Most deceptive problems are difficult for genetic algorithms (GAs). This is because a GA changes its search direction based on the building block hypothesis. This paper shows that a GA with neutral mutations, which is proposed for nonstationary optimization problems, is also an effective approach to solving deceptive problems. Since the GA with neutral mutations adopts a redundant string representation, most genetic operations become to be neutral for an objective function. Another novel characteristic point is that the GA has various genetic operations for operons, which are defined as the clusters of genes, in order to bring about adaptive genetic changes, according to the state of the genetic search. This paper modifies the GA with neutral mutations on the following three points: (1) The value lists in the genes have the same fixed lengths. (2) In order to maintain the consistency to the above rule, the tail values in the lists are deleted in the case of applying a duplication or inserted in the case of applying deletion. The inserted values are determined by a predefined rule for avoiding the fixations of genes on specific values. (3) The population is divided into subpopulations only at the stage of selection and reproduction for maintaining the appropriate diversity. The optimization ability of the proposed method is examined with four well-known deceptive problems. It is observed that it overcomes deception by the emergent property of finding deceptive hyperplanes and escaping the population from them by using the large genetic transitions to their complements.