Er...do you mean the subjects with less students will get scaled a bit more?
As long as it's a hard subject, the criterion on the number of candidates doesn't affect much, I think. E.g. If everyone, let's say, chooses Ext2 Maths or English Ext2 (far out! it's my dream), however, around 90% fail (band E2 or under), then the 10% will still get scaled up like... Wheeeeeeee...especially ones who get E4. The same scenario occurs to Advanced English. You do well, scaled up, do badly, scaled down still.
However, the BOS is smart enough to recognise the improvements of candidates over each years or so and thus, they have some adjustments on the levels of difficulties for a particular subjects in order to achieve the most accurate results of each student's ability. Quite obviously we can see that after each year, students get more aids, more materials, more info, tips, etc. to deal with the HSC. The BOS doesn't like it in some way. That's why they have to change the syllabus every 6 years or st like that.
To unwind this, get an example: 2006: 10% get band 6 and non get under band 3 or less (don't forget it's an example), 2007: BOS gives a harder test: 5% get band 6 and 10% get 3 or less. This shows a clear range of capabilities, right? They are trying to find out the accurate number of those with higher ability and those with lower abilities. This doesn't mean if you get band 4 this year, you are better than band 3 last year!