this is one of the classic articles:

http://citeseerx.ist.psu.edu/viewdoc...=rep1&type=pdf

Human minds use heuristics to quickly estimate probabilities. K + T and others try to explain these heuristics with an argument from evolutionary biology, and show where the heuristics are biased. Many situations where our (evolved) heuristics suck are situations which are highly abstract.

In poker, doing an operation like estimating variance is very vulnerable to these heuristics and biases. Despite consciously trying to accurately estimate your variance and the accompanying BR management strategy you will almost invariably be too conservative or too aggressive based on recent history and other date which you are hardwired to overvalue in your calculations.