What makes an unsteady graph unsteady or why oil is oil? - page 2
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
Then I'll ask questions too.
-
Fourier describes history very well. Why is this kind of extrapolation so bad at predicting future prices? That's the 1st question.
Now the 2nd, most interesting question: if Fourier doesn't work, doesn't that mean that other types of extrapolation won't work either?
Hmmmm, what does this have to do with Fourier? Hmmmm, what's that got to do with ectrapola? Logic, logic - all too straightforward. No fantasy pushing back on reality.Во вторник поиск заработает, наверное :)
Nah... ...it's never going to work for some people. Why? He probably needs to have opinions addressed to him in person. Now, all who have posted on the subject, drop everything and personally, here in a branch, papadet their past posts. And write: Dear Sir, here's my opinion on the matter of interest to you.
And search... No - why would he be a sucker to browse through other people's threads? )))
Nah... ...it's never going to work for some people. Why? He probably needs to have opinions addressed to him in person. Now, all who have posted on the subject, drop everything and personally, here in a branch, papadet their past posts. And write: Dear Sir, here's my opinion on the matter of interest to you.
And search... No - why would he be a sucker to browse through other people's threads? )))
The search really doesn't work. Are you some kind of administrator or something?
:))) Everyone's an admin here except me :) In all seriousness, you should read the forum of course, there are a lot of valuable insights here. It is important to avoid the process of reinventing the wheel, it will save a lot of time and nerves. But, on the other hand it is difficult to find something "especially valuable" too.
Fourier describes history very well. Why is this type of extrapolation so bad at predicting future prices? That's the 1st question.
Now the 2nd and most interesting question: if Fourier doesn't work, doesn't that mean that other types of extrapolation won't work either?
Fourier has NEVER been a method of extrapolation and is not intended to be one - anyone not aware of this should consult basic mathematics. It is a method of INTERPOLATION.
What you can do with Fourier : reconstruct missing values of a function assuming that the function is periodic.
The same limitation makes this method meaningless for EXTRAPOLATION. If it's not clear why - see definition of periodic function.
Regarding the other methods: no interpolation method that is strictly dependent on right-edge conditions is suitable for extrapolation.
Good luck.
ZS This isn't actually a personal appeal.
Поиск действительно не работает. Вы типа админ что-ли?
No, it's not. It's just that the subject has really been dealt with a million times. And the search - yes, it doesn't work. Even via Google (stationary site:https://www.mql4.com/ru/) - apparently no indexing.
Yes I have nothing against the lam first time this is discussed. Just who among those who have written about it qualitatively a lakh times will want to write it again? Although...))) maybe they will.
Тогда и я вопросы задам.
-
Can I try to answer
1. There are no idols and Elder in this case did not say exactly. More accurately, no matter where the price goes, you must be ready for it (have a plan). It is true there was a whole branch on the forum long ago devoted to calculation of the optimum payoff from various parameters. So the forecast was included in the 4th degree there. It has the strongest effect on the value of profits.
2. You can approximate it very well with Fourier, but the further statement "the forecast with it is bad" is too arrogant. There is a winged phrase "maybe you just don't know how to cook it ?".
3. there are methods which are superior, mathematically superior, to a forecast constructed with Fourier alone.
May I ask you 1 question too, to understand how deep you investigated Fourier?
You didn't select all the components of the spectrum for prediction but some of them (it makes no sense to select all of them), if so, the optimal procedure is byes, but in practice it is not applicable most often due to high requirements for a priori data. And there is only 1 unknown parameter. signal to noise ratio. So my question is what is your signal to noise ratio and how did you find it?
Fourier has NEVER been a method of extrapolation and is not meant to be - refer to basic mathematics. It is a method of INTERPOLATION.
What you can do with Fourier : reconstruct missing values of a function, assuming that the function is periodic.
The same limitation makes this method meaningless for EXTRAPOLATION. If it's not clear why - see definition of periodic function.
Regarding the other methods: no interpolation method that strictly depends on right-edge conditions is suitable for extrapolation.
Good luck.
I'm not a mathematician and I'm probably forgiven, but the question is: what then is suitable for extrapolation?
May I ask you 1 question too, to understand how deep you investigated Fourier?
You didn't select all the spectrum components for prediction but some of them (it makes no sense to select all of them), if so the optimal procedure is byes, but in practice it is not applicable most often due to high requirements for a priori data. And there is only 1 unknown parameter. signal to noise ratio. So my question is what is your signal to noise ratio and how did you find it?
Honestly, I don't. I haven't even thought about it yet. But, I'll have to think about it, thanks for the advice.
That Fourier doesn't work at all is not true. It works, but my system results in big drawdowns (one) and after 5-10 profitable trades appear 1-2 loss-making ones (two), change of SL\TR ratio does not help.
Я не математик и мне наверное простительно, но вопрос: что тогда пригодно для экстраполяции?
In general, I got the impression that you personally understood this - so I corrected the post. It wasn't really addressed to you personally.
About extrapolation: you can use regression - the result will be more stable, but still not very good.
And about the trading: pay attention to Svinozavr's posts - extrapolation is not necessary for successful trading ;).
Good luck.