a trading strategy based on Elliott Wave Theory - page 213
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
... Порой, умираю от смеха, сталкиваясь с замечаниями воинствующих флудеров! Это просто цирк какой-то, им бы в детсад - математику подучить, ан нет, возраст не тот! Генералы, блин...
:) Pay no attention, it is a test sent down to us from above, by the grace of our father and his son and his spirit, as a test of the strength of our faith :::)))
Amen!
I am engaged in everything, but mainly, of course, from stochastic methods point of view. The same problem about decomposition, but apparently not in a pure form, as it is formulated by the classics.
North Wind, in a nutshell, please, about the "discord".
I have read this topic in full, with interest, at least because I myself went down this path. Personally I liked the "caterpillar" from time analysis methods. But again, I could not use the pure method of time series analysis.
And did the path lead to a dead end?
I, at one time, was also into the Caterpillar.... Obviously, the method is not suitable for predicting the exchange rate of currency instruments. It is based on algorithms for detecting the cyclic component and deterministic trends in the series under study, neither of which we have. With all that it implies.
But, it seems, is "decoupling" the same thing?
...And did the track lead to a dead end?
I, at one time, was also fond of the Caterpillar... Obviously, the method is not suitable for predicting the price of currency instruments. It is based on the methods of detecting the cyclic component and deterministic trends in the series under study, neither of which we have. With all that it implies. But the discontinuity seems to be the thing, doesn't it?
Not that it's a complete dead end. You participants in this thread seem to be beginning to arrive at the same thing, but from the other side.
Yes, there isn't stationarity and cyclicity in the whole series (except the one that says "the price comes back anyway"), but in some segments both stationarity (in the form of some general direction) and cyclicity (more or less visible in the price movement inside the "channel") are present. In terms of the discontinuity problem, these are the segments where a "stable" process changes its characteristics to other but "stable" ones. In the terms of money changers and speculators, these are not necessarily "trend" sections, they are also "flat" (the same trend but with zero shift) and even sections of changing "trends" they also have some "stable characteristics". There is only one question: to find those characteristics that adequately describe the state of "stability". On one hand these characteristics should be stable to "noise" (smaller and not so interesting price movements), on the other hand they should timely warn about changes of trends. Well, this is something like this, if we try to put the ideas into the categories that have already been mentioned here.
As a practical idea, we can see how the caterpillar behaves in the selected areas, we also selected trends using LR, so we can look at these areas. If there is a track algorithm. (I don't have it anymore, it's lost).
Sometimes, I die of laughter when confronted with the remarks of militant flooders! It's like a circus, they should be in kindergarten for maths, but no, they're too old! Generals, man.
And I was beginning to wonder if I'm the only one who likes it. After all, links like North Wind has long been given, but no reaction. And people there are really gloomy, ok illiteracy, it is always and everywhere militant. But this is a complete lack of culture of thought! After all, even a layman in general, at the level of ideas, can understand these simple and clear publications and learn a lot. But no! Everyone considers it his duty to show off for no reason.
Honestly, you're the first one who has been able to say/show something intelligible about the difference between timeframes and tick frames! Before that there were just statements by many along the lines of "give me ticks as the market is totally different on ticks and it's obvious". To my requests for more information on this issue I got "it's all over the web" or something like that. Well now you have simply provided actual evidence of this difference. Thank you!
I was interested in one phrase of yours from this post
http://forum.fxclub.org/showpost.php?p=594864&postcount=52
If it's not a secret could you share some information about some non-linear transformations that take into account the density of the tick distribution over time? By this you mean that having statistical data on distribution of ticks by time (just at a certain hour of the day) we could thus somehow transform the standard quote data of timeframes to a new type, which would bend this type of data distribution to a normal distribution or something else? It would be interesting to hear your ideas, if it is not already a trade secret.
Yurixx, unfortunately this is a common misfortune in our country. You start to understand it as soon as you go abroad. I think it all starts with pissing boys in lifts and sprouts up in almost all areas of life. I haven't been too many places, of course, but I think you only see this in Russia! There's something here that simply lies at the very foundation of society itself. But what it is and how to eradicate it - it is absolutely unclear. Probably in Russia it will only go away together with society itself :o(.
Also, I like the forum engine there better (although not ideal either), here it is too ascetic.
...
If it is not a secret, could you share some information about some non-linear transformations which take into account the density of ticks distribution in time? By this you mean that having statistical data about time distribution of ticks (just at a concrete hour of the day) we could thus somehow transform the standard quote data of timeframes to a new type, which would be able to fit the data into a normal distribution or something else? It would be interesting to hear your ideas, if it is not already a trade secret.
