anthony@utcsstat.UUCP (01/04/85)
I recently did some statistical tests on the random numbers
produced by an IBM PC, Apple II+, and a Commodore 64. I had each
machine generate 1,000,000 random numbers which where then transformed
to the digits 0 through 9 (inclusive). The statistical tests that
followed indicate that the Commodore 64 and Apple are horrible when
producing supposedly (pseudo) random numbers. The C64 was worse than
the Apple. The numbers produced by the PC where almost too good to be
true.
I guess any games that rely on random number generation (and
other software for modelling, simulation etc.) are not as unique each
time as one may be led to think.
Any comments by hackers out there ???
--
Anthony Ayiomamitis {ihnp4|decvax|utzoo|utcsrgv}!utcs!utzoo!utcsstat!anthony