A Mechanism for Pockets of Predictability in Complex Adaptive Systems

TL;DR

This paper identifies a mechanism in complex adaptive systems that creates transient pockets of predictability, where agents' actions become temporarily decoupled from historical data, demonstrated through synthetic and real financial data.

Contribution

It introduces a new mechanism explaining the emergence of predictable periods in complex systems, supported by empirical testing on financial time series and synthetic models.

Findings

Frequent prediction days indicate collective agent organization.

Transient herding regimes lead to pockets of predictability.

The mechanism is validated on both synthetic and real data.

Abstract

We document a mechanism operating in complex adaptive systems leading to dynamical pockets of predictability (``prediction days''), in which agents collectively take predetermined courses of action, transiently decoupled from past history. We demonstrate and test it out-of-sample on synthetic minority and majority games as well as on real financial time series. The surprising large frequency of these prediction days implies a collective organization of agents and of their strategies which condense into transitional herding regimes.