Obtaining a stationary BP from a price BP - page 16
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
I can send you a demo account.
>> Who can I sell it to?
>>Who can it be sold to?
Don't you have anything else to sell other than Reshetov's leaked stream?
I can send you a demo account.
So what was the mistake? >>State is not interesting, it is interesting to hear the analysis of the loss ;-).
>>Who can it be sold to?
I'll give it away for nothing. >>Yourself, that's your own problem.
So what was the mistake? State - not interested, interested to hear the analysis of the plum ;-).
Who the fuck knows? It's gone down and that's it. The market didn't care that there were no errors in the code and the formulas were all correct.
... It's gone, and that's it. The market didn't care that there were no errors in the code and the formulas were all correct.
A negative result is also a result.
And one more confirmation that even a successful forward (and not even one) does not guarantee success.
Nevertheless, he who looks will find, or at least has a better chance to find.
to LeoV
Проблема адаптивных ТС заключается в том, что они тоже переучиваются по некоему алгоритму, который в них заложен, но он может не совпадать с алгоритмом изменения рынка. То есть, на некотором промежутке времени алгоритм изменения рынка может совпадать с алгоритмом переобучения ТС, а потом может "разъехаться". Рынок меняется не по заданному алгоритму - в этом вся и проблема.....
Perhaps my allegory about the plane was prettier :o)
to Svinozavr
Exactly! Nothing to add! :о)to Neutron
All this time I've been investigating with great effort the problem of searching for the optimal length of the training sample and its relation to characteristic time of quasi-stationarity of processes on the market. It turns out that the required value is at the level of 5-10%. Then it has to be retrained. The brokerage company commission in its turn defines the minimum size of the price movement, and a gradual but sure increase of market efficiency with the growth of the trade horizon defines the operation field unambiguously. And I kind of settled on that.
I'm glad for your success. For example how DC defines the minimum size of price movement, but I'm sure you understand what you wrote :o(
As for the answer to your questions regarding BP conversions, I am not at all aware of the feasibility of such a conversion. Your "... it will allow you to use standard stat-processing methods ..." does not mean anything.
I am a little confused. At the very beginning you suggested all sorts of ways to reduce the influence of non-stationarity, but to my simple suggestion - to bring the series to a stationary one you faint. The only way to reduce the influence of non-stationarity is to bring the series to a stationary one. There are no other options, don't make it up. Everyone is doing it, don't be shy and don't get all worked up over nonsense. (Don't forget about stationarity of Fourier C's, for example) And here's your idea:
It seems to me that one way (perhaps the only way) to deal with the non-stationarity of the generating BP, is to use adaptive methods in the TS. To do this, the system must retrain no later than the characteristic time of stationarity existence (if there is no stationarity at all, then the attempt to outplay the market is pointless in principle).
It's just not working. You don't have the slightest idea of when the misalignment will occur, and the misalignment doesn't even have to be catastrophic! Once a month? Can you simulate the market a month in advance? Seryoga! I'm on my way to buy plasticine and create your monument!!! I have a week ahead at most (I've shown off several times in the pages of the wonderful https://forum.mql4.com/ru/20423/page73 branch :o)
Please state the concept itself.
The concept is simple, - to use techniques where they work (AR, Cs, ARIMA, FARIMA, etc.) It applies to NS as well. :o), and besides use methods of model identification. In your case - identification is impossible in principle. There is no reason to define it.
The market is a really hard model to identify. That is understandable. There are so many freeloaders in the market who want to make money. Take the current chart of eurjpy. A characteristic feature is that the big uncles have not divided something, judging by the chart. And when they start an argument, they do it by taking out all the reasonable stops of the petty ones. And they do not care what Papkin thinks about the stationarity of the market. This is reality. Everything else is a myth. The stupidity of the situation with the search for a gray method must simply be understood as futile. And work in a different direction. For example, I've decided for myself a long time ago that my personal experience in trading is much better. I met one person not so long ago, a girl trader. She uses her intuition. She often gives good points in the market when one can enter and take profit without any advanced statistical tools using ordinary MetaTrader indulators.
The market is a really hard model to identify. That is understandable. There are so many freeloaders in the market who want to make money. Take the current chart of eurjpy. A characteristic feature is that the big uncles have not divided something, judging by the chart. And when they start an argument, they do it by taking out all the reasonable stops of the petty ones. And they do not care what Papkin thinks about the stationarity of the market. This is reality. Everything else is a myth. The whole stupidity of the situation with the search for a graphical method must simply be recognized as futile, and you have to work in a different direction. For example, I've decided for myself a long time ago that my personal experience in trading is much better. I met one person not so long ago, a girl trader. She uses her intuition. In addition to that she often uses her intuition and it often allows to enter the market and take profit without any advanced statistical means using simple Metatrader's indukes.
For instance, I didn't say it's that easy. This method is difficult, but it will ease a lot. For some reason, stationarity is identified with profit, but it is not; it may be a necessary condition, but it is an insufficient one. And as for intuition, you cannot cheat nature, and it is non-stationary.) This is exactly the reason why I started treating forex as a business long ago.
Who the fuck knows? It's gone and that's it. I don't give a shit that there are no errors in the code and the formulas are all correct.
Well, that's some kind of unproductive approach. There should be work on mistakes - mistakes in the method, not in the code. Can we at least know the timeframe, sample size and grids? ;-)