Econometrics: State-space model forecasting - page 5
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
Well by one bar only. I have another difficulty in assessing - visually it is difficult to assess even the potential profitability. The truth is nearby - in the tester. However, it seems he (avtar) hasn't even looked into mql, and so he will make clever R nonsense for fifty pages, until someone like you offers a free mql-realization. Looks like his professor subcortex is counting on it... ))
In short - another Yusuf. )))
Not familiar with Yusuf.
You should test a properly designed mechanism, not the crap you use to convince yourself of the profitability of an EA in the tester.
PS. I'm not familiar with MQL, but I don't need any help with it - it's much simpler than R and you can't compare it to models at all. There's plenty of ready-made code here - I'll manage MQL someday.It is not a prediction error, it is the spread that you are trading. That's what you need.
The TS would be a grail if the price would return to the forecast all the time. But in this case you will have at least 50/50 and you will lose with the speed of spread.
Well, it's only by one bar. I have another difficulty in assessing - visually it is difficult to assess even the potential profitability. The truth is nearby - in the tester. However, it seems he (avtar) hasn't even looked into mql, and so he will make clever R nonsense for fifty pages, until someone like you offers a free mql-realization. Looks like his professor subcortex is counting on it... ))
(Well - another Yusuf...) )))
Now by one bar, on the next bar by one bar, etc.
About the tester is correct, the adviser should write, in the tester even half of what is on the picture will not be. But who of these R experts may explain at least approximately on the fingers and with interjections what we should count and how to do it?
In the first picture, almost all directions of change are correctly predicted, which is fantastic.
That's why I'm interested in the topic. But I don't know how randomly the picture was chosen to illustrate it.
It is reasonable to assume that the "most beautiful of the set" is chosen. // i.e. a variant of "fitting".
In the first picture, almost all directions of change are correctly predicted, which is fantastic.
This fantasy is of little use. Everything can be eaten up by the spread if there are enough pullbacks on the trend, and you can't do without them.
So the challenge is to find a threshold that says it's a "forecast" and not nothing, it's a real "pullback" and not nothing. I.e. a certain channel around the price, inside which the forecast is not taken into account.
That's why I'm interested in the topic. But I don't know how randomly the picture was chosen to illustrate it.
It's reasonable to assume that the "prettiest of the bunch" has been chosen. // i.e. a variant of "fitting".
Anonymus isn't saying anything about a function I don't have.
I'll have to write it myself, so I can guess the result without a tester.
Now by one bar, on the next bar by one bar, etc.
The tester is right, an advisor should be written, in the tester even half of what is on the picture. But which of these R-specialists can explain at least approximately on the fingers and with interjections what to count and how to count it?
That's why I got interested in the topic. But I don't know how randomly the picture was chosen to illustrate it.
It's reasonable to assume that the "prettiest of the bunch" has been chosen. // i.e. the 'fit' option.
Was watching faa's thread. He, too, stopped at the question of what to do with the forecast. Although the models he posted are not working. Somehow it's not there.
For me, one thing is clear, that just the forecast will not work.
I understand all that. I really need half a bar of 100% prediction to make a grail.