Please check out the previous series on evolution as introduction to this preface.
The best way to forecast and profit from knowing the future is to analyze reactions of financial market to the past events and draw conclusions from that.
Financial markets are best forecasting tools
My fascination with the subject of evolution as a forecasting algorithm, see Great Question Part IV - Evolution as an algorithm, resulted in my own desire to create an algorithm that forecasts events in the setting of financial markets. Financial markets are best forecasting tools that we have developed so far, whether when the forecasting is done 1 year or 1 ms ahead.The best way to forecast and profit from knowing the future is to analyze reactions of financial market to the past events and draw conclusions from that.
This is how an idea of the Newsdaq came from. I wanted to create an eco-system where the players who would make correct decisions would be rewarded without paying too much of a penalty, but have enough at stake to care for decisions they are making.
Newsdaq
Newsdaq was a both a shared economy marketplace for amateur research analysts and a forecasting tool for assessing market impact of the world events. People would classify and rank stories by their explanatory values and would get rewarded by enriching research reports.
This program was web based and was developed in 2003, before the concept of shared economy business model was developed on top of Airbnb and Uber and lasted up until 2011. The intended customers for the product were financial companies who would pay premium for enriched "alpha" signal for their strategy.
Ultimately this effort failed to find actual customers. I believe because the product was intended for consumption by electronic trading systems in order to enrich their trading strategies in a process known as "alpha blending". The problem is that most of the information is already "priced into" the movement of market prices and adding one more source of information without ability to vet or test it would lead to sub-par results.
What I should have done instead is to cater to human consumption, where the vast near real-time news feeds are processed and summarized in a human-centered UI.
More on the development of the NewsdAI in the next blogs...
This program was web based and was developed in 2003, before the concept of shared economy business model was developed on top of Airbnb and Uber and lasted up until 2011. The intended customers for the product were financial companies who would pay premium for enriched "alpha" signal for their strategy.
Ultimately this effort failed to find actual customers. I believe because the product was intended for consumption by electronic trading systems in order to enrich their trading strategies in a process known as "alpha blending". The problem is that most of the information is already "priced into" the movement of market prices and adding one more source of information without ability to vet or test it would lead to sub-par results.
What I should have done instead is to cater to human consumption, where the vast near real-time news feeds are processed and summarized in a human-centered UI.
NewsdAI
Rather than creating a marketplace for human analyst to compete with each other and feed information to AI systems, much better pivot would be to create a AI-based algorithm that would feed information to human analysts using their knowledge to refine and fine-tune itself. I called this new approach NewsdAI, derivative of Newsdaq but I believe a much better name.More on the development of the NewsdAI in the next blogs...
Comments
Post a Comment