Difference calculus, examples. - page 19

[Aleksey Panfilov]

From the indicators in theMT5,MT4 code base discussed in this thread, put together this "Pseudo Fourier" combo. ))

The wide red line is the bottom line, it is not redrawn.

I suggest to discuss in the topicEMA of the third degree, I'm learning to use intraday.

[Mihail Marchukajtes]

I will allow myself to comment on the subject as well. The animation is indeed impressive and quite suitable for demonstrating the frequency market analysis approach, but what is the practical point of it? Build signals on the kinks in your curvature and you will be bitterly disappointed with the trading results. Just an example: МАшка as already said with a simple average 2 already starts lagging, and this is the simplest transformation and the more complex the transformation is the more it lags, the exception is CSSA analysis, which in principle is what you are trying to repeat. I remember the year 2006 or so these works appeared in the form of the leading and not redrawing indicator, which basically increased the interest of the market community in general, but I think there is a good tool for creating a cointegration of time series, as the avatar shows the prognostic ability of one frequency to another. I remember I even managed once to get a perfect circle on Lissajous figure that says the frequency I found, namely the period and the harmonic are prognostic at the moment in this section of the market and for some time my TS worked, not long but it worked and surprised me. However, I failed to repeat this success, because I had to do it manually and Neroshel's optimizer did not want to work with this tool. I still find this direction in frequency analysis very relevant, because I've got needed settings for CSSA, even if by accident. The main thing was the direct evidence that at the moment the frequency and period are in front of the market, and not behind it. There were 5-6 signals, but what a signal. If you do not know, there is no one who has repeated this tool for MT or any other terminal. And looking at those pictures you've posted earlier in the branch, I now understand that such a tool for analysis may well be created, but only by a much more experienced programmer than me.

[Aleksey Panfilov]

Mihail Marchukajtes:
Позволю себе тоже высказаться по теме. Анимация действительно впечатляет и вполне подходит для демонстрации подхода частотного анализа рынка, но какой в нем практический смысл? Постройте по перегибам своей кривулины сигналы и выбудете горько разочарованы по результатом торговли. Просто пример: МАшка ка уже было сказано с простым усреднением 2 уже начинает запаздывать, а это ведь самое простое преобразование и чем преобразование сложнее тем он сильнее запаздывает, исключение конечно является CSSA анализ,что Вы в принципе и пытаетесь повторить. Помню еже так в году 2006 примерно появились эти работы в виде опережающего и не перерисовывающегося индикатора, что в принципе подогрело интерес к рынку сообщество в целом, но там вернее тут как раз таки думаю вполне годный инструмент по поводу построения коинтеграции временных рядом, как раз на аватарке инструмент который показывает прогностическую способность одной частоты к другой. Мне помнится даже удалось один раз получить идеальный круг на фигуре Лиссажо, что говорило о том что найденная мною частота, а именно период и гармоника являются прогностическими в данный момент на данном участке рынка и какое то время ТС работала, не долго но работала что меня и удивило. Однако повторить данный успех у меня так и не получилось, потому как там подбор нужно было делать руками и оптимизатор нерошеловский не хотел работать с этим иснтрументом. До сих пор считаю данное направление в частотном анализе актуальным, поскольку пусть и случайно, но я получил нужные настройки для CSSA. Тут главное было то что было прямое доказательство того что в данный момент частота и период идут перед рынком, а не за ним, там буквально сигналом 5-6 было, зато какие. Насколько мне известно до сих пор никто не повторил этот инструмент ни для МТ ни для другого терминала, я во всяком случае не встреччал. И гляда на те картинки которые Вы выкладывали ранее в ветке, я теперь понимаю что подобный инструмент для анализа вполне возможно создать, но только гораздо опытным программистом чем я.

There may be someone, and hopefully more than one, who will check it out. )))

P/S.

For discussion of the last indicator, better use the topicEMA third degree, I am learning to use intraday.

[Mihail Marchukajtes]

Aleksey Panfilov:

There may be someone, and hopefully more than one, who will check it out. )))

P/S.

For discussion of the last indicator it would be better to use the topicEMA ofthe third degree, I learn to use it intraday.

There's nothing to discuss, you have already been told it won't work, it's a dead end. Unless you use Lysajo figure + CSSA. In the end this frequency analysis, it all started with FATL and SATL, what a word I remember :-), waste of time, meanwhile, frequency analysis came to us from the circuitry and the device that measures frequency, what is it called? Right oscilloscope, that is essentially need to take a digital oscilloscope to determine the current frequency of the selected area, and put it on the axis XY and perpendicular to the axis, let it be XZ pick up a frequency that appears on the screen of the fucking oscilloscope circle, the more idealized its circumference will be better prediction, but honestly say that earnings in the market consider that 3-4 guaranteed signal is a victory. IMHO of course. Getting this figure implies that one frequency goes after the other. It should be defined not to confuse what is running after who, kotir or our with you CSSA or simply popularly "Caterpillar" Saw my writing Fokunik would shed a tear from the surging memories. FATL, SATL, Caterpillar. Those were the days. But as the saying goes, all new is well forgotten old, so go ahead, but there are no fish there :-(

