies.
this is why it is necessary to have an effective and shock-sensitive switch.
For example, baskets of strategies (not among forex traders, but among quants at exchanges, who are more experienced), such as trend counter-trend + some other, which are switched on and off, or strategy weight in the basket is changed. Of course, initial yields are not sky-high and baskets occasionally "tweaked", but they are good for decent capital management.
there's nothing special there, they just diversify as they can.
I've never seen strategies based on pure garbage models, neuronet or something else... at least not in the public domain. My idea is to trade on one or two inefficiencies, found at random, or a portfolio of strategies.
GARCH is full of them, including currency market. The rugarch package - there are plenty of links. Basic work: Engle. HIGH FREQUENCY MULTIPLICATIVE COMPONENT GARCH, 2005
GARCH is full, including the foreign exchange market. The rugarch package - there are plenty of links. Basic work: Engle. HIGH FREQUENCY MULTIPLICATIVE COMPONENT GARCH, 2005
again, more for indices and stocks within them as far as I understand... which may grow for years. HFT is also not for everyone because of complexity and high cost of infrastructure. For forex it is more complicated
But i will have to do some more reading, thanks... on the other hand, i'm looking for high-yield options that do not last long but bring a lot of profit....kinda my style :)
On the surface, the point of GARCH models is to combat non-stationarity. In reality, the point of these models is to construct robust models, which (robustness) refers tothe rate at which the model returns to a stationary state after a shock generated by the news.
I.e. the point is to minimise the relaxation time of model parameter changes. + depot should not suffer during this time, for which(Maxim Dmitrievsky) switch (deversifier) strategies.