I've suspected this for a long long time - when I was a young kid (around 5 years old) my oldest sister was at university studying psychology with an emphasis on child development - so for years I got bombarded with "intelligence" tests to the point where I could do them very easily and I used to get extremely high results in IQ tests in my teens. However, I've never thought I was particularly brighter than anyone else, just that I had done a lot of these silly tests and had acquired the skill of doing them.
I've thought about this quite a bit (I was motivated by those experiences to go into AI research - which I did for six years) and I'm not convinced that the kinds of problems used in the tests I did map too well to "real world" problems - in essence all of the difficulties have already been abstracted away. Of course, I have no evidence for this but the fact that the logical/symbolic approaches to AI failed to live up to their early promise more or less supports my views.
Did he acquire the ability to spot patterns and connections, or did he just spot the patterns and connections by getting used to the tests? It's very possible that the ability to identify patterns was an innate part of him, an "intelligence" he already possessed, which allowed him to benefit from repetition.
