Alternative Data
ADs provide a better picture of a company's situation, raw materials, currencies. It also allows us to assess the "current state" (nowcasting) of significant indicators. Those data make trading signals better, more precise, less risky and more profitable.
It is a revolution accompanying the AI revolution and even preceding it. In my opinion, it is more important today than AI, which is only in its early stages (despite many impressive achievements). In my opinion, through AD, funds can earn more and build their competitive advantage over others.
ADs are not part of Artificial Intelligence. An example of AD is credit card sales data. This data can be used to predict the financial performance of companies. If we have historical data, then in the simplest case, to make forecasts, all we need is a spreadsheet!
And when we are interested in more advanced indicators of future profitability, such as consumer spending patterns, brand loyalty, switching between products/brands, trending moods, competitors performance, models created using Machine Learning can come into play.
With the increasing number of data sources and the complication of forecasting models, traditional ones will be replaced in a considerable part or even entirely by AI/ML-based models.
For a broader discussion of Alternative Data, see the separate article in this issue.
Visualization
It is easier for humans to look than to think. "Analysis" by sight developed long before abstract thinking.
There is something severe behind this remark. It is much easier for us to understand a situation when it is shown using images rather than just a verbal description. Therefore, as much as possible, use visual aids - graphics, pictures, diagrams, or charts to illustrate data, situations and processes.
Indeed, it is good practice to consider what goal we want to achieve, define the target group and identify which parts of the message will benefit from such enhanced presentation. The same applies to respecting the simplicity of the message, playing with colours and ensuring maximum readability.
Another good practice is to provide a benchmark, or reference point, to which we compare some quantity. Our mind performs better by observing the differences between some benchmarks and the current indication.
An excellent practice is to make it easy for the audience to understand the situation quickly. Thus, when preparing visual aids, try to help them understand the situation as quickly as possible – for example: whether we are in the realm of "normal" or have already gone beyond it.
All key, critical processes should have some sort of graphic representation. It should allow for a quick assessment of the situation, especially in unusual or crises. So let’s say I give you a colour scheme, where green means everything is going well, orange – attention required, and red – we have a critical situation. Sound familiar? It should.
As AI matures, the amount of information and complexity of systems (and portfolios) will only increase. Therefore, using standardized metrics within a company to illustrate key processes is something worth developing as a valuable skill.
Let me say it another way to emphasize the particular importance of this topic - the ability to graphically present important processes for a company is a competence worth developing. It is worth discussing what indicators to use, what types of graphs, what colours, and what schemes to facilitate and enhance understanding, ability, and speed of decision making.
Visual communication is one of the essential elements of building and consolidating a company's structural intelligence.
Automatization
Automatization is the critical process underlying the use of artificial intelligence.
It involves gradually learning and automating more functions of human intelligence. The ultimate stage of AI development in trading is full machine autonomy with a level of perception, "thinking", decision-making far exceeding human capabilities in every aspect.
What does this mean for traders and funds now and in the future?
Now
Today, automatization is one of the main topics because it takes the burden of routine activities and responsibilities off the shoulders of traders. One of the main problems that traders complain about is excessive workload and information overload.
The primary candidates for automation are routine activities that require no intellectual input. And over time, more and more activities will be automated - and more about that in a moment.
Suppose we have a great trader. Only some of his activities add value, and he should focus on them. You can consider using supporting programs or someone else to help with the remaining tasks.
What should not be automated are non-routine decisions, decisions in exceptional or critical situations and those requiring synthetic expertise beyond the reach of AI tools.
Instead, you can automate the execution of decisions in critical situations with confidence.
In an extreme situation, the trader only presses the appropriate key. A program then tries to escape from the market as quickly as possible. It tries to use liquidity, reduce costs and minimize the negative impact of the large order it exits. In nine out of ten situations, it will do this better than the trader and, in the case of substantial orders, in ten out of ten.
Automatization will expand to include more and more activities, including non-routine ones, over time.
In the future
To understand what automatization in a fund will look like in the future, we must first learn the decision-making process of a discrete trader or automated system.
The decision-making process consists of all the elements that lead from the initial analysis (what to trade and where to trade it) through the choice of location, entry, position management, exit, to post-trade analysis.
There can, of course, be many more of these steps if we take a more detailed approach (and the largest funds do).
Automatization here is about taking a single element of the decision-making process and trying to refine it first (to find the best practices) and then automate it.
It would also be beneficial to provide a feedback channel so that we and, in time, the AI system can improve this element based on the incoming and analyzed data. In short, we want the system to learn on its own.
In short, we automate best practices at each step and provide feedback so that the system learns and improves.
On the other hand, entry automation may involve breaking positions into smaller ones, examining order structures above and below, creating and executing entry strategies to minimize cost and adversarial price moves. Hiding positions and maximizing positions for the best signals may also be part of the automation.
Summary
We have discussed six of the "hottest" topics currently occurring in the Artificial Intelligence field. Two are sure to be the most important: XAI and Alternative data.
The first - because it opens up a powerful new trend of adjusting the latest tools to a trader's level of understanding. We already know that a gradient descent on a differentiable manifold tells him nothing. The second - because it is alternative data that gives traders and funds their main competitive advantage today.
In conclusion, it is worth repeating one important thought: the AI revolution is just beginning. It will completely change our world and ways of investing. This process is incredibly fascinating. The New City Trader was born out of a desire to share this fascination.
If you liked this post, give it a boost 🚀 and drop a comment so we know to publish more for you. Cheers!
