Confabulous, the difference between your experience and GAPAN's is that you only have experience of one candidate; they have experience of thousands. The statstical validity of your sample is non-existent; theirs is highly significant. In other words, if it came to deciding whose experience to rely on, you wouldn't feature highly in the shortlist.
That said, the trouble with any large-scale testing is that it tends to produce findings that rely on generalisations - well-founded ones, but generalisations nonetheless. As with all generalisations, there will be exceptions. That doesn't make the generalisation untrue.
GAPAN's tests are the same as those the RAF use, and they have a very high correlation with success in that kind of training. However, very few commercial students undergo a course of training that is as intensive and unforgiving as the RAF's, which means that a lower-aptitude student may well still pass. In fact, with the less-intensive course structure, the less able (by GAPAN's methods) student may well come good. In any case, when you are paying for the course (with some exceptions), many schools will allow you to keep throwing money at them until you pass! As you are the paying customer, how many FIs are going to tell you that you are below average? Not a recipe for continued business success...
Airlines, on the other hand, are more like the RAF. They don't want below-average people on their flight-decks. Their initial type training is intensive and has little flexibility for those who don't come up to speed quickly. If they don't have their own training system, they may not have a choice - until the first linecheck, anyway - but they will always try to get the best people they can. Some use the GAPAN system themselves, others use another way of testing aptitude before a candidate starts training. Such methods are inappropriate for those who've already undergone flying training, but sim checks and other checks achieve the same end.
At the end of the day, these things are just a guide. They are valuable when used in the right way. There is nothing wrong with an assessment of average; after all, you don't have to reveal that assessment to anyone if you don't want to.
Scroggs