Econometrics: State-space model forecasting - page 20
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 won't say. I will post all information from the tester in response to a similar model.
has the model been tested in real time ?
No. It's a base to work from. I can predict the size of the window - the result is very dependent. Apart from this is the EA itself - you can work on the size of the loss.
Join in.
I will post additional results on similar ones. So far it is much more candid than the prediction branch at faa.
No. It's a base to work from. I can predict the size of the window - the result is very dependent. Apart from this is the EA itself - you can work on the size of the loss.
Join in.
I will post additional results on similar ones. So far it is much more frank than the prediction branch at faa.
i have to choose the smallest timeframe possible for the model used (model timing) and run more bars in real time ...then compare using the same data with the currently used methodology in the program... results should be 1 to 1...if 1 in 1 then move on... no - look for flaws... I think there are flaws... imho all of course...
thanks no... the ideology is different...
for a dialogue with people to be more constructive... it's worth attaching a similar (and not a grail)) to the model used (although few people use R as I understand (personally, I have not seen it in the eye and have no desire to) + or at least attach a data file to each chart... where the columns-date, what predicted, what you shoved, what you got ...
+ there's a possibility that similar results may be obtained by other ... simpler methods and someone will let it slip )))... it's a matter of honesty and correctness in casting the fishing rod ...
choose the smallest timeframe possible for the model used (model calculation time) and run more bars in real time ...then compare on the same data using the current methodology in the software... results should be 1 to 1...if 1 in 1 then move on....no - look for flaws... I think there are flaws....imho all of course...
thanks no...ideology is different...
for a dialogue with the people to be more constructive... it's worth to attach a similar (and not a graal)) to the model used (although few people use R as I understand (personally I've never seen it and have no desire to) + or at least attach a data file to each graph... where the columns date, what predicted, what we shoved, what we got ...
+ there's a possibility that similar results may be obtained by other ... simpler methods and someone will let it slip )))... it's a matter of honesty and correctness in casting the fishing rod ...
Saw figures somewhere: 700 downloads of the R wrapper, and 1200 downloads of the autoregression forecast indicator. The topic was opened under these people.
R contains many models. The complexity of using them programmatically is about the same - you refer to a function. What matters is WHAT the model displays in the quote. The advantage of the one given is that it is a dynamic model for a non-stationary process. My understanding is that it is impossible to implement such a model in TA. As well as the indicator for R posted in kodobase.
I saw figures somewhere: 700 downloads of the R wrapper, and 1200 downloads of the autoregression forecast indicator. The topic was opened under these people.
R contains many models. The complexity of using them programmatically is about the same - you refer to a function. What matters is WHAT the model displays in the quote. The advantage of the one given is that it is a dynamic model for a non-stationary process. As I see it, it is impossible to implement such a model in TA. Just like the indicator for R posted in kodobase.
Everything where formulas are used is TA...econometrics is one of them...
do p1 (realtime)... it will be useful...
everything where formulas are used is TA...econometrics is one of them...
do p1 (realtime)... it will be useful...
There is a plan for the work to go to real. It is being implemented. Only part of the material has been published. I have written more than once: I am interested in the source of income for years to come. Hence the interest in R.
You are wrong about TA. There is a fundamental difference.
1. In TA, the characteristics of the kotir are never considered.
2. TA does not consider whether the quotient characteristics are correctly mapped by a particular TA tool
3. There is no concept of "evaluation" in TA - everything is substituted by the tester. And there is no question how much trust can be given to the test results. The forward test confirmed the test - that's it, the truth.
You somehow misunderstand what TA is.
TA is any analysis of a clean chart.
3. There is no concept of "evaluation" in TA - everything is substituted by a tester. And there is no question of how much you can trust the test results. Forward test confirmed testing - that's it, the truth.
You somehow misunderstand what TA is.
TA is any analysis of a clean chart.
No, they're not. The performance evaluation is essentially the same -- the real profit. Before that - who can and who knows how. And what is a forward test and how to do it correctly know hardly 10 percent of those who deal with the tester, or even less.Well, why wrong. When there were "chartists" from the word "chart" and then there was "technical analysis" something close to guessing by coffee grounds.
By the way faa showed it in his article here on the site.
In TA the significance of TS parameters is not checked at all, in contrast to econometrics, where everything starts with the significance of model parameters. In my model all parameters are significant, i.e. it can be trusted, as those parameters which I see, are such.
Well, why wrong.
There is a plan for the work to go to the real world. It is being implemented. Only part of the material is laid out. I have written more than once: I am interested in the source of income for years to come. Hence the interest in R.
You are wrong about TA. There is a fundamental difference.
1. In TA, the characteristics of the kotir are never considered.
2. TA does not consider whether the quotient characteristics are correctly mapped by a particular TA tool
3. There is no concept of "evaluation" in TA - everything is substituted by the tester. And there is no question how much trust can be given to the test results. Forward test confirmed the test - that's it, the truth.
What a load of nonsense... it's unclear where you and the faa got it from... not to mention that you can make indicators by 1 and 2 p... and they do exist... I saw them in one form or another long time ago...
In the meantime, the near-scientific conversations and funny pictures will go on endlessly... and the model will then go down the drain... imho...