From theory to practice - page 687
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 I found a new way of calculating process variance, simply - the width of the channel around the mean.
The genius and simplicity of the formula there is that you don't have to think about quantiles at all. Everything is calculated by itself.
I am still testing it out.
I've got it all counting by itself too, and no TV at all.
And everything works great. There is no calculation period even everything works by itself.
Here, you're welcome to use it. I haven't seen anywhere better for trading and probably never will.
It really works in practice for me.
Unlike theory.
All you need is :
a) start of bar calculation on the current chart
b) analyzed timeframe
The further from the current bar the start of calculation, the higher the probability of the maximum adaptation to the price movement. Adaptation itself "adjusts" itself need only to take the beginning of calculation away from the current bar more than 1000 bars
everything.There are a lot of talented people in this thread, so why don't we move on to practice at last?) And use the chance to use collective intelligence?
Because I don't think, I'm even 99% sure, that you'll find something dramatically better than this algorithm in the next 100 pages.
I could make it multi-timeframe, so that it shows all the latest levels of the last channel on all TFs.
But I already did it...it was of little use to the robot. That is why I simply threw the idea away and did not use it again.
Thanks, buddy! But, I personally don't need ready-made solutions (like AUTOMATIC_CHANNEL.ex4). I need ideas, or a research task.
Most of your suggestions relate to MM. I am interested solely and only in the physics and mathematics of the market.
You may post in this thread the task: what to investigate and with what purpose, and I, if I'm able to do it, of course, will help.
In the process of fighting the truth, delusion exposes itself...
I see... Okay.
I sincerely wish you something someday... I can only wish you blind luck in your search on such a dangerous road as you have chosen...
In the process of wrestling with the truth, delusion exposes itself...
I see... Okay.
I sincerely wish you something someday... I can only wish you blind luck in your search on such a dangerous road as you've chosen...
Be that as it may. Amen.
I also have everything counts by itself and without a TV at all.
And everything works just fine. There's not even a calculation period and everything works by itself.
Here, you're welcome to use it. I haven't seen anywhere better for trading and probably never will.
It really works in practice for me.
Unlike theory.
All you need is :
a) start of bar calculation on the current chart
b) analyzed timeframe
The further from the current bar the start of calculation, the higher the probability of the maximum adaptation to the price movement. Adaptation "settles" itself, we only need to take the beginning of calculation away from the current bar more than 1000 bars
everything.Well, yes. It is clear - the later we start to understand the process, the earlier we understand it.
There is no time in your formula
And that's the right thing to do. Everyone has heard of kagi, but you can also trade just by time, the price will be hidden.
Okay. More seriously.
Investigated the Wiener and Ornstein-Uhlenbeck process models for their relevance to the market for the "return to average" strategy.
The result on the real at the moment is "-3%" profit for 7 months of trading (about 100 trades).
Reason:
- Lack of market antipersistence parameter. Hurst coefficients, skewness and kurtosis of distributions are not working.
Now in operation:
1. Processhttps://en.wikipedia.org/wiki/Variance_gamma_process.
2. Gann theory.
To look at the chart over the last period and determine whether it is trending or flat?
Well, look at the graph with your own eyes and try to determine whether it is trending or flat. then transfer your understanding to machine code.
For the five hundredth time, I am publishing the Grail:
The process variance makes sense to me.
sigma^2 is the usual variance of the sliding window increment distribution
theta^2 is an unusual variance, namely = 2*(b^2), where
nu is the order of gamma distribution, and if we are talking about the Laplace distribution, nu=1.
But expectation, may the thunder strike me, is not clear to me...
I reread the correspondence between Automat and Vladimir - options, saturation functions... Fainted and fell asleep...
I tried to build a variance channel relative to the MA and the median, the results improved by about +10%, but that's not it... Wrong, as it were...
Continuing to dumb down...
it's just more nonsense, i.e. it's a toy for fun, but not a tool for work.
To read:
https://elis.psu.ru/node/337977
same thing -- bullshit.
It is not a tool for market work.