Not much wisdom in crowds of stockbrokers

As Farmer and co gently put it,

The traditional paradigm in economics is one of rational utility maximizing agents. Recognizing limitations in human cognition, economists have increasingly explored models in which agents have bounded rationality. We take this direction even further here by testing a model of trading in financial markets that drops agent rationality almost altogether.

They show that, while many very clever individuals work in the financial sector, much of their effect in the aggregate is to cancel each other out. The researchers want to eliminate this noise from the analysis to find out how much rational agents actually contribute and where the asymmetries with profound effects are. This somehow reminds me of the early universe, in which a lot of matter and antimatter in proximity resulted in much banging and flashing and afterwards a small amount of leftover matter (the physically observable universe) turned out to be terribly important.