What little level-of-difficulty difference there is from one exam
to the next is offset by adjustments to the score curve. A tougher
exam gets a more "forgiving" score curve, where it takes
FEWER right answers to get a given score. A statistically easier
exam compensates with a tighter curve, where you must get MORE right
answers to attain the same score.
And no, you're not being assessed against the other people
taking that exam on that date -- so it's not as if smarter
people taking a given exam date hurts you or lower-scoring people
taking another exam date gives you a hidden benefit. The score curves
are established statistically, before the exam is even given, by
the performance of individual questions in pervious years'
Finally, not even your %-ile ranking is measured against your own
exam date's test population. If you notice, they always label
that %-ile ranking against a running tally of all the people who
took the test over the THREE YEARS preceding your own exam date.
There is ONE little exception to all this, though and that is:
the February exam is generally not published, so you don't get to
see your own performance breakdown as you would on any other exam
date. But February is such an off time anyway, this is usually not
a big deal.