The Economic literature sees institutions as constraints that shape human interaction and limit the set of allowed actions. Institutions are also described as devices that increase the capability of an agent to predict social behaviors, and, in this way, they allow more flexibility. Strategic uncertainty is, so, at the origin of institutions. Following this line, the present paper aims at understanding what happen to individual behaviour after institutions are established. The central question is how an institution can reduce uncertainty and at the same time allow flexibility. To answer this question, the paper takes into consideration how persons perceive institutions. The starting point is the model proposed by Heiner (1983). This model is based on the reliability condition, a relationship that allows to understand whether an agent can expect to benefit or not from undertaking an action that is novel in respect to a set of already employed options. In Heiner’s view, agents display new behaviours in simple or well known situations, but in uncertain contexts (when the gap between the skills required to perform a task and its difficulty - labeled competence-difficulty, C-D, gap – is high), they consider just few variables, well know and tested in the past, and rely on previously employed strategies. Whereas the received microeconomics theory links predictability with simplified environments in the presence of certainty or risk, Heiner, therefore, see predictability as a by-product of uncertainty. This article applies Heiner’s model to a dynamic situation, in order to study how predictability and stability of behaviors are affected by the set up of an institution. In this perspective, the idea that institutions and rules simplify the environment and reduce the C-D gap is just a starting point. If rules arise because of the need for reliability in an environment with an elevate C-D gap, when an institution simplifies the environment, there could be no more need to follow the rule. In fact, with a lower C-D gap, innovative choices might become reliable. In this way, a paradox comes out. Institutions emerge to reduce uncertainty, but, in reducing uncertainty, they make themselves less necessary and so they increase uncertainty by means of the higher flexibility allowed. Considering this paradox, the present work tries to analyse if institutions can determine a real simplification of the environment, or if they are devices that make uncertainty more manageable, but not necessarily reduced.