Reciprocal translocation is a nature-inspired crossover operator that is proved to be a useful alternative for the conventional multipoint and uniform crossover operators. This paper compares the performance of reciprocal translocation against the well-known conventional crossover operators for the solution of provably difficult GA-deceptive functions. This study also aims to discover the linkage learning capability of reciprocal translocation since it is an important requirement to be successful on GA-deceptive functions. A number of well-known hard GA-deceptive functions are considered for experimental evaluations and several GA implementations with reciprocal translocation and other conventional crossover operators are used for their solutions. The obtained results show that the performance achieved with reciprocal translocation is much better than that of any other conventional crossover operator.