Yes, your last post was clear - don't worry! And I certainly don't know whether it was the GAI that made the application successful. The rest of it may have simply outweighed the various ability-test scores.
About the reverse scenario you imagine - I haven't heard of something like it, but it's certainly possible, for example in the case of an underachieving gifted kid. And I would think Davidson would take that possibility into account in reviewing applications with borderline achievement scores.
