Remembering veterans: Box and Jenkins - 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
All you want to do is to get rich.
Econometrics is not about self-interest, it's about the world crisis.
It does not matter exactly what the indicator is and it is not the point that it is meaningless. The most important thing for a nerd is to calculate everything that comes to hand. But the programmer has finally mastered EViews and is happy that he can plug some data into the program and get in return some meaningless numbers. The main thing here is not the result, but the joy of the process.
The econometrician wants us to rejoice together with him. So let us not spoil his mood with mercantilism and rejoice at the meaningless possibilities of econometrics.
Let us also honour the memory of Jenkins and Box, the instigators of the world crisis, with a minute's silence, i.e. we should try to live a whole minute without swearing at them. It will not happen at once and not at all, but you have to try.
Reshetov writes well, you cannot even be offended.
Everything I read on econometrics is not an acknowledgement of the SB and the efficient market. As far as I am concerned, the crisis was created by these people, based on them Nobels, Markowitzes and various funds with a capitalisation of thousands of trillions of dollars. Box and Jenkins are not one of them. They solved a very simple problem - to move from an unsteady market to an almost stationary market and to account for noise other than a trend. Everything.
In picture form, the above equation.
Everything I read on econometrics is not an acknowledgement of SB and the efficient market.
Do econometricians have mathematical proof of the inefficiency of markets? Don't tell me that if Jenkins and Box cannot find such a proof, then that in itself is proof, because even the outright nonsense of the founding fathers of the econometric sect is not subject to any doubt, but is the truth in the last resort.
As a picture of the above equation.
Malevich rests.
And on OOS to continue the same equation without changing anything in it, is it boring?
Malevich rests.
Can I continue the same equation on OOS without changing anything in it, breakdowns?
Do econometricians have a mathematical proof of the inefficiency of markets? Don't tell me that if Jenkins and Box couldn't find such proof, then that in itself is proof, because even the outright nonsense of the founding fathers of the econometric sect is not subject to any doubt, but is the truth in the last resort.
One can prove it and one can take it on faith. I have seen the proof, if I find it I will give you the link. The premise of Box and Jenkins is the non-stationarity of the market, the presence of memory in it. They built their models not like I did, out of the blue, but after analysis of ACF and CFC + 1 more nuance, about which I will write later on occasion.
And would it be too much trouble to use OOS to continue the same equation without changing anything in it?
Read Brukova, you don't want to fuss. Then just remember (and for many on the forum a revelation) the people who have had a huge impact on the analysis and forecasting of economic data.By the way the forecast picture.
By the way, a picture of the forecast.
By the way, there has already been a reminder that Malevich rests. You do not need to study any econometrics to make such "forecasts", it is only necessary and sufficient to construct an envelope from the price plus minus a kilometre.
For special econometricians, nobody asked for forecasts, but a request to continue the old picture on OOS, i.e. to show both the data on which the so-called meaningless indicator was adjusted and the data on which it was not adjusted, all in the form of one chart.
By the way, there has already been a reminder that Malevich rests. You don't need to study any econometrics to make such forecasts, you just need to construct an envelope from the price plus minus a kilometre.
For special econometricians no one asked for forecasts, but for continuation of the old picture on OOS, i.e. to show both the data on which the so-called meaningless indicator was fitted, and the data on which it was not fitted in the form of a single chart.
The answer is the same, Brukow. You have to solve a whole series of problems for this picture. Quite time consuming.
I just remember respected by many people and it is not part of my plans to teach how to build profitable systems. Personally, you can do it better than I can.
I answered the same question, Brukow.
What is the use of Brukiv with his so-called "predictions" of the exchange rate of the quid, when there is a more sensible book: "Mathematics in Economics" by S. Yudin. V. Yudin?
At least Yudin doesn't get hung up on econometrics and gives examples of various applied economic (and not just economic) problems. Moreover, the software in Yudin's book is not proprietary, i.e. one does not have to buy a pig in a poke to solve some meaningless econometric problems.
For the particularly gifted econometricians, no one asked for forecasts, but a request to continue the old picture on the OOS, i.e. to show the data on which the fit was made
Above was my answer to the question and unsuccessful. Thinking that's what Box and Jenkins would have answered.
All sorts of tricks with forward tests make sense for a non-stationary market. The ARMA model necessarily includes a stage to determine the stationarity of the quotient - the so-called unit root test. Here is the result.
By this test we strictly reject the hypothesis that our quotient is stationary. Hence, we conclude that we cannot trust the forward test even if it gives a positive result - it is just a coincidence and nothing more, and no forward test for this model will not provide any guarantees for the future.
I have written about this many times. Forward test is a very dangerous thing.
What to do?
Box and Jenkins teach: you should differentiate a quotient, i.e. take the difference of neighboring bars. This is what is called "integrating". Let's do a unit root test for the first difference.
For the difference, the test shows that that difference is stationary, hence we should build a model for the price difference.
But that's later.
What is the use of Brukiv with his so-called "predictions" of the exchange rate of the quid, when there is a more sensible book: "Mathematics in Economics" by S. Yudin. V. Yudin?