Econometrics: one step ahead forecast - page 127
![MQL5 - Language of trade strategies built-in the MetaTrader 5 client terminal](https://c.mql5.com/i/registerlandings/logo-2.png)
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
No doubt about it. It's called state space. There was a guy here, he made some promises, and no promises. Willing to cooperate on this subject with everyone. So far I only know the dollar index from the state space. Attempts to include stock exchange indices have failed.
I told you a long time ago - I won't give you any explanatory lectures - just so you don't shit all over me as an ad man, you smart-ass...
If you want to understand state-space, open a book and don't waste your breath.
to faa
Неверно. T-статистика = коэф/СКО
Well yes, got a bit confused from memory with the concordant stats, it doesn't matter (always a correction), I've been working with it for a long time. But then your cases are even worse. That statistic says that 80%(!!!!) of your odds are untenable.
Exactly the first one describes it. We need 100 / t-statistics and get the % error. But that doesn't remove the problem with the other coefficients. no trend. HP is smoothing to get noise in the residual.
I have a growing understanding that you don't know what EW is doing. Either you don't know the functionality itself, or the particular implementation of the model that HP uses. You see, if your coefficient is 280 (!!!!!!!!!!), it means that some kind of "trend" (in quotes) model is used, most likely some kind of polynomial. And in this case everything you do in EW has no practical meaning.
It's supposed to be correct. DW is about two, which means that the residual is normally distributed. There is also regression error = 11 pips, but the error of the dependent variable = 212 pips
No not correct (!!! and you're not listening, I won't explain anymore), how do you reconcile this in your head between two errors of 11 pps and 200 pps????. The remainder of what is normal, the HP forecast, but then it doesn't make any sense. The remainder of the price (forecast) cannot be normal. It is guaranteed to be so. Most likely you have "drawn" a polynomial and EW is not showing you the forecast, but a fragment of the identification (in this case the fit) on the local plot.
Please note that the average error % = 5.7%!!!!
As Carlson said to the baby, "WAKE UP!!!!" What 5 percent error?????? Don't you have any idea what you're writing? For such a percentage in 100 counts you should get a Schnobel Prize!!! Even two. And this is not a joke or mockery.
PS: I looked at your coefficients, they're random. Somehow I'm losing interest in your astrolabe. Ok, I'll put EW on, see the functionality in more detail. There's a lot of confusion in your conclusions.
to faa
Well yes, got a bit confused from memory with the concordant stats, it doesn't matter (always a correction), I've been working with it for a long time. But then your cases are even worse. That statistic says that 80%(!!!!) of your ratios are untenable.
I have a growing understanding that you don't know what EW does. Either you don't know the functionality itself, or you don't know the specific implementation of the model that HP uses. You see, if your coefficient is 280 (!!!!!!!!!!), it means that some kind of "trend" (in quotes) model is used, most likely some kind of polynomial. And in this case, everything you do in EW has no practical meaning.
No not correct (!!! and you're not listening, I won't explain it anymore), how do you do it in your head to correlate two errors of 11 pp and 200 pp????. The remainder of what is normal, the HP forecast, but then it doesn't make any sense. The remainder of the price (forecast) cannot be normal. It is guaranteed to be so. Most likely you have "drawn in" a polynomial and EW is not showing you the forecast, but a fragment of identification (in this case the fit) on the local plot.
As Carlson said to the baby, "WAKE UP!!!!" What a 5 per cent error?????? Don't you have any idea what you are writing? You should get a Schnobel Prize for such a percentage in 100 counts!!! Even two. And this is not a joke or mockery.
PS: I looked at your coefficients, they're random. Somehow I'm losing interest in your astrolabe. Ok, I'll put EW on, see the functionality in more detail. There's a lot of confusion in your conclusions.
Here are the statistics for the residual from the equation
According to Jarque Berg, it is impossible to reject the hypothesis that the residual is normally distributed! And the probability value is very large.
Therefore almost all the numbers can be trusted.
to faa
Ok, I'll put EW on, I'll look at the functionality in more detail. There's a lot of confusion in your conclusions.Here are the statistics for the residual from the equation
According to Jarque Berg, it is impossible to reject the hypothesis that the residual is normally distributed! And the probability value is very large.
So you can trust almost all the numbers.
no you can't!!!! A normal (professional) system and adequate statistics (man, who the hell is Jacques, where have you gone and why aren't you using the really proven statistics) should give you the conclusion from your GLISTOOGram that it is impossible to unambiguously determine belonging to a distribution. That's for starters.
I can throw in a program on EW that calculates everything, there's quite a lot there, not just parameter estimation.
Here are the results on H4
The value of the coefficient is even worse. In addition to this, it is not possible to reject the hypothesis of zero coefficients for a series of coefficients.
The residual from the equation had ARCH, which was simulated.
The descriptive statistics of the residual are killer - nothing can be trusted.