Why is the normal distribution not normal? - page 14
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
Yeah, make a button and hand it out to everyone.
>> I second that. And who can't get enough of one on request to give out Fto:)On the subject of GARCH to simulate the volatility of stock indices
That's right, as I wrote about, the volatility in the model is the RMS or variance, not the H-L, (which is a completely different series). But the correlation of all this with the increments in different combinations is very small.
(pensive) Where's the "bath time" button?
Sometimes the moderators come in on their own.
That's right, as I wrote about, the volatility in the model is the RMS or variance, not the H-L, (which is a very different series). But the correlation of all this with the increments in different combinations is very small.
On the nuances, it's too lazy to argue, and most people aren't interested. Large increments are followed by large ones with some attenuation. There is a dependence on the absolute value of increments. For FX it is necessary to analyse not lower than a day, because volatility is very periodic and depends on the time of day. More can be discussed, but not in this thread.
Sometimes the moderators come in on their own.
Thanks for the visit. >>)) It's just that it would be easier to pay attention with a pushbutton. No one would bully with it - he could get into the sauna himself, and you would have less work to do - sniff the forums.
It is too lazy to argue about the nuances, and most people are not interested. Large increments are followed by large increments with some attenuation. There is a dependence on the absolute value of the increments. For FX it is necessary to analyse not lower than a day, because volatility is very periodic and depends on the time of day. More can be discussed, but not in this thread.
Ok
Plot the tangent of the slope of the straight line drawn using the method of least squares through the cloud of data where the increments of the price series are plotted along the axes. If there is no dependence between increments, the line lies horizontally and its slope angle tangent equals (nearly) 0. If the dependence is 100%, then it is 1 (or 45 degrees of slope). If the dependence is counter-directional, the sign of the coefficient is negative.
Let's construct the RRR for minutes and find the correlation coefficient (CC). For the EURCHF series it is -0.4. There is a dependence (of some kind). Let's carry out the same for two-minute bars. Then for 3-minute bars and so on. We obtain the following dependence:
You can see at once that there is dependence between adjacent samples on all TFs up to TF=500 min. and this dependence decreases with increasing TF.
Now let's talk about statistical significance of the obtained result. What is statistical significance is available here.
That's all for now - my wife is driving me to bed!
If you believe your figure, the larger the timeframe, the higher the Correlation.
There are even positive correlations on larger...?
It is too lazy to argue about the nuances, and most people are not interested. Large increments are followed by large increments with some attenuation. There is a dependence on the absolute value of the increments. For FX it is necessary to analyse no lower than a day, because volatility is very periodic and depends on the time of day. More can be discussed, but not in this thread.
It's all clear. For those who are not on the subject you can read the respected Wikipedia https://ru.wikipedia.org/wiki/Волна. It's all laid out in a nutshell. There's just literally nothing more to add. Or maybe there is? I'm not a physicist or a mathematician, I'm an artist:))
Eh, I blabbed:((((
Well the cumulative is of course indistinguishable visually, but lascu.roman is right the change in volatility is missing.
and on a larger scale.
:)
Let's start with the fact that I lied a little bit:
The link is actually alternating. Let me explain.
Let us assume that we consider forex arbitrage. For EURUSD, for example, arbitrage loops of the "first generation" = EURGBP*GBPUSD, EURJPY*JPYUSD, EURCHF*CHFUSD, etc. form really negative feedback chains. The second generation for EURUSD, it is EURGBP*GBPJPY*JPYUSD, EURJPY*JPYCHF*CHFUSD, EURAUD*AUDGBP*GBPUSD etc. They form chains of positive feedback.
The number of chains in this case increases from generation to generation. For a limited currency basket it is not difficult to calculate their number and take them as coefficients,
But in practice the basket is unlimited, though obviously not infinite, i.e. the coefficients will have to be somehow shamanized (which is not bad, let good shamans prosper). :)
Next... you have to keep thinking genially and experiment a bit.... :))
zy. I'm sure the same principle remains valid for funds, despite the lack of such explicit arbitrage relationships in it.
zzy. I modeled a similar scheme for only three generations and obtained series of "quotations" with distinctly seen Elliott Waves. That indicates the presence of dough in the model. ;)
How's that?