Econometrics: let's discuss the CU balance sheet. - page 28
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
Avals, a question for you too, you seem to understand something about the subject from your posts, but then what are you doing here (in this thread), are you really interested?
and what are a lot of people doing in the https://www.mql5.com/ru/forum/7355 thread ? So as not to forget elementary things and not to get brain-dead))))
That's right, that's right. But you can "fantasise" a little further. For example, consider the hypothesis: "The measure of the dependence of the assets may be the difference of the distribution of their ratio increments from the Cauchy distribution".
But we won't do it in this thread. :)
Here, if we were dealing with the ratio of normal distributions, then yes - Cauchy. Because we have a fraction HP1/HP2 and when the denominator takes values close to zero, the result of division tends to infinity. Hence Cauchy with its uncertain variance and MO. In crosses (and in other instruments) we deal with increments. That is, even if these increments were distributed normally and independently, we would deal with (X+NP1)/(Y+NP2). And naturally we would not get a killer zero in the denominator and Cauchy would not add up. I.e. even with independence Cauchy would not be obtained, although the tails would be thicker.
Among the data provided byMetaDriver I found a curious specimen, quite large.
Here is the balance chart:
All in all a successful TS. Our task in the early stages is to recognise its quality.
We take the first plot of 171 observations. Here is the graph:
In this area of 131 observations TS is unprofitable and it should be rejected simply due to results.
Let us look at the statistical characteristics of the residual from the straight line fitting - this is the blue line in the graph.
For this plot the probability that this plot is not stationary = 3.67%, i.e. the plot is stationary
.
Let's take the next plot 131-417. here is the graph:
This is a profitable plot. For this plot the probability that the residual is not stationary is 8.1%.
Sad conclusion:
We cannot recognize the profitability of the CU in any way we know: neither by visual inspection, nor by analyzing the residual for stationarity.
If you have any ideas, I am willing to do the calculations. I now have plenty of varied stat material.
I have no ideas
If you have any ideas, I am willing to do the calculations. I now have plenty of varied stat material.
I have no ideas
Why is the balance and not the equity calculated? There are far fewer statistics in the form - it hides intra-dealer drawdowns.
Why is the balance and not the equity calculated? There are far fewer statistics in the form - it hides intra-dealer drawdowns.
And how do you get the raw data. These are derived from the championship. I had equity in pips.
And how do you get the raw data. These are derived from the championship. I had equity in pips.
It's a bit tricky on the report. I need to write a script that would export equities, not deals, according to the report and the appropriate symbol history to excel. Or maybe there is such a script.
Where does the "Sad conclusion" come from?
Where does the "Sad conclusion" come from?
Well, like why. The tester says nothing about the TS, the statistics are no better.
Well, like why. The tester doesn't say anything about TC, the stats aren't any better.
Well in principle yes. The lag is critical. You can use statistical methods to find the MO confidence interval, etc., but more or less interesting results will come from a very large volume of statistics.
I wrote a slightly different approach in a private message.
faa1947 I have written to you in private.