Author: Robert Hyatt
Date: 06:57:59 10/29/03
Go up one level in this thread
On October 29, 2003 at 05:07:31, Dan Andersson wrote: > What you are saying isn't totally correct for most PRNGs. Even if you use the >Mersenne Twister the number of bits that you can safely use decrease if you just >combine the output of a series. In the MT case I think the number of safe bits >decrease to about 53. But this is very paranoid limits. > >MvH Dan Andersson \ You are thinking too big. We only need 12 * 64 random numbers for Zobrist hashing. It won't matter whether you generate 768 64-bit numbers, twice as many 32 bit numbers and concatenate them, or even 49152 single-bit random numbers and combine 'em 64 at a time. The sample size is way small. If you are throwing darts with a Monte Carlo trial, I'd be more concerned when you start tossing 2B darts. But not 768.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.