Random Flow Theory and FOREX - page 28

New comment
 

Prival, you have restored it wrong, that's all.

  1. Берем преобразование Close[i]-Close[i+1]
  2. Берем обратное действие Close[i]+Close[i+1]

What kind of amateurism is this, Mr. Military Radio Technician? Stand up, fall down and do 20 push-ups! What's the real point of point 2? Well, it's a good thing you're not adding kilograms to meters...

If returns(i) = delta * Close(i) (delta is a non-centred operator of the first difference), then Close(i) = Close(Bars-1) + Sum(returns(k), k = Bars-2..i) - or something similar. Returns - these are increments that need to be accumulated from the very beginning of history to restore Close. And add the initial value, Close(Bars-1), to everything that has been added. The reconstruction turns out complete and unambiguous, the formula ideologically identical to a simple taking of the first form.

P.S. Pardon my familiarity. But you also do not suffer from political correctness...

P.P.S. Well, thanks to Neutron, he already answered you.

 
Mathemat:

Prival, you recovered it wrong, that's all.

  1. Take the conversion Close[i]-Close[i+1]
  2. Take the reverse action Close[i]+Close[i+1]
then Close(i) = Close(Bars-1) + Sum(returns(k), k = Bars-2..i) - or something similar. Returns - these are increments that should be accumulated from the beginning of history to Close
Yes, the second action is written correctly (I have not written it down accurately), I seem to have done as you wrote, by your formula. Recheck my calculations. I need an area with a clear trend, try to reconstruct it.
 

Please!

Red is the original series, blue is the restored series, green is the linear trend.

 

Okay. Got it. Thank you. (chuckles) I'll be looking for a fault somewhere. I always thought it was retrievable, too. I wrote about it on the last page.

Prival 12.12.2007 22:24 PM corrected| delete

lna01:
Prival:

Candid I have a request if it is not difficult to check ACF pic.3, if the same, then acceleration check does not make sense and already BGS, if so the system is SRS will consist of 2 equations.

First question: why did you take returns for the original row and not for Y-mu?


You're right not to take returns in pure form :-(, it kills the trend. I can't restore the initial process back. I never gave it much thought, I thought it was always possible to go back to the exact constant.

Got an unexpected result, couldn't find the error. If I find the error and everything fits, I will be able to post the trajectory (synthetics) first more or less workable version by the evening.

 
Prival:
You're right not to take returns in their pure form :-(, it kills the trend. It's impossible to restore the original process back. I didn't even think about it, I thought it would always be possible to go back to the exact constant.

I got an unexpected result and could not find the error. If I find the error and everything fits, I'll be able to post the trajectory (synthetics) the first more or less working version by the evening.

I think you still have an error. Such an effect could also give insufficient accuracy of the calculations, but since Y-mu came together, it is unlikely. There should also be no mysticism :).
I meant with my question that the transition to another coordinate system(Y-mu) should suppress the parasitic and strengthen the dependencies we need.
 

Prival, are you serious - are you ready to lay out the synthetics yet (in the sense that I said)?

 

One more thing, I don't know exactly what you call it in your terms, like a public offer. In military terms, it's the word of an officer. I owe Mathemat a bottle of cognac.

And I owe him for putting my own brains in my place. Try hanging two balloons from a rubber band (let's say on the chandelier) and ask the kids to pull them. Now imagine what properties the weapon on 1 ball must have to be guaranteed to hit the second ball. Or even more complicated, there are weapons on both balloons and you must be guaranteed to defeat the enemy. Now move these balls at least 200 km away and naturally increase the amplitude of oscillation. A crude model, but this is what I have studied and what I have been doing for over 20 years.

So in these terms it is easier for me to analyse the currency curve, for me it is the trajectory of the opponent. But even that is not the main point. Many trading books say that we should not treat it as a game. We do not believe it (or rather do not fully realize it), and convinced that we are already good at working on a demo account, we switch from demo to real account and get results. They (the books) don't tell us how to treat it then. Now it is war for me, it is easier for me. This is the psychological stability I need, to switch from demo to real. Figuratively it's that I can get hurt now too, and it's not a demo anymore.

Thank you that many have contacted me in person, written letters trying to support me, thinking that something bad has happened. On the contrary, only good things have happened, I have learned what and how to do. I need a system to point weapons at the enemy's trajectory and take the most advantageous point of launching missiles.

Don't think that I am an evil and bloodthirsty military man. I am a kind-hearted man and I have not hurt a fly. I have simply been trained for many years, and I take it seriously, knowing one thing - the enemy will think 10 times before attacking, knowing who he is up against. I have tried to be a good enemy, even a very good one, so that he would think not 10 times, but at least 11 times. Now I will notplay by his rules with the market, but I will wage war against him, and this is my rules - my battlefield.

P.S. Enough lyrics, or I'll flood my own thread. We need to do business. Thanks again to all.

P.P.S. Neutron, Mathemat . I made a mistake. Made calculations in separate file for forward and reverse conversions, everything fits together. I take back my statement, although I just now realized that it is nonsense to claim that an integral of the derivative does not restore the original function to within a constant. I should look for an error in myself, I've already made 12 sheets in MathCade to get a trajectory model, but it still fails. Maybe that's just the error.

 
Mathemat:

Prival, are you serious - are you ready to lay out the synthetics yet (in the sense that I said)?


Yes. If I understand correctly. You are a noble knight, you need it (synthetics=trajectory model) for a noble purpose. I need it for my scope calibration :-). Only there will be questions to you, how to check analytically exactly adequacy of the model, and what do you mean by adequacy if the model is a system of stochastic diff. equations ?
 

We have some students here who are not our peers in statistics and random process theory, and they can give us some advice. We just wait and try to do something ourselves...

P.S. Well I'm interested in it too for calibrating the scope. I just want to have a tool that gives me confidence that it is not random, this calibration. So you want "7 out of 10" and I want "99% sure that 7 out of 10, say, within the next year".

 
Prival:

You can't treat it like a game. ...It's war for me now,


Yeah, well, you can't have a partner who's trying to take money away from you. And it's not good to try to take it away from your partner either.
1...212223242526272829303132333435...85
New comment