I thought that the question was a bit vague and would have been fairer (not necessarily any more accurate) as two questions; splitting up the seeded and non-seeded bits. Essentially this is an argument of how random a psudo-random number is - not about a point of Sql.

But to be honest with everyone I need to work out a lot of the questions by thinking about what is actually being asked, rather than what at first glance appears to be the question! I guess that there are so many clever people here that some tricks are employed to liven things up.

My query to get an answer was:

select rand(), rand(), rand()

select rand()

select rand()

select rand(1), rand(1), rand(1)

select rand(1)

select rand(1)

As all the seeded RANDs gave the same answer I figured that the question poser was calling this deterministic; pseudo-random number generation (as we all seem to know!) isn't very random, but I think random enough to derive the answer the questioner was looking for.