From theory to practice - page 432
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
Maybe ticks at a certain reading interval will work, but who knows, we'll see Alexander's results.
He tested it last week and everything was fine, but on the minutes this week is also good, but on the long term it is worse.
Is it a grail, which has been described in 430 pages?
There are better signals in the trading books.
You know, like MAs and stuff like that....
I advise you to reread the thread. You've got the wrong code.I suggest you reread the thread. That's not what you put into the code.
Well, Alexander looked, as he said, so he did.
The formula has already been written here many times.
Here is the code. This is what Alexander has from the second graph. I will not publish what is on the third graph without Alexander's permission, as he has not written anything about it here.
I did not see a modulus before addition in your formulas (Alexander's formula had SUMM(ABS(return)), you have SUMM(return). In addition, I did not understand why the timeframe was a minute one, while Alexander's was about one second step.
Then it should still be called a process with memory. Unfortunately, this is the accepted terminology, although I don't like it either.
For processes with "partial" memory, it seems, simply there are no terms, because the academic science has not yet grasped such phenomena) By the way, precisely for this reason all "Nobel laureates" etc. -physicists-mathematicians and can't earn anything :)In random process theory, so-called memory is a consequence of stochastic dependence between process values (random variables) at different points in time. It does not make much difference how far apart these moments are. A random process is said to be given if all possible finite dimensional distributions are given and hence we can always calculate the dependencies between its values at any given time and easily see what and when it remembers. This is in theory. When it comes to practice (statistics on random processes) we always find that we have much less data than we need to reconstruct distributions. Therefore we always have to resort to all sorts of simplifications and assumptions, and avoid desirable but unsubstantiated conclusions. Besides any well working theory will sooner or later change the market in such a way that it becomes unprofitable (there are famous examples).
I think that laureate physicists, like other people, do not earn anything (or sometimes do) for roughly the same reasons :)
I didn't see your formulas taking a modulus before adding (Alexander had SUMM(ABS(return)), you have SUMM(return).
Sorry, I overlooked it.
But still... Did Alexander start working on minute steps? You do understand that Abs (Open(i)-Open(i+1)) with one minute increments is not at all equal to the sum of Abs(return) for 60 seconds included in a minute?
Sorry, I overlooked it.
But still... Did Alexander start working on the minutes?
He asked to check on the minutes. No need to read the thread from the last post, then everything will be clear.
The growth graph looks like a soundtrack doesn't it?
How does this track sound? Using some program found on the net I converted this picture to sound files Noise.wav and Sine.wav
I attach archive with files and program.
Note, before listening make speakers volume to minimum, otherwise I am not responsible for your health!
Forward Testing! True, I have launched it on a demo account.
Alexander gives good advice, but we have to implement it properly. The formula is the same, but the approach is different. All variables are dynamic.
P.S. It has nothing to do with my previous posts where converted price charts are concerned. I had not elaborated on that idea yet.