Econometrics: State-space model forecasting - page 25
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
I missed it with all the discussion - so what's the average transaction size in pips?
1. Could you give me the type of autoregression, the function on the basis of which the prediction is made.
2. I have to use up to 1000 bars of history, so cases >100 bars cannot be excluded. I should consider cases > 1000 bars but for some reason my Expert Advisor ignores these cases even though the indicator may display even 10000 bars. What is the reason in the Expert Advisor, I do not know. I cannot find the 1000-bar limit in the code. Perhaps this is a system constraint?
I don't have a regression. A state space model is used. See above answer Mathematics. Above I gave a general view of the model.
I distinguish between the size of the window over which the parameters are calculated and the sample size over which the result is calculated, which is given above in pips excluding the spread. We see a fairly smooth balance sheet growth line. Maybe I'm wrong, but for me it's very important to have a smooth balance line.
Dear Yusuf!
I, to my shame, have not been able to understand your model - my knowledge is very limited to university curriculum and we have not read anything like that.
At the same time, that you use gamma function (and gamma distribution?) is very interesting as these functions are widely used in economics.
Here I took the ZZ reversal distances in bars and got the following histogram
Very similar to the gamma distribution.
Balance = 0.1780, i.e. 1780 4-digit pips.
Judging by the last picture here, there are about 1000 trades. Therefore it is less than 2 pips.
Balance = 0.1780, i.e. 1780 4-digit points.
The last picture here shows about 1000 trades. Accordingly, it is less than 2 pips.
Balance = 0.1780, i.e. 1780 4-digit points.
Judging by the last picture here, there are about 1000 deals. Correspondingly, it is less than 2 pips.
A total of 1038 bars. The trades are not on every bar. A continuous one colour (red or blue) is one trade.
Balance = 0.1780, i.e. 1780 4-digit points.
Judging by the last picture here, there are about 1000 deals. Correspondingly, it is less than 2 points.
Here are the statistics:
0.0001 0.0004 0.0008 0.0011 0.0014 0.0121 661
From the last column: out of 1,038 bars, the system was 661 bars out of the market.
To this we should add that the model enters/exits the pose when crossing the threshold.
Here are the statistics on the value of the upper threshold
Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
0.00000 0.00011 0.00026 0.00030 0.00044 0.00131 38
By the way, mean is comparable to spread and mah=13 pips...
The question needs to be answered: how many minimum history bars are needed for the trend to persist to the next bar? The probability of the trend persisting at 10+1 bars is much higher than at 50+1 bars and one might not consider 100+1 bars at all.
Every econometrics course (are you really an econometrician?:)) tells you what the variance of model parameter estimates is and the speed of convergence of estimates to true values: the smaller the sample size and if there are no structural changes in the series - the larger the variance of model parameter estimates. As the sample size increases, the variance (most often:)) decreases as eps*sqrt(n), eps>0, n being the number of observations.
Parameter estimation errors contribute to the error of any model. Therefore the lower the accuracy of parameter estimation - the higher the error of the model.
On the other hand, a small window allows for adaptation to parameter changes. In practice this problem is much better solved by solving the decay problem for model parameters rather than reducing the window size.
I don't have regression. A state space model is used. See above answer Math. Above I gave a general view of the model.
I distinguish between the window size over which the parameters are calculated and the sample size over which the result is calculated, which is given above in pips excluding the spread. We see a fairly smooth balance sheet growth line. Maybe I'm wrong, but for me it's very important to have a smooth balance line.