Econometrics: one step ahead forecast - page 24
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
For example, we need to get XP on 10 points, so we take:
1) points 23, 22, 21, 20....
2) points 40, 39, 38....
Another thing is that with a new point, the HP curve calculated from the new "old" point will change.
I do not share the opinion of the topicstarter and believe that in the way he does, KP should not be used.
Also, I speak "in defense" not of ideas and methods of the topicstarter, but of XP, and to say that this filter uses "future" values, I think it is not correct. In the tester, it may be possible to look into the future, but in real-time - no (where will it get these "future" values?). So, the statement "KP uses data from the future" is profanity.
I don't share the topstarter's opinion and think that the way he does it, KP should not be used.
Some glimpses of a "future" candle are found here https://forum.mql4.com/ru/44275/page11
A standard TA technique is pattern work.
Exactly!
So put aside your econometric snobbery. Go back to the previous page. And think about what I said.
And since you don't understand where the "future" values come from in the KP formula, here's a hint
This tay(t+1) is the "future".
It's more interesting. How do I use it and why not?
As far as I understand, you are using an estimate of the prediction error of the deviation from KP. And you shouldn't do that because that very deviation from KP on the same bar, but with a new one, will be different.
If it works, why not use it to your advantage, you just need to accumulate statistics and then try to teach it to an advisor.
As far as I understand, you are using an estimate of the prediction error of the deviation from KP. And you shouldn't do that because that same deviation from KP on the same bar, but with a new one, will be different.
I'm interested in the forecast one step ahead.
Forecast = extrapolation of KP from the previous 4 bars + linear extrapolation of the residual from the last two bars.
On the arrival of a new bar this is repeated.
What difference does it make what happens to the forecast calculation on the previous bar that we have already worked out?
If we take non-calculated smoothing, will the quality of the forecast improve?
Where does it follow from?
No question about it. All TAs are set up that way. But advisers go stale over time.
I'm interested in the prediction one step ahead.
Forecast = extrapolation of KP from the previous 4 bars + linear extrapolation of the residual from the last two bars.
When a new bar comes, it repeats.
What difference does it make what happens to the forecast calculation on the previous bar that we have already worked out?
If we take uncalculated smoothing, will the quality of the forecast improve?
Where does it follow from?