Filter without delay - page 8
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
There's nothing in the box. Maybe try again?
>> I'm repeating the private message.
Your indicator has generated undoubted interest.
Allow me to make a few prncy judgements from my point of view.
All your indicators are designed to work in stationary markets where price forecasting is possible. But the market is not a stationary market - it's generally accepted.
When you make a prediction, the frequencies have changed, and more importantly the frequency has changed.
Without agreeing on the model of the rinket it is useless to talk about the Expert Advisor.
The following model is proposed:
there are several trends on the market, which are created by the group of dominating players for a certain period of time. Apart from them, there is some noise.
The dominating trends are in different correlation with each other, but they always create a trend which is clearly visible on the ZZ. The ZZ spread may be larger (interesting for the position) or smaller (not interesting - consider it a sideways trend). Therefore it is the market reversals that are of interest, not the future price forecast.
I have an EA built on Berg, but I don't know how to work with the phase, which I believe predicts market reversals. I'm interested in wailets which have a time duration apart from the ACH. But I would like, maybe with your help, to finalize my Expert Advisor. It has good characteristics with P/F up to 10 but not less than two. The situation is worse with recovery factor: the market goes away and the factor starts to fall and very often it is less than one.
Alex.
Gentlemen, let's think about what a digital filter is and what a real filter is. First of all, we need to understand what a real filter is.
A filter is something that transfers the input "something" to the output. :)
For example vibration - and a spring is a filter, a shock absorber in a car is a filter, a piece of microporous rubber is a filter.
Vibration, in the example - as it is from the realm of mechanics and you can touch it with your hands.
There are active filters, those that apply external energy, as if actively fighting the vibration. Passive filters are springs and the same spring dampers but with a different damping coefficient.
Phase is the moment at which the input is initiated. Phase distortion can be linear or non-linear from the frequency, and this is how much of a delay there is in the filter.
There is no such thing as a filter without phase distortion. It's just that the distortion always depends on the frequency of the filter. For example with vibration - there is a frequency of 1 hour - like the ship is rocking, and there is 0.01 seconds - like the engine... Distortion at frequency of 1 hour in microporous rubber will be mega small and in relation to the frequency itself, more precisely to the period - will be negligible. And over a period of e.g. 0.00000001 second the vibration is practically not transmitted at all. Why the phase is important :) Because you have one wave, not 300. And you can not miss it, and if there are 300 of them, so there is one half-wave.
Anyway, all of you and I are building some kind of filter. :) It's just that their transfer function is VERY, VERY complicated and so are neuro filters. :)
I'm repeating the message.
Your indicator has caused obvious interest.
Allow me to make some strong judgement from my point of view.
All your indicators are designed to work on stationary markets, where price prediction is possible. But the market is not a stationary market - it's generally accepted.
When you make a prediction, the frequencies have changed, and more importantly the frequency has changed.
Without agreeing on the model of the rinket, it is useless to talk about the Expert Advisor.
The following model is proposed:
there are several trends on the market, which are created by the group of dominating players for some time period. Apart from them, there is some noise.
The dominating trends are in different correlation with each other, but they always create a trend which is clearly visible on the ZZ. The ZZ spread may be larger (interesting for the position) or smaller (not interesting - consider it a sideways trend). Therefore it is the market reversals that are of interest, not the future price forecast.
I have an EA built on Berg, but I don't know how to work with the phase, which I believe predicts market reversals. I'm interested in wailets which have a time duration apart from the ACH. But I would like, maybe with your help, to finalize my Expert Advisor. It has good characteristics with P/F up to 10 but not less than two. The situation is worse with recovery factor: the market goes away and the factor starts to fall and very often it is less than one.
Alex.
Ok, I'll see what's up with the wavelets, I'll post later. Can you give a precise definition of what phase is? It seems to me that we mean different things by this word.
