Tics: amplitude and delay distributions - page 5
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
2 Mathemat
My research has shown roughly the same thing.
Nevertheless, you can make money on ticks, but not in the market that is usually implied here.
yes
[edit] More precisely, where quotes depend on order flow
NorthernWind, I have a question for you. I remember in http://forum.fxclub.ru/showthread.php?t=32942 you hinted at the possibility of converting the distribution to Gaussian. Have you got any results in that direction? I'm not talking about ticks so much as bars.
P.S. That's where Prival will really unfold with confidence intervals...
NorthernWind, I have a question for you. I remember in http://forum.fxclub.ru/showthread.php?t=32942 you hinted at the possibility of converting the distribution to Gaussian. Have you got any results in that direction? I'm not talking about ticks so much as bars.
P.S. That's where Prival will really unfold with confidence intervals...
Thank you for your faith in me being so boundless :-). I am very much afraid of failing, because my knowledge is meager and my ignorance boundless :-(( There is some material somewhere (by Tikhonov, I think), how to have a random variable with its p.d.f, get another random variable, related to the first one, but having another p.d.f. If this is what you need I can look it up. There are even algorithms on how to do it.
Yes, I believe in you, Prival: your undisguised wariness is bound to change into a new quality someday. Regarding transformation of distributions: I did it when I was making a normal distribution in MQL4 from a uniformly distributed one. I needed a function that was the inverse of the integral function of the normal distribution. It turned out that I could not find an expression that would work decently, say, in the range of plus or minus 5-6 sigmas on the Internet. Strator's 'Probability Library' helped me out. In general, it would be interesting to look at some general algorithms...
P.S. Say we have a histogram of initial distribution (p.d.f. is unknown in analytical form). And we want to convert it to another with a given function (here - Gaussian). Find a numerical algorithm that produces a table of transformation function values.
Yes, I have faith in you, Prival: your undisguised wariness is bound to change into a new quality someday. Regarding transformation of distributions: I did it when I was making a normal distribution in MQL4 from a uniformly distributed one. I needed a function that was the inverse of the integral function of the normal distribution. It turned out that I could not find an expression that would work decently, say, in the range of plus or minus 5-6 sigmas on the Internet. Strator's 'Probability Library' helped me out. In general, it would be interesting to look at some general algorithms...
It would be desirable to post links here, and pick up literature on Rushen.
NorthernWind, I have a question for you. I remember in http://forum.fxclub.ru/showthread.php?t=32942 you hinted at the possibility of converting the distribution to Gaussian. Have you got any results in that direction? I'm not talking about ticks so much as bars.
P.S. That's where Prival will really unfold with confidence intervals...
The question of converting one distribution to another, besides its acadimical value, has practical application. As far as I understand, it became especially interesting in connection with the construction of pseudorandom number generators for coding and encryption. That's where we need to look for it. In principle, the simplest version of converting a uniformly distributed value into a normally distributed one is in Excel (as a function). But as far as I understand it's not what you need. Unfortunately I don't have time for it, therefore I will briefly tell you my results. For bars - I got nothing interesting, because of their heterogeneity in time. For ticks - there is a result, but it is inside spread and it is specific, not suitable for DC quotes.
SZY. And "plus or minus 5-6 sigmas" is very good, for practice you don't need more, as it seems to me. However, I did not study this question rigorously, so I will not vouch for it.
Sorry again that I can't satisfy your curiosity, I'm just really busy at the moment.