Machine learning in trading: theory, models, practice and algo-trading - page 1716

New comment
 
Reg Konow:
Is news fundamental data? If so, their digitization (for use in technical analysis) is subjective.

The balance sheet is technical data. The same as price, volume, OM can be put directly into formulas, but statements of politicians and global and local events of all kinds that influence people's minds and make them rethink the value of traded objects are fundamental data. In short, everything that is expressed in numbers is technical data, and in words and meanings is fundamental. My understanding is this.

I'm not going to convince you. I am more close to the classic concepts of professors HSE and Schumpeter.))) According to your logic, the number of products in the warehouse, the productivity of machines, the number of workers and ITRON, this is technical data, and the statements and orders of management are fundamental.

 
mytarmailS:

I wrote this just for the sake of example, you don't have to delete it.

In general, the task of searching for cycles within price changes, decomposing it into various one-type movements, in my opinion, has perspectives and is more correct than searching for regularities simply by price. But on the other hand it is a complicated probabilistic problem. Even if there are solutions, their accuracy is hard to estimate.

 
Valeriy Yastremskiy:

In general, the task of searching for cycles within price changes, decomposing it into different single-type movements, in my opinion, has prospects and is more correct than looking for patterns simply by price. But on the other hand it is a complicated probabilistic problem. Even if there are solutions, their accuracy is difficult to estimate.

There are no cycles in the price, there are cycles in volatility, but it doesn't show buy or sell direction.

 
Valeriy Yastremskiy:

...According to your logic, the number of products in the warehouse, the productivity of machines, the number of workers and ITF, this technical data, and the statements and orders of management are fundamental.

Right, about the technical data, but a correction for the fundamental data. Orders and instructions to be obeyed are also technical data, because they do NOT imply freedom of interpretation and possibility of non-fulfillment, but the way they affect the consciousness is the domain of fundamental analysis.
That is, FA is an analysis of subjectively significant events, words and numbers and an arbitrary, unsystematic, approximate interpretation of their meaning within some, personal ideology and worldview.

There, twisted...))

In short, any data can be considered technically (numbers, formulas, dependencies, patterns) and fundamentally (good, bad, disaster...).

FA operates with logic, psychology, economics, and technical analysis is guided by numbers, patterns, indicators, and predicts repetitions by statistical probability.
 

I have a fix idea )) But I need data.


Take all pairs with dollar and all pairs with euro

Decompose all pairs with the dollar and all pairs with the euro

And we'll find the same component repeated in every pair of the dollar, it'll be a real dollar index.

Do the same with the euro.

and then graph the dollar index against the euro

What do you think of the idea?

 
Tag Konow:
In short, any data can be viewed technically (numbers, formulas, dependencies, patterns) and fundamentally (good, bad, disaster...).

Except that good, bad, and a catastrophe... these are subjective notions, and a figure is a figure.

For example, the market fell by 10% for someone it'sa disaster, and for someone it's good because he sold, while the figure is 10%!

 
mytarmailS:

I have a fix idea )) But I need data.


Take all pairs with dollar and all pairs with euro

Decompose all pairs with the dollar and all pairs with the euro

And we'll find the same component repeated in every pair of the dollar, it'll be a real dollar index.

Do the same with the euro.

and then graph the dollar index against the euro

What do you think of the idea?

Any idea needs validation. I wouldn't call the component of the pattern I found an index. And would try to link it to external data. Maybe we will be lucky and it will be found))))

Of course a more complete study of all courses will give a more complete picture. And over a long enough time period.

Why is there no service for checking the Expert Advisor on all of the currently available data? For all currencies and for all available timeframe. The data is there. ))))

 
mytarmailS:

Hmii ... except that good, bad, disaster... these are subjective concepts, and the figure is a figure.

For example the market fell by 10% for someone it'sa disaster, and for someone it's good because he sold it, while the figure is a 10% figure!

Right. This is the difference between TA and FA. Although, the FA has a huge impact on the stock price, e.g., at moments of significant or inflated events. That is, in a "cold" market, the TA forecast is more accurate than the FA, but when something happens, the FA is irreplaceable.
 
Retag Konow:
Right, about the technical data, and a correction on the fundamentals. Mandatory orders and instructions are also technical data, because they do NOT imply freedom of interpretation and the possibility of non-compliance, but how they affect the mind is the domain of fundamental analysis.
That is, FA is an analysis of subjectively significant events, words and numbers and an arbitrary, unsystematic, approximate interpretation of their meaning within some, personal ideology and worldview.

There, twisted...))

In short, any data can be considered technically (numbers, formulas, dependencies, patterns) and fundamentally (good, bad, disaster...).

FA operates with logic, psychology, economics, while technical analysis is guided by numbers, patterns, indicators, and predicts repetitions by statistical probability.

I understand your approach. Today even the numerical data does not know how to screw with the analysis with a sufficient degree of probability of prediction. And if you can digitize and systematize by your FA, the Nobel Prize would be at least))))

 
Valeriy Yastremskiy:

I understand your approach. Today even the numerical data do not know how to screw with the analysis with a sufficient probability of prediction. And if you manage to digitize and systematize by your FA, the Nobel Prize will be at least))))

This is almost equal to the creation of AI. ))
1...170917101711171217131714171517161717171817191720172117221723...3699
New comment