A brilliant insight just came to me:
Achieving true market prediction means having the capacity to create and predict the behavior of an artificial intelligence.
Here are the reasons why:
1. The description of AI from Caprica suggests consciousness can be built given a sufficient amount of recorded on-line activity.
2. If you know how the parts of a market works, you can know how the whole works.
3. A person is a part of the market.
4. How a person interacts on-line is part of a market, e.g. people build and use a system for products use.
5. An AI can imitate on-line market behavior such that it’s indistinguishable from a person.
6. If the program allows for predictive behavior, then you can predict a part of market behavior.
7. Since you can predict the part, you can predict the whole.
8. Therefore, in order to predict how the market will work, you need an AI.
This implies that the more AI-like or better your modeling, the more you can predict how the market will behave.
If AI proves to be un-predictive behavior, then perfect market prediction is impossible. (It might just be quantum in nature.)
If AI proves to be impossible, then perfect market prediction is impossible. (See the Chinese Room Argument.)