Curving for national exams is done at the population level. Yes, there may be differences in ability from year to year in individual classes in individual schools, but at the population level differences are negligible.From my recollection, exams were pretty much of a muchness. I don't recall seeing past papers that were excessively hard or easy (we did try past papers during each course)
I see no reason why 10% must pass at an A grade. This would imply that someone, one year, who should have received a very good B grade might receive an A, and conversely, someone who should have received an A received a B.
Given that there are differences between each year's ability in different subjects, surely there must be years when 9, 11, 8 or 12% should pass...
"10% pass at an A grade" was an example I used for simiplicity; the actual curving calculations (as I said) are more complex and I'm not sure of the exact algorithms used by the examinations board. At universities it works slightly differently (although the basic principles are the same) because universities don't examine at the population level.