Author: Uri Blass
Date: 22:47:08 10/30/01
Go up one level in this thread
On October 30, 2001 at 20:36:49, Dann Corbit wrote: >On October 30, 2001 at 20:04:41, Uri Blass wrote: >[snip] >>Deeper blue did not use null move pruning and I believe the algorithm of the >>chess programs of today is superior(every top program that I know except Junior >>is using null move and even Rebel started to use null move pruning). >> >>Junior also has other pruning rules. >> >>I do not believe that the deeper blue team were right in not using null move >>when all the programmers of today are wrong. > >Some time ago, there was an America's cup race between New Zealand and the US, >with the US using a [cough] Catamaran. Seems rather unfair, but -- there it >was. > >The New Zealand skipper said something like, "If he's going to race a catamaran >against a single hull, he ought to go all out." > >Connors replied that he was just being extra careful with the equipment. > >The point is, when you have a ludicrously unfair advantage, you can enjoy an >enormous safety margin. > >Surely, Deep Blue could have searched a lot deeper with a nullmove heuristic. >But then, it would have been prone to a large class of errors that it was not >prone to with ordinary Alpha-Beta. > >Maybe, with nullmove, Deep Blue would have won the match 7-0. Or Kasparov might >have found a tactical blindside and won the match 7-0. Playing it safe was the >smart thing to do. > >Why take a risk when you don't have to? not using null move pruning is taking a risk. null move pruning does not mean playing unsound moves when the time to calculate is not important because it is possible to use zunzwang detection. Some programs like yace already do it. Uri
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.