Is Benford's law a good instrument to detect fraud in reports of statistical and scientific data? For a valid test, the probability of 'false positives' and 'false negatives' has to be low. However, it is very doubtful whether the Benford distribution is an appropriate tool to discriminate...