Bayesian regression - Has anyone made an EA using this algorithm? - page 20
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
Do you have any other textbooks ? :)
You seem to be missing the point.
SanSanych has not answered the question of what residuals should be stationary when testing the TC.
Apparently, this is a big mystery... :)
It's no mystery... Just forgot the gist of it.
This should not be forgotten - these grains of truth should be kept!
I am interested in the question of TC stability for a long time, I came across only a criterion of "fit" of TC to the test sample.
At a certain value of the criterion TC is considered fit and unsuitable for work.
Apparently yes, I studied from textbooks that you don't recognise. :)
Why are you attached to ARIMA?
If you trade in deviations, it may be ARIMA, but you still need to prove the stationarity of the series to which you apply this ARIMA. And there is such a tricky thing there....
And some skilled traders try to trade trends, to which they apply ARIMA, i.e. they predict deviations and transform them back to the trend.... and then they start to scold econometrics.
For those who like to trade trends, to be more precise: not trends, but trading absolute values of kotir (probably a level breakdown or something?).
There is a package called forecast. It's very well-known and widely used. I've been using it myself to forecast the sales of my own products.
So this package decomposes the initial quote (I want to stress - the initial quote) into three parts: the trend, deviations from the trend and the cyclic component. Then it predicts these three parts for n steps ahead and gives the result.
I'm ready to cooperate on this subject as well. If I find any developments, I will do it as soon as possible.
This follows directly from the non-stationarity of the original series.
Sit down, two.
Write a couple of formulas for the entities in question (the simplest ones, don't get too complicated) -- and try to figure out what the quandary is. Then you'll know where to take your irony.
In MQL4 forum your phrase was"Forward testing for a TS, the residue of which is not stationary, is self-deception... ".
This should not be forgotten - these grains of truth should be kept!
I am interested in the question of TC stability for a long time, I came across only a criterion of "fit" of TC to the test sample.
At a certain value of the criterion TC is considered fit and unsuitable for work.
Yes, I did... I don't remember...
If we talk about the tester, this is the problem, in my opinion.
We take some sample and use the tester to calculate, for example, the profit factor. Then we take another sample and get a new profit factor value. Altogether we get two figures. Are these two figures the basis for statistical conclusions? These figures don't mean anything at all.
It must be solved and is solved differently.
A sample is taken. Some subset is randomly selected from this sample and counted as a profit factor on it. Then again a random sample is taken and so on, for example 1000 times. You will get 1000 profit factors. This set can already serve as the basis for statistical conclusions.
By the way, this method does not exclude using a tester.