Market prediction based on macroeconomic indicators - page 57

 
Vladimir:

To Peter: I am not predicting the S&P500 directly. The purpose of this paper is to predict recessions in order to get out of the market before they occur and improve the profitability of the buy&hold strategy. Although the S&P500 contains stocks of 500 companies, it is driven by institutional investors who buy and sell the index itself (or its options), not its components. 13% a year doesn't seem like much, but enough for big money where turnover is important. Bernie Madoff attracted his clients by promising them a modest 10% a year, which he failed to achieve.

...

Thank you, the position is clear.

But if you can, answer this question: why choose a mathematical method to analyse market movements using standard approaches, like numerical series prediction with trained neural networks, and why do we need maths and MO when economics and geopolitics rule the game? What formulas and data can predict all this?

Objectively, don't you think you are imposing an impossible task on NS and the choice of analysis method is fundamentally wrong in this particular case?

Let me explain: the 500 largest companies together are reflected in billions (and not 10 000 hours) of parameters, each of which has a long history of values, and if you've chosen a numerical analysis, then why do you randomly reduce its scale n-fold, continuing to claim for any reliability of predictions? Isn't this self-deception? With such an aspiration to objectivity, "cut off the shoulder" and "sculpt" as you wish in your parametric analysis. Isn't it?



 
Алексей Тарабанов:

So, what is the forecast for the S&P500 ?

Mathematical, economic, political or existential. Take your pick.

 
Реter Konow:
Mathematical, economic, political or existential. Take your pick.

Specific. Nearest movements and levels.

 
VVT:

I'm sorry, but all this for 5-13% a year??? It's not worth it.)

That depends on how you look at it.

 
Реter Konow:
Thank you, the position is clear.

But, if you can, answer this question: why choose a mathematical method to analyse market movements using standard approaches, such as predicting numerical series with trained neural networks, and why do we need mathematics and IR where economics and geopolitics rule the game? What formulas and data can predict all this?

Objectively, don't you think you are imposing an impossible task on NS and the choice of analysis method is fundamentally wrong in this particular case?

Let me explain: 500 largest companies together are reflected in billions (not 10 000 hours) of parameters, each of which has a long history of values, and if you have chosen the numerical analysis, then why do you randomly reduce its scale n-fold, continuing to claim for any reliability of predictions? Isn't this self-deception? With such an aspiration to objectivity, "cut off the shoulder" and "sculpt" as you wish in your parametric analysis. Isn't it?



I do not understand the question. It is like a philosophicalquestion: what was created first, the universe or the laws of physics by which it evolves? If you believe that we live in a simulation, then mathematics and the laws of physics werecreated before the universe. In any case, mathematics allows us to describe what is going on around us. About the mathematical approach: the approach is correct if it works. If my approach stops predicting recessions, I will change it. By the way, I don't use neural networks. You don't need them if you only have 3 indicators. Everything is quite prosaic: if one indicator moves in one direction, and another indicator moves in another, and the third in the third, then the recession signal is output. Of course you can do it by eye or intuitively, but it is easier on the computer.

 
Реter Konow:
Mathematical, economic, political or existential. Take your pick of any one.

Econophysics says otherwise).

 
Vladimir:

Higher mathematics is of course very fascinating, according to your forecast S&P500 from 2019 there is no sell and buy signal, but if you look at the chart you will notice that the price went up by exactly the NOT forecasted 13% per year, and if you consider the high volatility of this year associated with covids you could earn by buying and selling and buying again up to 50% per year.

What is wrong with the forecast?

 
Vladimir:

I don't understand the question. Sort of a philosophicalquestion: which was created first, the universe or the laws of physics by which it evolves? If you believe that we live in a simulation, then mathematics and the laws of physics werecreated before the universe. In any case, mathematics allows us to describe what is going on around us. About the mathematical approach: the approach is correct if it works. If my approach stops predicting recessions, I will change it. By the way, I don't use neural networks. You don't need them if you only have 3 indicators. Everything is quite prosaic: if one indicator moves in one direction and another indicator moves in the other, and the third in the third, then the recession signal is output. You can of course use your eye and intuitively, but it is easier on the computer.

I carefully reread your mathematical approach to analyzing and predicting the S&P500 index on the first page and was impressed by its serious, scientific content. However, quickly got to the essence of the described actions and discovered the author's blatant "arbitrariness", that leads the analysis as he likes:

1. limits the number of indicators to be analysed as he sees fit.

2. selects the "more/less" important ones by assessing them according to some kind of scale.

3. trims historical parameter value plots at will, etc.

By the level of apparent unobservable subjectivity, it rivals fortune-telling, though presented scientifically.

If at the same time, if the predictions turn out to be correct, it's great and good for the author. :)

I am only pointing out the subjecƟvity in research and prognosticaƟon, which falls out of the picture and is protruding from all sides.
 
VVT:

Higher mathematics is of course very fascinating, according to your forecast S&P500 from 2019 there is no sell and buy signal, but if you look at the chart you will notice that the price went up by exactly the NOT forecasted 13% per year, and if you consider the high volatility of this year associated with covids you could earn by buying and selling and buying again up to 50% per year.

What is wrong with the forecast?

Again, my method does not predict the S&P500. It predicts recessions. The 2020 recession is not over yet. There is no problem with the forecast.

50% a year steady is utopia. It may succeed in one year with a small investment, but then it is just as quickly drained.

 
Реter Konow:
I carefully reread your mathematical approach to analysis and forecasting of S&P500 index on the first page and was impressed by its serious, scientific content. However, quickly went to the essence of the actions described and found the author's blatant "arbitrariness", which leads the analysis as he wishes:

1. Limits the number of indicators to be analysed as he sees fit.

2. Selects the "more/less" important ones by rating them according to some scale.

3. Cuts historical parameter value plots at will and so on...

By the level of apparent unobservable subjectivity, it rivals fortune-telling, though presented scientifically.

If at the same time, if the predictions turn out to be correct, it's great and good for the author. :)

I am only pointing out the subjectivity in research and forecasting which is falling out of the picture and sticking out from all sides.

1. predictors are chosen based on their ability to predict recessions. The selection is made automatically, without my influence or opinion.

2. the scale of evaluation is how much more profitable the proposed buy&sell strategy is than buy&hold

3. the historical plots are limited to the depth of history of individual economic performance

The only possible criticism is that the historical results do not guarantee the accuracy of predicting recessions in the future. All the results of the chart shown have been fitted to history except for the last recession signal in December 2019.

For a constructive dialogue I suggest comparing the accuracy of my system/model with other fundamental or thechnical recession prediction systems. You could also compare the yield + drawdown of my system to other systems trading the S&P500.

Reason: