Using neural networks, Belokurov, Evans & Le Du (2003, 2004) showed that 7 out of the 29 microlensing candidates towards the Large Magellanic Cloud (LMC) of the MACHO collaboration are consistent with blended microlensing and added Gaussian noise. They then estimated the microlensing optical depth to the LMC to be between 0.3 x 10^{-7} and 0.5 x 10^{-7}, lower by about a factor of two than that found by the MACHO collaboration. There has been a recent independent claim of a low optical depth to the LMC by the EROS collaboration, who find 0.15 x 10^{-7} from 3 candidates (Tisserand et al 2006). Griest & Thomas (2005) have contested our calculations. Unfortunately, their paper contains a number of scientific misrepresentations of our work. We stand by our application of neural networks to microlensing searches and believe it to be a technique of great promise. Rather, the main cause of the disparity between Griest & Thomas and Belokurov et al. lies in the very different datasets through which these investigators look for microlensing events. We don't exclude the possibility that some of our non-microlensing designations may change with access to clean data. Whilst not everything is understood about the microlensing datasets towards the LMC, the latest downward revisions of the optical depth means that Griest & Thomas' microlensing puzzle is a roughly 1 sigma effect. Efficiency calculations can correct for the effects of false negatives, but they cannot correct for the effects of false positives (variable stars that are mistaken for microlensing). Therefore, the best strategy in a microlensing experiment is to eschew a decision boundary altogether. Rather, each lightcurve should be assigned a probability and the microlensing rate calculated by summing over the probabilities of all such lightcurves.