Author: Thorsten Czub
Date: 01:55:33 10/31/04
Go up one level in this thread
On October 31, 2004 at 04:14:50, Kurt Utzinger wrote: >No results, comments on how the latest >Chess Tiger is doing? >Kurt one reason is maybe that the results are not that good. it fails to beat s8 e.g. it might be stronger in some test-suites (wm-test e.g.) but this does not bring a much better play. also i am missing all the nice switches i can setup with pro deo. in ct2004 i can only change the style roughly, and the problem is IMO that the other styles far from normal are not really something that brings a different or better strength. i am very disapointed that christophe did not continued gambit-tiger2. ct15 was the effort to melt tiger with gambit tiger. and ct2004 seems on the same track. a new rewrite but IMO it is not as interesting as old gambit tiger was. gambit tiger HAD something. kind of femme fatale for computerchess freaks. but i miss this behaviour in ct2004. maybe this is the reason for the silence. or it might also be that not many people HAVE ct2004 so far. if christophe could make a tigre that comes back with the BEHAVIOUR of gambit tiger, i would be very pleased. IMO the mistake was to believe that you can combine an unsound playing engine with an accurate playing engine to get more strength. this is a wrong idea. if you bring 2 people in one room, and one of them is accurate and the other is freaky, it will not result in better decisions. what was the strength of gambit-tiger ? to play moves no other program would play. to direct the game into a direction others would not have been expected because they searched for the BEST move. IMO this was what CSTal did. it did not play the best move, and instead an unsound move the opponent was not able to refute instantly, but that made the game interesting. you cannot create an engine that plays like tarrasch AND tal.
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.