How can I tell the difference between a FOREX chart and a PRNG? - page 7
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
As for the second point - all this is child's play. To emulate volatility is elementary, it is enough to take the tick volumes of a real instrument and generate a random walk on their basis. There will appear thick tails and a volatility spike during the American session, and other other effects. But the SB will remain random and it will still be detected.
Disagree. If you take into account all the statistical patterns in price and generate a random series based on them, it will be impossible to tell the difference.
But even in this case, it must be said, the distribution on large TFs should tend to be normal. If, of course, the generator is of good quality.
It's not quite clear what "large TF" means in the context of an MF oscillator? There is no concept of time as such in oscillators. There is no difference whether we are looking at 100 closing prices of "minute" or 100 closing prices of "day".
I mean ternary. Who cares, the main thing is that the distribution is known, and the method described is universal anyway.
Why normal and not uniform?
And at the same time, if I understand it correctly, it can only work on a very large sample. If you take 1000 observations like I did, it's impossible to tell the difference.
I disagree. If you take into account all the statistical price patterns and generate a random series based on them, it would be impossible to tell the difference.
On the second point, it's all childish tricks. It is elementary to emulate volatility, just take the tick volumes of a real instrument and generate a random walk based on them. There will appear thick tails and a volatility spike during the American session, and other other effects. But the SB will remain random and it will still be detected.
Understanding such elementary stuff, I of course implied that you would be generating hourly bars at least with clustered volatility using either (and) some a la Garch or (and) real volume.
Understand, the type of distribution does not determine whether a series is random or not. It's just that non-random market series are not normal, and our primitive generators generate normal distributions in the simplest case. But to hypothesise on this that all non-normal series are markets and normal ones are random walks is flawed, since random walks can also be non-normal.
Yes, but all statistical patterns of price are somewhat more than all statistical patterns of volatility.
Of goats. The more regularities we have accounted for, the harder it is to distinguish the generated series from the real one.
The practical output here is the reverse process: once we have a generated series which we cannot distinguish from the real one, we can already assume that we know a good chunk of the real patterns. And hence we can try to exploit them.
How? A method?
The practical output here is the reverse process: once we have a generated series that we cannot distinguish from the real one, we can already assume that we know a good chunk of the real patterns. And so we can try to exploit them.
OK, let's say we generate a series with an ACF identical to the real one. What next? Can we make money on the ACF of the real market? I tried it - even without a commission it failed. So the question is - what is the power of this knowledge? We cannot tell the difference between the SB and the market by this indication, but we still cannot make money.
It's difficult to explain in two words. Besides, if I were you, I wouldn't trust me. That's why I suggest you check.
You're beautiful! You've been rubbing it in for so long, you don't know the method or .... but you're rubbing it in. You're convincing me of something. At the end you write, "Besides, if I were you, I wouldn't trust me."
Even a blonde would envy that logic.
There was no method and there is no method at all. And the author was asking about it, about "how to calculate", not about the philosophical "can't you see".