Author: Uri Blass
Date: 02:25:15 10/27/02
Go up one level in this thread
On October 26, 2002 at 20:36:30, Dana Turnmire wrote: >The Genius programs have always been known in the past as top notch blitz >programs. Mr. Lopez posted that Ruffian had beaten Genius 6.5 by a score of +46 >=4 -0. At first I thought there had to be something wrong with the way he had >set the test up but now I believe his results could be accurate. > >I ran a 10 game match between Chessmaster 9000 and Genius 7 at Game/90 in which >Chessmaster only won by one point (+3=5-2). I have just finished a 10 game >match between the two programs at Game/5 with Chessmaster winning by a score of >+6=2-2. I assume Genius is basically the same engine as 6.5 but what I don't >understand is why would the Genius program now be stronger at longer time >controls on faster hardware when it was just the opposite just a few years ago >on slower machines? All my games were played with the default settings with >pondering off and each machine having 64 meg hash. Maybe Genius suffers from the big hash tables at blitz. Maybe Genius can perform better if you reduce the size of it's hash tables. I do not know and it is only a suggestion. The results may be also a statistical error. Chessmaster9000 is not weaker than Ruffian. +6 =2 -2 can happen also between equal programs so the result is not strange. +46 =4 -0 is result that it's probability to happen is so low that I simply cannot believe 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.