Behavioural ecologists have built complex models of optimal behaviour in simple environments. I argue that they need to focus on simple mechanisms that perform well in complex environments. This is, however, a difficult area for a modeller in that various modelling choices are not obvious. In particular, what complexities should a model world include and what sorts of rules should be considered? In my talk I consider an approach that is more limited but which sidesteps these difficulties; I will take rules that have evolved and ask what aspects of the environment make these rules ecological rational. I argue that various psychological phenomena can only be understood in adaptive terms if the environment in which the rules evolved was sufficiently complex. For example, rules may only make sense in adaptive terms if there are specific spatial or temporal heterogeneities. I illustrate this general theme using examples that show that lack of transitivity, contrast effects, aspects of prospect theory, and other behavioural phenomena which are often regarded as irrational can all have adaptive explanations.