[Aleksey Panfilov]

Mihail Marchukajtes:

There's nothing to discuss, you've already been told it won't work, it's a dead end, it's been tested. Unless you use Lysajo figure + CSSA. In the end this frequency analysis, it all started with FATL and SATL, what a word I remember :-), waste of time, meanwhile, frequency analysis came to us from the circuitry and the device that measures frequency, what is it called? Right oscilloscope, that is essentially need to take a digital oscilloscope to determine the current frequency of the selected area, and put it on the axis XY and perpendicular to the axis, let it be XZ pick up a frequency that appears on the screen of the fucking oscilloscope circle, the more idealized its circumference will be better prediction, but honestly say that earnings in the market consider that 3-4 guaranteed signal is a victory. IMHO of course. Getting this figure implies that one frequency goes after the other. It should be defined not to confuse what is running after who, kotir or our with you CSSA or simply popularly "Caterpillar" If you saw my writing focal point would shed a tear from the surging memories. FATL, SATL, Caterpillar. Those were the days. But as the saying goes, all new is well forgotten old, so go for it, Alexey, but there's no fish there :-(

Mikhail, thanks for your messages.

I have got you wrong.

Good luck.

[Mihail Marchukajtes]

Aleksey Panfilov:

Michael, thank you for your messages.

I misunderstood you.

Good luck.

I am ready to support the conversation in terms of practical application of such transformations. A smart approach to creating TS does not hurt at all. For example if you trace the expected volatility in the market and work only in periods of its growth, then I think the volatility level itself can be attached to the frequency or period by some artful manipulations and in this case the results of transactions can be improved. Naturally IMHO...

[Aleksey Panfilov]

Mihail Marchukajtes:

I am ready to support the conversation in terms of the practical application of such transformations. A smart approach to creating the TS would not do any harm at all. For example, if you trace the expected volatility in the market and work only in periods of its growth, then I think the volatility level itself can be attached to frequency or period by some cunning manipulations and in this case the results of deals can be improved. IMHO of course...

Sorry. This task is not interesting for me anymore. ((

[36A8C2F9]

Aleksey Panfilov:

The red broad is a sliding line constructed by interpolation with a fourth-degree parabola. It is not redrawn (the analogues were discussed at the beginning of the page to the seventh). The nodes, if I understood correctly, are the four previously drawn values and the new price by which the parabola of degree 4 is selected and the fifth new value is drawn on it.

The curve blue line (not redrawn can be considered a trace of the straight blue line) is the center line, each point of which is plotted on the last three points on a wide sliding line from the assumption that they lie on the sine wave of a certain period, just as each pointof the straight blue line is plotted on three points on the already correctly extrapolated sine wave (gray).

Only the extrapolated grey sine and the straight blue line are redrawn.

P/S.

If you implemented your idea to isolate the oscillations, you should have gotten a line very close to a sinusoid with variable amplitude and reversal (a kind of quantization).

Just for such a line to investigate the extrapolation by a sine wave is relevant.

[Aleksey Panfilov]

Intermediate results of the branch.

In an attempt to generate a timely (not lagged) signal, touched upon:

Interpolation by polynomial and extrapolation by polynomial or sine.

Polynomial interpolation and removal of the first or second difference, analogous derivatives.

Implemented byNikolai Semko in theBanzai.mq4indicator(post#57).

The figure shows the line averaged by the polynomial of the 4th degree and corresponding difference 1, 2 and 3 from top to bottom. As it should be for sinusoidal functions, we see a shift to the left, about a quarter of a period, in each successive derivative. To reduce the shift, we can increase the interval between value removal points.

I would like to point out another possibility to get a timely signal.

It is simply a sinusoidal averaging.

There are 7 first lines of the indicator averaged by the 2nd degree with the leverage of 50. Only the periods differ. The red one has period 1, parabola. Orange has period 150, sine wave near the constant. Yellow 130, sine wave near constant. Green 115, sine wave near constant. Blue 110, sine wave near constant. Blue 104, sine wave near constant. Purple 103, sine wave near constant.

It must be said that the higher the degree of interpolation, the more difficult it is to select parameters at which the averaged line behaves adequately. It has already been mentioned#64#97 that no reasonable averaging is obtained by a polynomial of degree higher than the fourth. The introduction of cosines seems to soften the coefficients of the difference equations and the degree of interpolation can be increased. On the other hand, certain ratios of period (cosine) to shoulder can "spoil" the averaging line even with a degree less than the fifth. For example, even with second-degree averaging(that's one sine near the constant) if the leverage increases, approaching half a period and beyond, then the amplitude grows to infinity. This is logical, because values half a period away on a sine near the constant must be equal, so there is a conflict between the last counted value and the value removed by the lever.Increasing the leverage to a reasonable amount shifts the averaged line to the left, that is, its lag decreases. An example is shown in the figure.

[Aleksey Panfilov]

Sine averaging.

The blue line is the construction diagram, the orange line is the result.

Leverage 72, sine period 153.