a trading strategy based on Elliott Wave Theory - page 262
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
... нам нужна альтернативная статистическая модель, которая имеет распределения с толстыми хвостами, проявляет персистентность и имеет нестабильные дисперсии.
Существует класс шумовых процессов, который соответствует этим критериям: 1/f или дробные шумы ...
...
1/f-шум тесно связан с процессами релаксации. Фактически 1/f-шум был постулирован Мандельбротом (Mandelbrot, 1982) как сумма большого количества параллельных процессов релаксации, происходящих на многих различных частотах.
Candid, can I get a link to this printed publication?
Also, such models exist and are excellent at modelling the behaviour of residuals in time series by the distribution function (fat-tailed distributions) and by the autocorrelation function (persistence). These are autoregressive models of infinite order. It is a great thing and it predicts the behavior of the simulated series very well but it has a limit in terms of the maximum yield - it barely covers the existing spreads. For example, if we keep the spread not larger than 1 point at EUR/GBP twenty-four hours a day, then the annual return of the AR-model will be 10 000 points! The same can be said about EUR/CHF (20000-30000 points a year). If spread at these pairs is 2 points, the annual return will fall to 200-400 points, if it is 3 points, we lose points. But for EUR/USD the border of profitability lies in the area of 0.5 points, i.e. unreal spread.
The book is called Fractal Analysis of Financial Markets by Peters. By the way, it was recently recommended by Rosh on the MQL forum. In it Peters also compares the fractal approach with AR-based models. As far as I understand it, he considers 1/f-noise to be the closest analogue to forex. As opposed to stock markets.
Basically, I had a non-linear model in mind from the start. Besides, there is no expectation that it will work in all market phases. Now, however, a new idea from "everything should be done the other way around" series has appeared, it is almost suitable to start coding, so it will be a priority for me for some time.
isn't it crowded in here in one thread?
I'm helping http://onix-trade.net/forum/ to administer...
we have an opportunity to create a separate forum on the mathematical approach due to the reorganisation of the categories...
i.e. i propose to create a separate forum for this discussion in the section dedicated to trading systems.
it will allow
- increase user-friendliness - the engine is functionally developed.
- increase the ease of publishing - you can create separate threads.
- easily upload any pictures and files.
- to be in charge of your own forum - additional rules/conveniences are up to you.
If the suggestion interests you, you can already start to write in
http://onix-trade.net/forum/index.php?showforum=67
(then I'll put it in its place)
(respected Metaquotes support, I think, will not mind to get rid of unnecessary work).
"MQL4: The Adverse Method solandr 06.04.2007 21:18
I just did it that way - I have answered the question asked in that thread. I have just provided a link to it here to keep it from getting lost in the website.
Candid, haven't you read Mandelbrot?
He writes that he was wrong about 1/f noise... (who doesn't get it right...)
Mandelbrot seems to have spent 40 years studying the market,
but judging by the reviews on amazon of his last book, (The Misbehavior of Markets)
hasn't been able to come up with anything complete yet.
i wonder... how people read Peters... i got the feeling that the book is a mess... (though maybe it's in my head.)
there's a lot of literature on the subject, just make sure the word "market" isn't in the title...
I just came across this link...
http://www.bearcave.com/misl/misl_tech/wavelets/hurst/index.html
I strongly believe that if the strategy is profitable, it will show a profit with any spread that exists in BC. Take a look at the results for pairs involving the Japanese Yen. That is, the mathematical expectation of the strategy's profit should be at least several times higher than the spread (as well as the resulting swap) paid for making (holding) trades. If it circles around the spread and slightly exceeds the spread (as well as the swap), then I believe the strategy is simply not viable for REALLY existing trading conditions. Looking at a strategy with zero (or normalized) spread is purely an academic interest, which is far removed from current realities.
And traffic fees are still collected only in Russia. The rest of the world has long been charging a fixed fee only for the width of the channel. This is why there are so many free file-sharing sites abroad, from which people download films and other large files.
Candid, haven't you read mandelbrot?
He writes that he was wrong about the 1/f noise... (who doesn't get it right...)
No I haven't. Can you quote me or give me a link to the text? It is important what exactly he considers a mistake. By the way, I note that further on in that quote it said that frequencies should be logarithmically equidistant. I haven't quoted its sequel not because of laziness or love to brevity :). The analogy is an analogy, but one should not identify the market with 1/f-noise beforehand.
mandelbrot seems to have spent 40 years studying the market,
but judging by the reviews on amazon of his last book, (The Misbehavior of Markets)
hasn't been able to offer anything complete yet.
Einstein, by the way, most part of his life was engaged in the unified field theory. And also it seems that he could not offer anything complete :)
i wonder... how people read Peters... i got the feeling that the book is a mess... (though of course, maybe it's in my head.)
In my opinion it is an overview book and in this quality I liked it.