Re: [GP] Re: quasi-random numbers?
- On Apr 30, 2010, at 10:30 PM, hoai nguyen xuan wrote:
> Essentially, the purpose of of pseudo-random numbers and quasi-If I may deviate from the discussion at hand: but isn't it the case
> random numbers are different. The aim of the former is to try to
> look random, i.e. given a historical sequence of numbers generated
> by a pseudo-random number generator, it is hard to predict what is
> the next in the sequence.
that predictability is perfectly acceptable for standard (non-crypto)
RNGs? For example, Mersenne Twister's output is quite predictable
given previous values in the sequence. Rather, standard RNGs should
produce statistically random, uniformly distributed sequences, correct?