Ballistic protection requirements often call out a minimum probability of no penetration against a projectile fired at a specified velocity with a required confidence level. For example, 90% confidence of 90% probability of protection at 2,900 fps. This paper describes a potential testing algorithm to maximize confidence in the resulting test data, and validation of the algorithm through Monte Carlo simulation. The algorithm uses a two-step process – a brief initial V50 test, followed by testing at velocities calculated to maximize the likelihood of validating the ballistic requirement within a given number of shots. The algorithm is based on straightforward confidence calculations based on the generalized likelihood ratio, does not require the generation of a zone of mixed results, and eliminates the velocity tolerance commonly used to discriminate valid and invalid shots. Monte Carlo simulations indicate that the algorithm may bias the confidence calculations.