From theory to practice. Part 2 - page 26
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
Rather, we should talk about the difference between a trend and a return.)
If we start discussing the general theory, nothing good will come out again. This is a good case where there is already a formalisation of return and we need to supplement it with a suitable (for this particular case) formalisation of the trend. The obvious option is unlikely to be a reversal of conditions)
My robots include both clearly trending and conditionally counter-trending ones. On a long interval both of them earn.
I have a lot of money in my trading robot, but it is not possible to increase its volume due to large drawdowns. The obvious idea is to supplement the existing system with a trend system for diversification.
The top-starter is also sort of earning, but because of the large drawdown it is impossible to build up volume. The obvious idea is to supplement the existing system with a trend system for diversification.
There! Diversification is the right word. Currently I have 4 different basic algorithms trading on my real account, two of them are trend-following, one is conditionally counter-trending, and one is conditionally scalping.
There are several modifications with different parameters.
All of them are profitable on average, all of them have decent drawdowns.
The top-starter is also sort of earning, but because of the large drawdown it is impossible to build up volume. The obvious idea is to supplement the existing system with a trend system for diversification.
Or think of such a sign of price reversal against the existing position, so that when this sign appears, you can close the position with minimal losses.
If we start discussing the general theory, again nothing good will come out. This is a good case where there is already a return formalisation and we need to supplement it with a suitable (for this particular case) trend formalisation. The obvious option - reversal of conditions is unlikely to work)
Er, wait a minute. Where do we have"already have a return formal isation"?
The top-starter is also sort of earning, but because of the large drawdown it is impossible to build up volume. The obvious idea is to supplement the existing system with a trend system for diversification.
The profitability of the topicstarter, by the way, can be reduced with a light hand, but I think the public statement of methods is redundant).
Uh, wait a minute. Where is it that we"already have a formalisation of recoverability"?
Well, it would be more correct to say that there is a formal rule to determine the moments of time, which quite often (but not always) turn out to be close to the moments of price reversal.
The author's drawdowns, by the way, can be reduced with a light movement of the hand, but I think the public statement of methods is redundant).
I should have given it to some girl for the eighth of March)
Well, it would be more correct to say that there is a formal rule of thumb for determining the moments of time which quite often (but not always) turn out to be close to the price reversal moments.
Is that the formula with the root of the period? Why should it be believed?
It's not like the top starter has even identified the process. He simply postulated from his head that there is a return, and that is the formula.
And secondly, it is not clear, why to calculate volatility limits by the process formula, if it is more accurate and easier to determine them directly from the current implementation, by measurement?
All the more reason to do so with the middle one)Is that the formula with the root of the period? Why should it be believed?
The original author didn't even identify the process. He just postulated out of his head that there is a return, and that is the formula.
And secondly, it is not clear why to calculate volatility boundaries by the process formula, if it is more accurate and easier to determine them directly from the current implementation, by measurement?
Especially since this is what he does with the average)Generally speaking, there's a pretty simple matstat there, although I'm not sure I can articulate it clearly. It means testing of the null hypothesis that expectation of increments is equal to zero. The distribution for the null hypothesis has zero expectation and either Gaussian or Laplace form. This statistic is similar to the one used in the Student's tester, but instead of the roots of the sum of the squares of the increments, the sum of their moduli is used. The distribution of the statistic depends on the type of incremental distributions, or perhaps an asymptotic Gaussian distribution is used (if the sample size of the increments is large enough). Some p-value for rejecting the null hypothesis is chosen, which gives the "magic" value of the coefficient 2.578 in Shurik's formula.
If the null hypothesis is rejected, the initial model is abruptly forgotten and it is said that the average increments for the real prices must tend to zero and for that they must soon change their sign and therefore a position is opened to the corresponding side which is closed when the "equilibrium is restored".