Two experiments were carried out in order to test the effects of task sequence on the speed of rule learning and problem solving. Expt. I involved numerical rule‐learning tasks. One group of subjects was presented with the tasks using a simple‐complex sequence while another group was presented with the complex rather than the simple task initially. The simple‐complex order resulted in a far superior performance. Expt. II tested the effect of task difficulty and task precedence using problem‐solving tasks. Speed of problem solving was not found to be significantly affected by task order.These contrasting results were discussed in terms of hypothesis testing theory and learning hierarchies.