Why is the normal distribution not normal? - page 9
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
I agree with that! I'm working in that direction myself at the moment.
It's not like you were arguing with me. At least I haven't found anything in your posts that I don't agree with. It's just that my writing style seems to be such that the reader misses individual words and even sentences. Let's improve!
And here's one from a neighbouring thread:
You listen, you listen to the matho lovers and wait, wait, wait.
What do you expect him to do? Get yourself a MA and admire the MO. )))))))))
Probably, my style of writing is such that the reader misses some words or even sentences. Let's improve it!
A jerk :-)
Check it out:
The very top of the RPR probability density turns out to have a "thin structure"! You can see that there are no EURUSD minute bars with an amplitude of 1,2,4,5 and 8 pips. A mystery though...
Maybe Fibonacci? And you can also see significant failures at 12 and 16 points.
Can you comment on this picture in more detail? It suggests bad ideas about consortiums and other impure machine forces.
Colleagues, I can share a "piece" of my research, or rather a concept. I understood that this includes the generation of synthetics with characteristics close to the real series. So, "conceptually", I came to a conclusion (it was long ago) that such series can be formed relatively easily by product of random variables with ND, with some coefficients (how should I say it - in general case). The only subtlety is that the variance has to change according to some "trend" or otherwise "shape" or "formula" (as you like) for the time samples. Below, a very simple demonstration of the idea itself, i.e. the essence (without the final formulas, which still need to be checked).
A certain chunk of EURUSD, M15, (H+L)/2, 5000 samples
Trivial algorithm and roughly selected parameters for a "similar" plot:
(B - vector with the first element)
Synthetic series:
A are "gliStograms" of logarithms of the ratio of increments:
PS: I note only that the change of dispersion has a certain "wave" character
Sergei, give the same data in logarithmic scale on the vertical axis, because most of your experimental data are at "zero" and are not informative.
As for the method you mentioned for bringing the distribution closer to the market, I also did it. I even introduced a fractional value of the exponent (it was fun to do).
joo писал(а) >>
Can you comment on this picture in more detail? It leads to bad thoughts about consortiums and other impure machine forces.
Sergei, please give the same data in logarithmic scale on the vertical axis, because most of your experimental data is at "zero" and is not informative.
As for your method of approximating the distribution to that of the market, I've also done so. I even entered a fractional value of the exponent (it turned out cool).
Yeah, it's funny. I was having fun with Hearst's exponent.
I can cite as much evidence as I like to prove the existence of statistically significant dependencies between price increments. In other words, I argue that there is a relationship between the previous price movement of an instrument and its movement in the next step.
I'll bet you $5,000 that this statement is false.
Sergei, give the same data on a logarithmic scale on the vertical axis, because you have most of the experimental data at "zero" and are not informative.
As for the method you mentioned for approximating the distribution to the market, I did the same thing. I even entered a fractional value of exponent (which turned out to be fun).
In all seriousness, this is most likely a result of the digital filter of the particular DC. It's up to them to decide what constitutes "noise" etc.No.
I fell upon it when reviewing Hearst index spectra for HF signals. Then I looked at other markets and found it. Yes, it exists in those markets that are controlled.
This effect is not described anywhere. I suppose it is caused by the discreteness of volatility. And I draw conclusions: for Eurusd the most stable volatility is 10 min timeframe after large price movements; the only practical profit is in small stops when pipsing after the bump. However, pips are unprofitable and there is nothing to catch.
1999-2001 (no more than m15 in Excel)
Update. Excel 2007 allows you to work with 1 million lines! This removes most of the limitations of the software.