Dependency statistics in quotes (information theory, correlation and other feature selection methods) - page 58

 
IgorM:

There was an online show on this forum a couple of years ago where a man went from 1000 to 50,000 in a month, and unlike in your case, he even made a couple of mistakes

Where is this guy now? How's he doing?
 
DmitriyN:
Where is this man now? How is he feeling?
Dima, isn't that what Gerczyk's been telling everyone? The 12-year-old unicum.
 
Mathemat:

Hehe... Question in reverse: why can you believe the ACF on a large number of bars?

And anyway - you don't get it. We are looking for any kind of information dependencies - not just linear dependencies, which only the ACF reveals.

What do you mean any? I'm only looking for the ones that can be eaten. I calculate the ACF. There is, it means something can be smoothed out. Successful, analytics to the piggy bank is a deterministic thing. More ACF to the remainder. What about you? I found that the current bar depends on 500, so what? What is it meaningful? And if we shift it? There are a lot of questions. After all, there are special algorithms that calculate the shift in lags. 10 is a large value. Dependencies greater than 50 lags are on the other side of good and evil.
 
Mathemat:

Double up. For every 1,000 deposits, 1 lot. And the growth will be exponential.

But this is kamikaze trading. Not convincing and looks like a cheap trick.

Of course it is. I was the one who reckoned that any movement made a profit. No losses.
 

Here's a screen with a consistent pattern repeatability. There are a lot of smart, savvy programmers here. Try to formalise it. I think you won't succeed.

The task is as follows.

Based on the first two movements of the impulse-correction calculate the probability of formation of the third part of the impulse.

That will be one move. On the screenshot, the run length and arrival time to the end point of the orange and green impulses.

 
IgorM:

Honestly, your friend is a wuss )))))

Here on the forum a couple of years ago was an online show, the man went from 1000 to 50,000 in a month, and unlike your case, he even made a couple of mistakes

Ah, guys!

My first Expert Advisor ran over 10.000 of deposit up to 6 million on M1 in 10 days! I threw it away, I did not know what value it had on the forum.

That's the trouble, that's the trouble.

 
Mathemat:

Double up. For every 1,000 deposits, 1 lot. And the growth will be exponential.

But that's kamikaze trading. It's not convincing and it sounds like a cheap trick.


Every one of us would not mind repeating this stunt.

And then, to increase the deposit from 50 bucks even once a year to 10000 is not so bad. Then divide it by 200 and drain the other three quarters at 50 each - just a dream. 10000\150==... that's a hell of a percentage.

I suggest we stop measuring the width of our loafers.

 
faa1947:
What do you mean any? I'm only looking for ones that can be eaten.

Again I don't get it - or are you holding on to your beliefs?

There's no algorithm laid out here for how to eat it. But that doesn't mean it can't be invented.

Dependencies over 50 lags are on the other side of good and evil.
The trick is that TAdv is on the other side of good and evil too, but it works, if you believe ... .
 
VNG:

Here's a screen with a consistent pattern repeatability. There are a lot of smart, savvy programmers here. Try to formalise it. I think you won't succeed.

The task is as follows.

Based on the first two movements of the impulse-correction calculate the probability of formation of the third part of the impulse.

That will be one move. On the screenshot, the run length and arrival time to the end point of the orange and green impulses.


well, if a lot of savvy ones don't formalise, then it's a gut feeling))
 
Avals:

Well, if a lot of smart people don't formalise, then it's a gut feeling))

What's the flair? What is the doubt? In the stability of the pattern. Then analyze the history yourself and you will see. There is no gut feeling here, the rules are ridiculously simple.
Reason: