The paper develops a theory of biases in decision making. Discovering a strategy for solving a game is a complex problem that may be solved by decomposition; a player decomposing a problem into many simple sub-problems may easily identify the optimal solution to each sub-problem: however it is shown that even though all partial solutions are optimal, the solution to the global problem may be largely sub-optimal. The conditions under which a decomposition process gives rise to a sub-optimal solution are explored, and it is shown that the sub-optimalities ultimately originate from the process of categorization that governs the creation of a decomposition pattern. Decisions based on a strategy discovered by decomposition are therefore frequently biased . The persistence of biased behaviours, observed in many experiments, is explained by showing the stability of different and non optimal representations of the same problem. An application to a simplified version of Rubik cube is finally developed.