Ward 6 - page 29

 
Dr.Drain:
Thanks, for the adequacy :) I.e. as far as I understand it, even with TP=SL for a large number of trades the probability outweighs in favour of TP. Have you modelled this on a large number of trades, I mean over 100...200 trades?
 
Dr.Drain:

I have been working on this for some time and with a probability of well over 50% it is possible to predict changes in volatility without too much trouble. what the movement will be and approximately how big it will be. but the sign of this movement (up/down) is much harder to predict, it is a different order of difficulty.

There is no special need to predict the sign. We can always reverse by increasing the position volume, to compensate for the losses of the previous reversals.
The only question is how soon the trend will occur, how much we will lose on reversals and how long the price trend will be.
 
khorosh:

Everything will somehow average out and everything will be fine.)
I'm afraid to disappoint you, but that's how it's going to average out. Also, does it not occur to you that my analysis algorithm is such that I won't enter the market in real "long non-trends", which are especially fraught with the fears you describe?
 
DmitriyN:
There is no particular need for predicting a sign. You can always turn around by increasing the volume of the position
Martin. Such a martin. The result is known.
 
DmitriyN:
... have you modelled this on a large number of trades, I mean over 100...200 trades?
I "check" it right before your eyes on a demo account :-) Yes, apart from this demo visible to you there are "over 200" trades that also show a probability overweight. But all that wouldn't matter if there wasn't a kernel - the theoretical basis of the algorithm - due to which it follows purely from mathematical and logical construction that there should be an overweight. If there were no such thing, I myself would declare the preponderance on some hundreds or even thousands of trades to be a fluke.
 

I'd like to understand the meaning of the Swinosaurs filter using this example of his circuit.



I know how diodes, resistors and capacitors work.
Let's assume that we feed the Close price to the input, what do we then get at the output? Where is the mathematical and physical meaning of this circuit?

 
Dr.Drain:
I'm afraid to disappoint you, but on average it will be like that. Also, does it not occur to you that my analysis algorithm is such that I will not enter the market in real "long non-trends", which are particularly fraught with the fears you describe?
You are assuming that you know the time interval of a possible imminent non-suspending trend, because you are not going to enter this interval. Then you'd better open a new topic: "Predicting a no-reversal trend" It would be very interesting, especially for lovers of martingale using averaging.
 
DmitriyN:

Let's say we give the Close price as input, what then do we get as output? Where are the mathematical and physical implications of this scheme?

The physical one is directly drawn :-))) Well look, let ground has a potential of conventionally zero. The potential itself has no physical sense as we know, as it is defined to an arbitrary constant. So, ground is zero. Next, some potential, positive with respect to ground, is applied to the input. Current goes through the diode VD1. The capacitor is charged. You know the simple equation. So you understand that voltage across the capacitor has started to increase. That is the potential at the output. But - with some delay, of course. A transient process. Time constant for it is determined by R1 (because C1 is the same both for the case in question and for the case when negative potential is applied to the input). So, yes, negative potential to input is all the same, only now the time constant is R2*C1. And R2 is not equal to R1. All the same, it's obvious. The discharge process is made physically equivalent to the charge process. From the circuit it is obvious. If so, it's easier not to understand the µl code which author wrote, and just write equations of charge-current-voltages for this scheme on the piece of paper and calculate what will be the output. Write it as a piece of code yourself and go ahead, feed it to the input price chart. It is even easier for me than to understand author's calculations (which I didn't do and was initially happy with just this picture when I read it).
 
DmitriyN:

What then do we get out of it?

a curve with a non-linearly distributed lag with respect to the input signal. that is the interest. it is, you see, a non-linear filter. already a step forward from linear or linear filters with slowly changing characteristics. for such a filter there is already no AFC and FFC there. Already there is shamanism in trying to understand what is on the output in terms of linear filters (at least in the definition of "lag"), although the circuit is primitive.
 
khorosh:
You assume that you know the time interval of a possible imminent reverse trend because you are not going to enter at that interval.
I am not assuming. And I don't know. But I can see it in this geometry: price will not jerk strongly, but quickly go down without jerking - that is, my filter is down and price is down jerking weakly around it. And why should I enter the market? When such a pattern ends, then I will enter.