Intensive statistical complexity measure of pseudorandom number generators

A Statistical Complexity measure has been recently proposed to quantify the performance of chaotic Pseudorandom number generators (PRNG) (Physica A 354 (2005) 281). Here we revisit this quantifier and introduce two important improvements: (i) consideration of an intensive statistical complexity (Physica A 334 (2004) 119), and (ii) following the prescription of Brand and Pompe (Phys. Rev. Lett. 88 (2002) 174102-1) in evaluating the probability distribution associated with the PRNG. The ensuing new measure is applied to a very well-tested PRNG advanced by Marsaglia.