Econometrics: let's discuss the CU balance sheet. - page 4

 
Demi:


I understood all of that.

Did you understand my question - why detrending? to analyse the results of the TC for a certain period? There is balance sheet chart - WHY do we do "detrending"????????????? What for?


To get statistics. That's what the topic is about. Trying to assess the balance through statistics.
 
faa1947:

To get statistics. That's what the topic is about. Trying to assess the balance through statistics.

and what kind of statistics can a balance curve after detrending give about TC?
 
Demi:

and what statistics about TC can the balance curve after detrending give?

Read from the beginning of the thread. Everything is given in graphs and figures.
 

I know only two methods of evaluating the performance of a TS over a certain period - capital gains (a priori above the "risk-free" rate) and risk-adjusted capital gains

 
Demi:

I know only two methods of evaluating the performance of a TS over a certain period - capital gains (a priori above the "risk-free" rate) and risk-adjusted capital gains


Well, there are more, in the tester for example. I'm setting a limited objective. Formulated it in the beginning. I don't think anyone has looked at the account balance that way. Might be something interesting.
 
faa1947:

Well, there are more, in the tester for example. I'm setting a limited objective. Formulated it in the beginning. I don't think anyone has looked at the account balance that way. Might be something interesting. I'll figure out why the program doesn't work on a large sample and continue.


Do you realise that you are not drawing this conclusion from the time series statistics, but from the analysis of the balance curve? - "Trading decisions can be made when there is a 14 pips deviation from the smoothing".

So you make trades first, then you get the balance series and smoothing and then you make a trading decision??? You make a trading decision in relation to the balance curve?????? You trade it???????

 
Demi:


Do you realise that you are not drawing this conclusion from the time series statistics, but from the analysis of the balance curve? - "Trading decisions can be made when there is a 14 pips deviation from the smoothing".

So you do the trades first, then you get a series of balances and smoothing and then you make a trading decision???


Yes, no. I type in the stats and set the parameters of the TS. TR and SL are pretty obvious, but 14 pips is not. There's just a number and how to use it is the question.
 
That's all for today.
 
Demi:


That's all I got.

Did you understand my question - why do we need "detrending" to analyse the results of the TS over a certain period? There is balance sheet chart - WHY make "de-trending"????????????? What for?


By analysing the residuals after detrending, faa actually checks that there is a trend in the balance curve. That is, that it is not a stochastic trend (i.e. random as in SB for example), but a deterministic one. Only I don't quite understand why some kind of complex detrending is used rather than linear detrending. After all, the basic model is the preservation of MO, which is the linear trend model. I.e. the ideal balance is a straight line at the angle of the corresponding MO. Econometricians, in order to check the model for consistency, analyse the residuals (or errors as they are called) - i.e. the difference between the model value and the real value. Only if this residual has a normal distribution with mo = 0, the hypothesis passes that our model fits. I.e. there is a deterministic trend in the balance - stable mo. And in general, the robust system equity model is a SB model with positive drift.

Most do this analysis more simply. Some people just visually look for the balance curve to be "good" and everyone likes the balance curve as a straight line into the sky)) They look to avoid deep drawdowns, which in balance analysis usually results in a "heavy tail" in the negative zone, which is a sign of lack of normality in balances. That is, this econometric analysis can be replaced by simpler expert methods and indicators such as PF, FS and maxdd. But their application depends on some nuances and is strictly unformalized, which is not suitable for theoretical sciences.

 
Avals:


By analyzing residuals after detrending faa actually checks that there is a trend in the balance curve. That is, that it is not a stochastic trend (i.e. random as in SB for example), but a deterministic one. Only I don't quite understand why some kind of complex detrending is used rather than linear detrending. After all, the basic model is the preservation of MO, which is the linear trend model. I.e. the ideal balance is a straight line at the angle of the corresponding MO. Econometricians, in order to check the model for consistency, analyse the residuals (or errors as they are called) - i.e. the difference between the model value and the real value. Only if this residual has a normal distribution with mo = 0, the hypothesis passes that our model fits. I.e. there is a deterministic trend in the balance - stable mo. And in general, the robust equilibrium model is the SB model with positive bias.

Most do this analysis more simply. Some people just visually look for the balance curve to be "good" and everyone likes the balance curve to be straight to the sky)). They look to avoid deep drawdowns, which in balance analysis usually results in a "heavy tail" in the negative zone, which is a sign of lack of normality in balances. That is, this econometric analysis can be replaced by simpler expert methods and indicators such as PF, FS and maxdd. But their application depends on some nuances and is strictly unformalized, which is not suitable for theoretical sciences.


Guys, seriously, come on! Stop mocking statistics and theorists! Stop it! It's not even funny anymore!

Well there is a globally accepted classic of all time - the deterministic and stochastic component of the series.

Well, what "deterministic" and "stochastic" trends? Where is this nonsense coming from?

If you like calling, for example, a straight line a "deterministic trend", fine. But "stochastic trend" is already "hard soft" ......

SB = "stochastic trend" - this is nonsense ......