Re: Unpredictability in self-recursive systems

From: Anders Sandberg (
Date: Wed Oct 03 2001 - 06:51:38 MDT

On Wed, Oct 03, 2001 at 02:37:54PM +0800, Chen Yixiong, Eric wrote:

> I hereby propose the Recursive Complexity theorem: "In any recursively
>complexity system, the amount of system chaos increases with the
>accuracy of the prediction system used by different participants
>coordination in such a way as to nullify any long term benefit of using
>the more advanced prediction system."

I hope you are aware that theorems are defined to be rigorously derived
statements in a system, you cannot just propose them because they seem
likely. The above statement is a conjecture (like the Goldbach

That said, you have merely rediscovered something that has been studied
in economics for a long time (where there *is* a theorem to roughly this
effect). However, you might want to read up on the papers that led to
the foundation of The Prediction Company and the field of applying
learning systems to economics - there are interesting loopholes due to
lack of information that can give advanced prediction systems at least
some profit.

Note that you implicitly make the assumption that the agents are playing
a zero-sum game in your weather and business example. If it is a
positive-sum game you instead get stabilisation out of the adaptation.

Anders Sandberg                                      Towards Ascension!                  
GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y

This archive was generated by hypermail 2b30 : Sat May 11 2002 - 17:44:11 MDT