Confirmation bias

Seeking and weighting evidence that supports what you already believe, and discounting what does not.

It is the most damaging bias in research work because it operates invisibly and feels like diligence. An analyst who has written a positive thesis reads the next disclosure looking for confirmation and finds it, because most disclosures contain something for everyone.

The structural defences are the ones that work: writing the bear case before the bull case, seeking out the best argument against the position rather than the weakest, and recording in advance what evidence would falsify the thesis.

Find the person who disagrees and understand why.