From theory to practice - page 501
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
Reread all of Orlov...(see attached file).
Basically - it all boils down to the fact that nothing in the market matters and nothing works, except two things:
The sample volume and the distance between the centres of the sample distributions over the time interval tau.
Essentially, he suggests waiting for price to move out of the variance in a particular sliding window and after a time interval << the size of the sliding window (!!! not like mine in the same window, but much smaller!!!), to close the trade using a counter-trend strategy. That's it.
Actually, correct. In other words, traders are waiting for a long candle, a large percentage of price change, not the speed of change, but the percentage change is important, there is a high probability that the rate will not go back. The ADX(14) indicator notices such things and the ZIGZAG appears. This could be the beginning of a trend. The start of another long candle on the higher timeframe. A significant deviation. And a narrow point in time for traders to make a decision on the different strategies being applied. The pies have been finished, we got high, we climbed into the wagons. The train has moved. Maybe.
Reread all of Orlov...(see attached file).
Basically - it all boils down to the fact that nothing in the market matters and nothing works, except two things:
The sample volume and the distance between the centres of the sample distributions over the time interval tau.
Essentially, he suggests waiting for price to move out of the variance in a particular sliding window and after a time interval << the size of the sliding window (!!! not like mine in the same window, but much smaller!!!), to close the trade using a counter-trend strategy. That's it.
Why, Alexander, has Fokker-Planck been put aside?
After all, the initial idea of converting the distribution that took place at the time of opening a trade into the distribution that will come later was so interesting. You're now only talking about ways of identifying when to enter a trade, and I think you've refined the Bollinger Band method a bit. And the problems arise in the question of when to close it... There's so much room for research there, and the mathematical apparatus is very unconventional - trajectory integration. Have you found anything?
Maybe at least in the formulation of the problem: what do you want from the distribution of increments to the end of the transaction for the purpose of profit? I remember that you talked about the extremum of the probability density function. If we know what is needed, we may look for those, again by formal methods of statistical analysis, for which there are many programs and there is no need to write them again. Then, you may be lucky, and we will be able to identify one or more typical prehistory.
P.S. By the way, I've found somewhere formulas for an optimal breakpoint (analog of the bridegroom problem) on a random walk time series. Seems to be Shiryaev's school. Look it up, where?
P.P.S. Have you become disillusioned with the methods of probability theory at all, isn't it time to analyse sequences rather than distributions?Why, Alexander, has Fokker-Planck been put aside?
After all, the initial idea of converting the distribution that took place at the time of opening a trade into the distribution that will come later was so interesting. You're now only talking about ways of identifying when to enter a trade, and I think you've refined the Bollinger Band method a bit. And the problems arise in the question of when to close it... There's so much room for research there, and the mathematical apparatus is very unconventional - trajectory integration. Didn't you find anything?
Why not - no, I haven't.
In fact, the working ones are:
1. Calculating dispersion (diffusion coefficient) via "the root of time"
2. using medians.
Have statistical advantage over counterparts:
3. exponential time scale
4. ACF
Unworkable really is the moment of exiting a trade...
If you paid attention to my last hypergeometric histograms, they directly indicate (and Orlov confirms it) that the time of decision to exit a trade should be less than the dimension of the observation time window. In fact, it is a short-term prediction of a "return to the mean".
Study, calculate, distribute
the increments
without understanding the physical meaning of the increments
is handwashing and brainwashing
Study, calculate, distribute
the increments
without understanding the physical meaning of the increments
is handwashing and brainwashing
P.S. Yes, by the way, I met somewhere formulas for an optimal stopping point (analogous to the picky bride problem) on a time series with random walk. Seems to be Shiryaev's school. Look it up, where?
P.P.S. Have you become disillusioned with the methods of probability theory at all, isn't it time to analyse sequences rather than distributions?Shiryaev speaks about a particular case of optimal stopping - the problem of random process decay. In this thread I posted a link to a video with his story about it (and that he published a book on the subject last year).
TC has a problem with the basics of matstat. For example, he talks about the same distribution of two samples, without confirming it with agreement/uniformity criteria.
Vladimir, if one day you have a desire to work in the VisSim system, I can send you my TS model - you just need to write one phrase: "Send it" :)) Let it be a draft, giving 0% profit (in fact, the basic model) - but, it is easier to finalize their ideas, than do it all over again. IMHO.
...
TC has a problem with the basics of matstat. For example, he talks about the same distribution of two samples, without confirming this with the agreement/uniformity criteria.
But the matstat also has a problem with the basics for forex rate series, these series do not have the property of statistical stability (relative frequencies tendency to probabilities), why the laws of large numbers are not fulfilled. What are the criteria?