Ward 6 - page 18
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 am not interested in your processes. Of course, if I take a sine wave as a "process", I can predict it into the future and filter it out with a prediction. I asked for an example of a linear filter algorithm, not an example of a filtered signal.
Dear Michael, I understand that you have a cool friend who is an expert on filtration and you believe him and all that. However, I am afraid, you just heard the bell, but did not understand where from. The linear filter theorem says that there is no linear filter that gives a non-lagged response for any input process. But this does not mean that there is no corresponding linear filter for that particular process. The same can, with some caution, be extended to quasi-linear filters (those with floating or adjustable parameters).
Strictly speaking, the class of processes for which no non-delayed linear or quasi-linear filter exists is very narrow: At least all (!) partial autocorrelation coefficients (for purely linear filters) must be identically equal to zero and the variation of partial ACF (for quasi-linear filters) must be zero. So don't talk about a sinusoid.
A clue in a more explicit way
Seems fantastic. On the difference we make a (statistically valid) "naive" prediction - a straight line to the future, and in relation we observe a "probability preponderance sign" :-)) Who will refute it?
The linear filter theorem states that there is no linear filter that gives a non-lagged response for any input process.
Strictly speaking, the class of processes for which there is no non-delayed linear or quasi-linear filter is very narrow.
We do not want to filter a sine wave or a meander. clearly, there is no lag-free linear filter for a stream of quotations.
Where's the sci-fi, I don't get it? And what kind of sci-fi?
Strictly speaking, the class of processes for which there is no non-delayed linear or quasi-linear filter is very narrow: they must at least be identically equal to zero (!) all partial autocorrelation coefficients (for purely linear) and zero variation of partial ACF (for quasi-linear). So don't talk about a sinusoid.
What kind of nonsense is that? Do you agree that there is no lag-free linear filter for the EURUSD chart? And in your opinion, something there is identically equal to zero, and moreover, not one but several of this something are equal to zero. Are you out of your mind? You have discovered a fundamental property of any quote stream that you think it should have, so that you do not get upset?
Where's the fiction, not clear? And what fantasy?
A clue in a more explicit way
Seems fantastic. On the difference we make a (statistically valid) "naive" prediction - a straight line to the future, and in relation we observe a "probability preponderance sign" :-)) Who will refute it?
Dr.Drane, and the MASHA sometimes shows that. ONCE IN A WHILE. So does your filter, sometimes.
The problem with false inputs hasn't gone anywhere. Here you say it's not a lagging filter. So what artist does it have a divergence with the price!
khorosh:
Can you define what "lagging" is? What is "smoothing" is intuitively clear. It is a reduction in the volatility of the filter output compared to the input. You have to "smooth out" but not acquire the lag. What is "lag" is not clear (strictly - not clear, intuitively, on a domestic level - quite clear). All filters have these two values rigidly connected. If you smooth it, you get lag. And vice versa. A good example is simple moving averages (SMAs).
As a result of long communication with a specialist in the filtering theory, I realized:
1. This is only true for linear filters. For non-linear filters, unlike linear filters, there is no strictly proven prohibition in principle on the existence of a "non-delayed smoothing filter".
2. About your "although I do not know how to express it in numbers" - it is not only your problem. The very concept of "lagging" cannot even be formulated in traditional language. For linear filters, everything is formulated in terms of the transfer function (AFC & IF). Absolutely all results described in the literature for non-linear filters belong to a narrow class of filters - "linear filter with slowly changing parameters". For such filters the notion of AFC/FF also exists (they also change slowly), accordingly the creation of a "non-lagged smoothing filter" is also impossible.
3. For arbitrary nonlinear algorithms the notions of AF/MF are meaningless, so the concept of lag, and the measure of lag are not defined in any way. And creation of a non-delayed (at least in common sense, "by eye") filter is not forbidden.
But I can't give strictly any definition of "not delayed". You can use a clever trick, go to the end, define it axiomatically, for practical application it is enough. Namely, let's call non-lagging a filter, in terms of which a game with TP=SL will show profit. It is elementary. Conversely: if khorosh: thinks that my filter is lagging, let him take any other lagging (even smoother) filter - for example a regular SMA - and try to repeat my tricks in public .
For nonlinear filters, unlike linear filters, there is no strictly proven prohibition in principle on the existence of a "non-lagged smoothing filter". What I call NDNRF - No Delay & No Redrow Filter.