As far as I know, VCAA has not released their exact treatise on how they scale SACs.

They have merely mentioned that all of those above things are important.

Cohort ranking is about ranking everyone who has taken the same SAC (i.e.: in the same SAC cohort). What they then do is analyse how the cohort performs in the exam -- this will determine how harshly the SACs of that cohort will be scaled. This will vary depending on your ranking within the cohort.

The way it works roughly is supposed to be something like this:

Student A | Rank #1 | SACs: 100% | Exam: 99% => Scaled SACs: 99% (#1 Exam Score)
Student B | Rank #2 | SACs: 95%  | Exam: 80% => Scaled SACs: 90% (#2 Exam Score)
Student C | Rank #3 | SACs: 90%  | Exam: 90% => Scaled SACs: 80% (#3 Exam Score)

So the short answer to question 2: No. Your SAC ranking matters, otherwise why would you even bother to perform well in SACs, if your exam mark just represents your SAC score?

But as you can see from the example, it seems a bit unfair (student #3 getting student #2's poorer exam performance) -- and I'm betting that VCAA don't release their exact policy on this either because it is unfair, or because there may be some special algorithm they don't want schools to see (otherwise they'd try to game the system).

This algorithm most likely includes the GAT, which is said to only affect your scores positively (and never negatively). The most likely way this works is that they use your GAT to check for disrepancies (i.e.: getting an extremely low SAC but you had a high GAT score for that are) -- and if there are disrepancies, they may reconsider your case.

But in truth, this is mainly speculation with no solid sources or evidence. Hopefully other students or teachers can fill in my gaps, and point out where I'm wrong.