This paper presents a new approach to statistical similarity assessment based on sequence alignment. The algorithm performs mutual matching of two random sequences by successively searching for common elements and by applying sequence breaks to matchless elements in the function of exponential cost. As a result, sequences varying significantly generate a high-cost alignment, while for low-cost sequences the introduced interruptions allow inferring the nature of sequences dependence. The most important advantage of the algorithm is an easy interpretation of the obtained results based on two parameters: stretch ratio and stretch cost. The operation of the method has been simulation tested and verified with the use of real data obtained from hardware random number generators. The proposed solution ensures simple implementation enabling the integration of hardware solutions, and operation based on only two sequences of any length predisposes the method to online testing.