Digital ACSTrend - page 29
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
Check here:
Check here:
https://www.mql5.com/en/forum/175639/page2
...
Nice to see a comparison to a random walk for market but it has been already proven by Lo and MacKinlay that a lot of the markets are not a random walk, especially in spot forex. Spot foreign exchange have more trading opportunities (where market is inefficient) the more you scale down the time frame. Contrary the popular belief that longer periods are better for trading.
The absence of prof is not the prof of absence.
which picture is from a random time series?
They all represent a series with different distribution from arandom walkgenerator. John Last you should be careful to make conclusions from phase space trajectory. It is only a space in which all possible states of a system are represented - but how do you make that distribution is mostly significant.
...
"Никто не обнимет необъятного" - Козьма Прутков
Hi
You need also to plot a correlation dimension graph and to explore how it progresses with the embedding dimension.
"1. The time series is random when the correlation dimension remains roughly equal to the embedding dimension. There is no saturation of the correlation dimension with increasing embeddings. The returns follow a random walk process. In this case, neither short-term nor long term predictions of the series are possible
2. The time series is chaotic when the correlation dimension clearly saturates with the increasing embedding dimensions of the phase space. The phenomenon of chaos inherently implies that no long-term prediction is possible. However, it is possible to make short term predictions
3. The time series is correlated when the correlation dimension remains well below the value of the embedding dimension and keeps increasing with increasing embeddings without ever ever saturating. In this case it should be possible to make long term predictions."
Trading on the edge: neural, genetic, and fuzzy systems for chaotic financial markets.Guido Deboeck p.298
1/Largest Lyapunov exponent is our predictability horizon
Other thing to do is to estimate the Hurst exponent and you can make a double check with randomized buckets.
Anyway for the series in question the Hurst exponent was above 0.5 (iVAR is robust enough, accordingly below 0.5)
https://www.mql5.com/en/forum/178285/page6
How did you manage to make random walk with all negative Lyapunov exponents?
На вкус и цвет товарищей нет
...
Живу в Тамбовском лесу
Серый
Don't try to teach my Grandma to suck eggs.
dass der dicke Dachdecker die dicke Dame durch den dicken Dreck trug.
traders joking
Assumptions
Nice to see a comparison to a random walk for market but it has been already proven by Lo and MacKinlay that a lot of the markets are not a random walk, especially in spot forex. Spot foreign exchange have more trading opportunities (where market is inefficient) the more you scale down the time frame. Contrary the popular belief that longer periods are better for trading.
Please beware of scientific papers about the market. They can give a snapshot about a particular market but they cannot reveal the "nature of the market". Here Pava is very correct with its citation.
I use market papers as hints and ideas but everything has to be tested life.
A team helps a lot, and I see this thread as a team.
The reason is pretty obvious the markets are not a scientific phenomenon.
They are changing all the time. The market today is not not the market yesterday and the market tomorrow would be different.
An interesting feature of the market is that it wants to be technologically neutral, the market wants to give the same opportunity to a very advanced and to a very basic system (I know I have some crazy ideas and it is one of them).
What we want to do is to hack it, and only to participate when the market makes that the odds are in our favor, and that he would do from time to time, and inversely not to bet when the odds are not in our favor. Beating the market anytime, anywhere is not possible, at least at my retail level.
You can't prove anything, you can just observe and adapt.
By the way the quant strategies maybe historically have lost more than any other kind of trading strategies. And that is because their authors have proved something.
All our assumptions are probably incorrect.
Justfor fun
Yep
I agree: Adaptation is a proof of cleverness
Still, calculating the largest Lyapunov exponent, should be made by calculataing the Eigen values, on a small portion of datas, o/w ....
This largest_Lyapunov_Exponent and the calculation of the correlation are our goals
Just for fun, I put this Hurst_Differnce_Total indicator, showing this value (from CLOSE, HIGH, LOW, and OPEN),
I tried to calculate the Hurst difference from the weighted price, but there is a weird difficulty, that I do not understand where it comes from
Regards