Follow: https://www.tradingview.com/u/drSwierk/
ADs provide a better picture of a company's situation, raw materials, currencies. It also allows us to assess the "current state" (nowcasting) of significant indicators. Those data make trading signals better, more precise, less risky and more profitable.
It is a revolution accompanying the AI revolution and even preceding it. In my opinion, it is more important today than AI, which is only in its early stages (despite many impressive achievements). In my opinion, through AD, funds can earn more and build their competitive advantage over others.
ADs are not part of Artificial Intelligence. An example of AD is credit card sales data. This data can be used to predict the financial performance of companies. If we have historical data, then in the simplest case, to make forecasts, all we need is a spreadsheet!
And when we are interested in more advanced indicators of future profitability, such as consumer spending patterns, brand loyalty, switching between products/brands, trending moods, competitors performance, models created using Machine Learning can come into play.
With the increasing number of data sources and the complication of forecasting models, traditional ones will be replaced in a considerable part or even entirely by AI/ML-based models.
For a broader discussion of Alternative Data, see the separate article in this issue.
Visualization
It is easier for humans to look than to think. "Analysis" by sight developed long before abstract thinking.
There is something severe behind this remark. It is much easier for us to understand a situation when it is shown using images rather than just a verbal description. Therefore, as much as possible, use visual aids - graphics, pictures, diagrams, or charts to illustrate data, situations and processes.
Indeed, it is good practice to consider what goal we want to achieve, define the target group and identify which parts of the message will benefit from such enhanced presentation. The same applies to respecting the simplicity of the message, playing with colours and ensuring maximum readability.
Another good practice is to provide a benchmark, or reference point, to which we compare some quantity. Our mind performs better by observing the differences between some benchmarks and the current indication.
An excellent practice is to make it easy for the audience to understand the situation quickly. Thus, when preparing visual aids, try to help them understand the situation as quickly as possible – for example: whether we are in the realm of "normal" or have already gone beyond it.
All key, critical processes should have some sort of graphic representation. It should allow for a quick assessment of the situation, especially in unusual or crises. So let’s say I give you a colour scheme, where green means everything is going well, orange – attention required, and red – we have a critical situation. Sound familiar? It should.
As AI matures, the amount of information and complexity of systems (and portfolios) will only increase. Therefore, using standardized metrics within a company to illustrate key processes is something worth developing as a valuable skill.
Let me say it another way to emphasize the particular importance of this topic - the ability to graphically present important processes for a company is a competence worth developing. It is worth discussing what indicators to use, what types of graphs, what colours, and what schemes to facilitate and enhance understanding, ability, and speed of decision making.
Visual communication is one of the essential elements of building and consolidating a company's structural intelligence.
Automatization
Automatization is the critical process underlying the use of artificial intelligence.
It involves gradually learning and automating more functions of human intelligence. The ultimate stage of AI development in trading is full machine autonomy with a level of perception, "thinking", decision-making far exceeding human capabilities in every aspect.
What does this mean for traders and funds now and in the future?
Now
Today, automatization is one of the main topics because it takes the burden of routine activities and responsibilities off the shoulders of traders. One of the main problems that traders complain about is excessive workload and information overload.
The primary candidates for automation are routine activities that require no intellectual input. And over time, more and more activities will be automated - and more about that in a moment.
Suppose we have a great trader. Only some of his activities add value, and he should focus on them. You can consider using supporting programs or someone else to help with the remaining tasks.
What should not be automated are non-routine decisions, decisions in exceptional or critical situations and those requiring synthetic expertise beyond the reach of AI tools.
Instead, you can automate the execution of decisions in critical situations with confidence.
In an extreme situation, the trader only presses the appropriate key. A program then tries to escape from the market as quickly as possible. It tries to use liquidity, reduce costs and minimize the negative impact of the large order it exits. In nine out of ten situations, it will do this better than the trader and, in the case of substantial orders, in ten out of ten.
Automatization will expand to include more and more activities, including non-routine ones, over time.
In the future
To understand what automatization in a fund will look like in the future, we must first learn the decision-making process of a discrete trader or automated system.
The decision-making process consists of all the elements that lead from the initial analysis (what to trade and where to trade it) through the choice of location, entry, position management, exit, to post-trade analysis.
There can, of course, be many more of these steps if we take a more detailed approach (and the largest funds do).
Automatization here is about taking a single element of the decision-making process and trying to refine it first (to find the best practices) and then automate it.
It would also be beneficial to provide a feedback channel so that we and, in time, the AI system can improve this element based on the incoming and analyzed data. In short, we want the system to learn on its own.
In short, we automate best practices at each step and provide feedback so that the system learns and improves.
On the other hand, entry automation may involve breaking positions into smaller ones, examining order structures above and below, creating and executing entry strategies to minimize cost and adversarial price moves. Hiding positions and maximizing positions for the best signals may also be part of the automation.
Summary
We have discussed six of the "hottest" topics currently occurring in the Artificial Intelligence field. Two are sure to be the most important: XAI and Alternative data.
The first - because it opens up a powerful new trend of adjusting the latest tools to a trader's level of understanding. We already know that a gradient descent on a differentiable manifold tells him nothing. The second - because it is alternative data that gives traders and funds their main competitive advantage today.
In conclusion, it is worth repeating one important thought: the AI revolution is just beginning. It will completely change our world and ways of investing. This process is incredibly fascinating. The New City Trader was born out of a desire to share this fascination.
If you liked this post, give it a boost 🚀 and drop a comment so we know to publish more for you. Cheers!
Follow: https://www.tradingview.com/u/drSwierk/
