this compensates for known Random bug (just ask 'Morduun' )
I searched Morduun's posts and could only find that s/he had problems with sets of random numbers from the same seeds not being reproducible on some hardware (i.e. the same program would run differently on some machines.)
From what I read, Morduun ported a Mersenne Twister algorithm to Blitz as a solution to this problem. I didn't see anything about reseeding being necessary or the first call to Rand() not being random.
Rook, please cite something.