From: mhx@iae.nl   
      
   On Thu, 19 Sep 2024 0:28:33 +0000, Krishna Myneni wrote:   
      
   > On 9/12/24 21:10, Krishna Myneni wrote:   
   [..]   
   > A bit more involved test of the same 64-bit PRNGs starts to show the   
   > possibility of defects in Marsaglia's 64-bit kiss prng (RAN-KISS)   
   > compared to a simple 64-bit LCG prng (RANDOM).   
   [..]   
      
   I wonder if it is at all possible to really prove something   
   about the PRNG *with tests of this type*. Intuition wants us to   
   believe that the longer we run the simulation, the closer   
   the result must match the expected outcome. Shouldn't we   
   compute the probability that after a certain size run the   
   result does NOT match the known result (given an ideal PRNG),   
   or how unlikely it is that the result has a given error?   
      
   Example: say the result of PRNG-a consistently has one of   
   its bits (say bit 0) stuck at zero. Would the test under   
   consideration detect this specific problem at all?   
      
   -marcel   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)