Author: David Rasmussen
Date: 02:18:04 04/06/01
Go up one level in this thread
On April 05, 2001 at 22:23:17, Robert Hyatt wrote: >On April 05, 2001 at 18:41:37, David Rasmussen wrote: > >>What I really don't understand, is how you can be sure that you have a high >>hamming distance in Crafty. You don't check for it, you just generate in one >>pass. Maybe you have done some work to find a seed value that generates a good >>sequence. Is that it? In that case, you're lucky. Most random sequences will >>have a minimum hamming distance of ~15. Is that high enough? > > >What you don't realize is that I tested several "seeds" and did test the >hamming distances between every pair of random numbers. I picked the seed I _do_ realize that, because that was what I meant by "done some work" :) >that gave the best overall result. Then I no longer need to test since the >RNG produces the same stream of random numbers each time the program starts >up. :) > Of course, that was what I meant. Why else would one want to find a good seed? >a hamming distance of 16 is not very good for 64 bit values... It is very hard to get above 16 with just a random sequence. If, one the other hand, one keeps generating a particular needed random number until it fits all the previously generated numbers in having some predefined minimum hamming distance, then one can relatively quickly get to 23. But thats not one streak from a given seed. That is very unlikely. Do you remember what your minimum hamming distance is with the crafty numbers? I can of course check myself in 5 minutes.
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.