In Constraint Programming, enumeration strategies play an important role, they can significantly impact the performance of the solving process. However, choosing the right strategy is not simple as its behavior is commonly unpredictable. Autonomous search aims at tackling this concern, it proposes to replace bad performing strategies by more promising ones during the resolution. This process yields a combination of enumeration strategies that worked during the search phase. In this paper, we focus on the study of this combination by carefully tracking the resolution. Our preliminary goal is to find good enumeration strategy blends for a given Constraint Satisfaction Problem. © 2012 Springer-Verlag.