Probabilistic model for comparing the effectiveness of counterfeit deterrent features

As new counterfeit deterrent features are considered for new document designs, the need exists for evaluating candidate features' performance in a systematic and objective manner. To this end, an application program has been developed that is based on a probabilistic model of the counterfeiting process as determined by interview of experts in the field. The probabilistic approach attempts to capture the variability in understanding and replicating a security feature by representing the steps of the counterfeiting process with a Markov chain, and by maintaining the total cost incurred by the counterfeiter probabilistically. The application provides an estimate of the probability of understanding the security feature given the amount of resources the counterfeiter has at his disposal. The relative effectiveness of security features can be ranked based on this information, and on a scale that also provides information about the resources required of a potential counterfeiter of a note with that feature. An investigation found that the problem of determining the transition probabilities of the Markov chain is well defined. Further, sensitivity analysis shows that the model results are not highly sensitive to variation in model inputs by different users.