The research and practitioner literature has generally established that frequent, well-publicized sobriety checkpoint programs produce reductions in drunk driving and related social ills through increasing the perceived risk of apprehension in the general population. As such, the accepted measure of whether a checkpoint program ‘works’ is general deterrence. However, infrequent sobriety checkpoint programs do not produce general deterrence and there is no established alternative for defining the effectiveness of such programs. Given that infrequent checkpoint programs are a fiscal reality in many locations, we argue for a simple specific deterrence metric for monitoring the impact of such programs. Specifically we compare the mean number of Driving Under the Influence arrests on checkpoint days versus non-checkpoint days. We use data from a police department in the Upper Midwest to demonstrate the practical use of such a metric and discuss the implications for maintaining such programs or shifting resources towards saturation patrols.