Maybe not directly, but I will try to answer your question. Two tasks. Just converting one distribution to the other. In principle this is a pretty well-known thing, and the methods are not new. In the same Excel, you can use standard tools to convert uniformly distributed data to normally distributed data, etc. This is about the distribution of data values directly, but you can also look at the rate of ticks over time in terms of distribution. Data speeds up and slow down depending on time of day, but if these speeds can be reduced to a familiar distribution (i.e., normal), then it's probably possible to judge something... In general, in this case, a series of prices transforms to the biparametric form in the form of distributions, and in essence, "time" ceases to exist, and accelerations and their derivatives become important. As a consequence, if suddenly a change in the distribution begins, we can probably talk about a breakdown of the previous process... etc.
And just now I've put together the first draft of my "astrolabe". Exactly the draft, because I haven't entered the whole logic of data processing into the system yet (4 modules out of 9), and of course, not everything is perfect yet, not all the criteria are optimal and there is a lot of searching and research ahead.
How does
work? Elliott theory, Hearst theory, potential function, potential and ... common sense. It is important to say that the system has no input parameters at all. None at all. Initially there is no predictable number of counts for which the system will make a prediction. It could be 7, 30 or 3,000 samples. I can't tell with what probability the system will make a prediction, it will choose the best one according to current data.
Right now I have only one parameter in MathCAD on historical data - the current bar (I use it to walk through the history). It is clear that it will not be available in the production version. From the reference selected as the current one, the system searches for a "trend" in history. This is one of the weaknesses so far. Well, it's looking somehow - it's the first iteration on the selection of historical data.
The system is "dissecting" the data memory, by the way, Sergey I admit, your Hurst score was better than mine. And another very important point, in terms of our current discussion: the calculation is done only on the clock, no minutes, let alone ticks.
We select several channels (together they may form a fan or a cone by their marginal rMS) for which the potential is calculated and then every channel "runs" a wave and not just a wave but a wave inheriting the fractality of the current data, as a result - the possible forecast values. An array of new prediction values forms new channels to be corrected by a "channel" wave of a higher order (no correction is taking place at the moment).
To clarify possible extrema (reversal zones), it is time to use parabolas just to the right of the red line based on the forecast data with possible data capturing from the left part. And here's where the optimization problem really arises due to the simple reason we get a matrix of forecast values - one count of x may correspond to several counts of y.
Prediction example
First tests and first results. The left part from the red vertical line is the data that has been selected for analysis. The red marks in the right part are forecasted price values in new channels, respectively.
Readout 6200
Readout 6205
Readout 6210
Forecast accuracy
You can see that the forecast is slightly wrong, though the new structure more or less repeats (for critics, "more or less" is the key word in this sentence). The general trend is downwards, which is exactly what has not been taken into account, i.e. the correction by the "high order channel" wave has not been done. And there must be not just a "linear" downward shift in the data, but taking into account the potential energy of the new channel, and the "binding force" of the new readings.
First results
Got what I expected. The system is not "wrong" (within accepted constraints of course) if "similar structures" of future movements exist in the selected historical data. If not, it lies, in the most brazen way, but sticks to the "party course".
PS
...
I have never tested my strategy on history, because I believe that
the future and the past are NEVER repeated.
Just like you cannot enter the same river twice... :)))
Alex, history, always repeats itself. You enter the same water in the river. Read Hirst, he will tell you, including about precipitation..., after all he wasn't a geophysicist for nothing.
Sergey, the dimension of volatility and spread should be the same. If it's meters - then meters, and if it's kilometres - then everything is in kilometres:-)
I use the "ideal" TS model in estimation calculations which comes to predicting only one parameter - the direction of the expected jump in price. The amplitude of this jump can be assumed equal to the volatility of an instrument in a selected timeframe or its standard deviation, which is almost the same. Taking into account that FAC can be interpreted as a relative value of prevalence of one type of price movement over the other (the opposite and counter-directional jumps), then we can state, without loss of accuracy, that the TS, based on the "ideal forecasting indicator", will NOT make a mistake when choosing the direction of the opening position, with the probability proportional to the absolute FAC value, underlying in the "ideal forecasting indicator". Profit or loss in pips from each trade is reasonable to estimate the value of the standard deviation of the instrument. Then the profit of TS on a sufficiently long time interval can be estimated as the difference of all successful trades and unsuccessful ones, each of which is multiplied by the volatility. Further, let's relate the obtained gross return to the number of executed trades and get the average estimation s of the "ideal" TS return per one trade:
s(TF)=Volatility(TF)*{(n+)-(n-)}/N=FAC(TF)*standard deviation(TF), where (n+) is the number of deals with positive balance, (n-) is the number of "negative" deals, N is total number of deals.
Which was required to prove.
PS: tell me the limits of volatility values for eurosd. I just don't calculate volatility at all. And right now I can't do such calculations.
If you can't estimate volatility, estimate standard deviation.) There will be no difference.
Sergey, thank you, he explained everything very clearly.
I only had the guts to elaborate on the smallest question, he was not a geophysicist, he was a hydrographer. He monitored water levels in open basins, hence his method of filling and emptying the reservoir. Well, that's just a small thing.