The commonly used penalty methods (such as Varicomp or varying the thickness of the out-of-line barriers), for determining the detonation transfer probability for an explosive interface, may become less practical or convenient the more insensitive explosive is the acceptor. The method proposed here suggests replacing the design donor by an explosive (substitute donor) of a lower detonation pressure. Both the acceptor and the mechanical hardware of the test system should be identical to those of the design system. It is assumed that the detonation pressures of both the design donor and the substitute donor are normally distributed. The same assumption holds also for the threshold detonation pressure required for initiating the acceptor. Following these assumptions we present a method to determine a minimum number of tests required to demonstrate a given reliability at a given confidence level. The relative standard deviation for each detonation pressure is assumed to be equal for all donors. We show then that the required number of tests to demonstrate a given reliability (at a given confidence level), practically depends on the ratio of the detonation pressures of the design donor and its test substitute. A similar method can be applied to demonstrate low probability of detonation transfer (high safety), by replacing the design donor by one of a higher detonation pressure.