Bayesian regression - Has anyone made an EA using this algorithm? - page 14

New comment
 
Dmitry Fedoseev:
Polynomial.
Take it, calculate it, compare it.
 

If you add noise to the quotes, you get this distribution:

And how would that help trading?

 
Yousufkhodja Sultonov:
Take it, calculate it, compare it.

Why should I? I don't care, you can stay in your fog as dense as you like for as long as you like.

Besides, this proposal of yours looks very strange. Since you are presenting yourself as such a unique expert and inventor, you must know polynomial regression and know its properties.

There is absolutely no need to calculate it, there is an indicator in the codebase, you can even change the degree of polynomial, and that's really power.

 
Event:

If you add noise to the quotes, you get this distribution:

And how would that help trading?

It doesn't. But according to local regressologists - since the distribution becomes normal, you can now apply all the methods that, in their opinion, can only be applied to a normal distribution. (just kidding of course)
 
Dmitry Fedoseev:

Why should I? I don't care, you can stay in your fog as dense as you like for as long as you like.

Besides, this proposal of yours looks very strange. Since you are such a unique expert and inventor, you should know polynomial regression and know its properties.

The polynomial needs to be adapted to actual data each time, while in case of (18) you don't need to do anything, it adjusts itself in the best possible way. You just don't have the courage to admit that a better model than (18) has not yet been invented in every sense.
 
Yousufkhodja Sultonov:
The polynomial needs to be adapted to the actual data every time, and in the case of (18) you don't need to do anything, it adjusts itself in the best possible way. You just don't have the courage to admit that a better model than (18) has not yet been invented in every sense.

Why adapt it? It is the polynomial that adapts best on its own. Your curvilinear regression will only rarely fit the data. The situation here is quite different, your regression is not that it is the best or the best, it does not apply here at all.

It's also not quite clear what you call adaptation? The very essence of regression is adaptation. Why else would you call it butter?

How can you give an estimate to something you haven't tried?

 
Yousufkhodja Sultonov:
Why aren't you a millionaire then? You have a branch, talk about the charms of your (18), don't do it here.
 
Dmitry Fedoseev:

Why adapt it? It is the polynomial that adapts best on its own. Your curvilinear regression will only rarely fit the data. The situation here is quite different, your regression is not that it is the best or the best, it does not apply here at all.

It's also not quite clear what you call adaptation? The very essence of regression is adaptation. Why else would you call it butter?

The easiest way to shut my mouth is to show the workings of the polynomial model using this example. I am convinced that it has no predictive ability. It might show something at a stretch of entered factual data, but beyond that it will break away from reality.
 
Yousufkhodja Sultonov:
The easiest way to shut me up is to show the workings of the polynomial model with this example. I am convinced that it has no predictive ability. It might be able to show something at a segment of the actual data entered, but then it gets away from reality.
Otherwise, you'd think yours would be applicable for forecasting.
 
Dmitry Fedoseev:
Otherwise, you'd think yours would be applicable for forecasting.
Apparently, the market does not really care about the forecast per se, especially in the short term. In the long term, the forecast gives modest fruit in the form of 10-12% per annum, which many are not happy with.
1...789101112131415161718192021...55
New comment
Reason: