Author: Thorsten Czub
Date: 07:39:49 03/04/00
Go up one level in this thread
On March 04, 2000 at 09:27:44, blass uri wrote: >How did you compare it? by watching games. >I think that the right way is to give the same positions from your games of >Fritz6a that you guess that Fritz6a is weaker in tactics to Fritz5.32 and see if >Fritz5.32 can see tactics faster. NO- test positions with definitive key moves do not measure this. what test-suites measure is how good or how many extensions and search tricks for key-moves programs have. but these nice magic tricks do not work in chess-games. they only work for test-suites. >Impression based only on watching games can be wrong because it is possible that >Fritz6a was unlucky to get positions that Fritz did not like in your games. when a program more than another program loses tactically ... >I know that Enrique claims that Fritz6a is the best in his secret tactical test >suite. AND ?? this is not against me, this only proofs my point. in test-suites fritz6a may be better, but a chess-game is more than a test-suite. don't you understand the difference ? >If you have positions from your games when Fritz6a is weaker in tactics then >please post them. you don't understand my point. you want to measure tactical behaviour and mix it up with finding key-moves faster. these are 2 different topics. hiarcs e.g. is good in finding key moves. shredder is weak in finding key moves. cstal is good in finding key-moves but weak in tactics itself. the king is good in both. the king 2.2 was good in tactics. the king 2.54 was better in finding key moves but weaker in tactics than the king 2.2. MM5 was good in finding key-moves. better than many other 8-bit dedicated machines. but not very good in tactics at all. same for fidelity machines. >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.