In general, if you have an idea of the algorithm, or at least an outline of it, drop me an email at fxlab64 dog gmail dot com
I have my own development, platform for MTS, works through MT4, written in C#, there is a connection to Matlab and MS SQL Server 2008. So I have something to build it on)0
Ok, I'll see what's up with the wavelets, I'll post later. Can you give a precise definition of what phase is? It seems to me that we mean different things by this word.
In general, if you have an idea of the algorithm, or at least an outline of it, drop me an email at fxlab64 dog gmail dot com
I have my own development, platform for MTS, works through MT4, written in C#, there is a connection to Matlab and MS SQL Server 2008. So i have something to build the theme on)0
I am very glad to receive feedback. I suggest using terms from Matlab and corresponding TOOLbox only. There's a lot of literature on walettes, and again, I suggest using from Matlab, or we'll never figure it out.
Gentlemen, let's think about what a digital filter is
A filter is a tool, either that tool has been honed by a crowd of people for two hundred years (PF) or 30 years in the case of the wailettes. In either case, you can get something for free without having to invent it. Fourier and attempts to apply it to the market are widely known (finwares, for example). But none of the works publicly and systematically completed, as was the case with MESA. Kravchuk started and then gave up half way through. And Matlab allows the SPM to be evaluated. IMHO one should enter the market when there is an SPM area of one of the frequencies comparable to the sum of all the others.
But the fundamental disadvantage of Fourier is the stationarity of the signal. For non-stationary markets we suggest the wavelet. Very similar, but the frequency starts and stops - the trend started and stopped. Matlab contains more than a hundred wyelet functions. I'm convinced that market reversals are necessary, especially if we can calculate the confidence interval of such a reversal.
Он имеет непрохие характеристики с P/F до 10, но не меньше двух. Хуже обстоит дело с фактором восстановления: рынок уходит и фактор начинает падать и очень часто меньше единицы.
P/F > 2 and recovery factor < 1. How is that?
Фильтр - это инструмент, либо этот инструмент оттачивал толпа народу лет двести (ПФ) либо лет 30 - в случае вейлетов. В любом случая можно не изобретая получить что-то на халяву. Широчайше известен Фурье и попытки уго применения на рынкете (финвары, например). Но ни одна из работ публично систематически не доведена до конца, как это было сделано с MESA. Кравчук начал, а потом бросил на пол дороги. А Матлаб позволяет оценить СПМ. ИМХО входить в рынок надо, когда имеется площадь СПМ одной из частот, сравнимая с суммой всех остальных.
Но принципиальный недостаток Фурье - это стационарность сигнала. Для нестационарных рынков предлагается вейлет. Очень похож, но частота начинается и заканчивается - тренд начался и заончился. Матлаб содержи более сотни вейлет функций. Мо-моему убежденид на нажны развороты рынка, особобенно если мы сможем считать доверительный интервал такого разворота.
The question was about the logic of understanding the physical nature of the filter. And in particular, that there is always phase distortion. :)
In a word - attempts to do something with filters in the usual sense of the word (I mean filters) and all mathematics, which was worked out for it. For financial speculation utopia. :) Not dangerous, but utopia.
I tried to explain it here and elsewhere. Alas.
As for wavelets, they are a thing for another purpose. You simply don't need to use it here. It's not impossible or impossible - we just don't need it. And that is how it is done. They are for something else.
P/F > 2 и фактор восстановления < 1. Это как?
It's a titanium butt with carbon fibre inserts.
2 VDev: Can I ask for another picture? All the same as on the first one + on top of two more lines (or more) which are drawn at the points of values of each of the two indicators
at moments of current time (i.e. before re-drawing). For the "or more" variant - through the points, equidistant from the present for small intervals (like 5-10-15 ....).
The aim is to see the process of redrawing in dynamics, maybe something interesting will turn up. I don't think it's easy for you, but still?
P/F > 2 and recovery factor < 1. How is that?
Profit factor = max profit / max loss. Recovery factor = max drawdown / max profit. These are different values and characterize the TS very differently