This paper discusses new implementations of the predictive alternate test strategy that exploit model redundancy in order to improve test confidence. The key idea is to build during the training phase, not only one regression model for each specification as in the classical implementation, but several regression models. This redundancy is then used during the testing phase to identify suspect predictions and remove the corresponding devices from the alternate test flow. In this paper, we explore various options for implementing model redundancy, based on the use of different indirect measurement combinations and/or different partitions of the training set. The proposed implementations are evaluated on a real case study for which we have production test data from 10,000 devices.