How to treat this? Can we consider the experiment a clean? Can we trust the findings? Primarily on the findings. Yes, of course, assume that the probability of success in each trial for 50% somewhat presumptuous. Among the predicted Paul matches were those whose results were pre-clear any more or less versed in football, man. But there were others, the result of which even on the field was decided at the last minute. Estimate the probability of success predictions in each case is difficult, but even if we remove from the sample 6 correct results (half of which hypothetically include all matches whose results were clear in advance without any predictions) and leave all wrong, then we obtain the probability of accidental coincidence of 10%, which is also beyond common sense, and in most scientific papers are no longer considered to be coincidental.
Clean experience. Well, what then can be clean? Can the trainer, based out of patriotic motives, put in a jar with a German flag a tasty oyster, and other rotten? So after Paul had predicted the victory of the opponents twice in Germany and in the final match Germany generally has to do with it. This is the first. Well, and secondly, if we return to evaluating the results obtained, that the prophet we have become not Paul, and his trainer, the conclusions are not changed. You can, of course, refer also to conspiracy theory, saying the results of all matches have been pre-painted with a hypothetical 'world government', a TV crew in some sideways gained access to the records of these insidious plans and used to further bubble cut.