While the value of unidirectional delay was used as the preference of a node in the previous research on the performance of the leader, this paper proposes the value of bidirectional delay as a preference of a node and analyzies election strategies in the case of the various number of candidates through simulation. We made a comparative study to evaluate the quality of leaders chosen by the bidirectional preference and unidirectional preference in terms of relative error and from the standpoint of the probability of choosing the optimal leader.