Author: Laurence Chen
Date: 20:37:16 09/25/99
Go up one level in this thread
On September 25, 1999 at 23:02:59, Marian Predescu wrote: >After > >1.e4 Nc6 2.Nf3 e5 3.Bb5 Nge7 >(Cozio Defense) >4.c3!? d5 >(I assume that this is theory. I can't recall anyone playing this on move 4 >against me before) > >My question i, how come Fritz 5.00 doesn't consider 4.c3 at all, when it is the >second more popular move according to the databases? > >Details: > >Old Fritz 5.00 itself likes other moves, not 4.c3, but: > >+=(0.31)4.0-0, or >+=(0.28)4.Nc3, >=(0.25) 4.d3, >=(0.16) 4.Qc2 > >After 4.c3 it gives: >=(0.19)4...d5 >=(0.19)4...a6 > >On a depth of about 14/40... > >b) An opening tree of the Ruy Lopez Cozio old database from Univ of Pitt shows >that actually that out of 406 games > >4. 0-0 was played in 202 games with >1-0 37% >1/2 29% >0-1 32% > >But the next most popular is: > >4.c3 in 90 games with > >1-0 41% >1/2 34% >0-1 24% > >4. Nc3 in 54 games >1-0 51% >1/2 31% >0-1 16% > >4. d4 in 53 games >1-0 32% >1/2 35% >0-1 32% > >Found more games in another database(added them to the tree), so this time out >of 1059 games, but... nothing changed much(maybe there was some duplication). > >Who is right in a case like this: Fritz or the databases? > >Just a question from the chess point of view, about the value of "infinite >analysis" of chess software like Fritz... > >How much is one supposed to read in either the computer or the database? The answer is database. The computer opening book may omit such moves because it may not lead to favorable positions which the chess engine would excel. All chess engines suffer from the horizon effect, and they are not very good in long strategical planning. So when using the infinite time to evaluate the position the evaluation will shift and depending on the position, if it's a non-tactical position, it may be very difficult for the chess engine to find the correct plan. Solution to the problem is to play the position out. Laurence
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.