Re: [computer-go] Re: Sharing Secrets

[Don Dailey <drd@xxxxxxx>]

GCP is  right.  Usually the "birthday paradox"  calculation is applied
to determine  the chance of  a collision, but  what we really  want to
know is  the chance that there will  be a collision AND  it affect the
root move choice.  This extra constraint  is kind of like having a few
more bits of key.

- Don


   X-Original-To: computer-go@xxxxxxxxxxxxxxxxx
   Date: Thu, 21 Oct 2004 18:06:50 +0200
   From: Gian-Carlo Pascutto <gcp@xxxxxxxxxxxxxxxxx>
   X-Accept-Language: en-us, en
   Reply-To: computer-go <computer-go@xxxxxxxxxxxxxxxxx>
   Sender: computer-go-bounces@xxxxxxxxxxxxxxxxx
   X-Scanned-By: MIMEDefang 2.42

   Frank de Groot wrote:

   > And, BTW, a 1::65556 chance of collision would be terrible in many cases.
   > When searching millions of tree nodes and a hash collision can mean the 
   > difference between pruning or not, you'll find such a hasher unacceptable.

   This is completely false. It's been demonstrated in the computer chess 
   community that a tree searcher can have an enourmous amount of hash 
   collisions and still not have a noticeable decline in playing strength.

   It's a result of the resilence of tree searching against single-node 
   errors. (Which is why the programs are so strong despite the evaluations 
   being so wrong so often.)

   The issue is different when you're using the hashes outside the tree 
   search. (e.g. fuseki library that's probed at the root, a single mistake 
   could kill you)

   -- 
   GCP
   _______________________________________________
   computer-go mailing list
   computer-go@xxxxxxxxxxxxxxxxx
   http://www.computer-go.org/mailman/listinfo/computer-go/

_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/