Author: martin fierz
Date: 07:14:27 07/20/04
Go up one level in this thread
On July 20, 2004 at 05:36:08, Uri Blass wrote: >On July 20, 2004 at 04:54:30, martin fierz wrote: > >>On July 20, 2004 at 02:44:33, Roy Brunjes wrote: >> >>>Earlier I posted about this position with the question "Is the exchange >>>sacrifice 13. ... Rxf4!? sound?" >>> >>>[D] r1b3k1/pp2n1b1/1qn1prpp/2pp4/5B1P/2PP1NP1/PPQ2PB1/RN3RK1 b - - 0 13 >>> >>>Some readers thought it should be, others were not so convinced they would play >>>it themselves in a game (not necessarily a vote for or against the soundness, >>>just that they might not play that move themselves in a game). >> >>just for the record, the strong players liked it very much, while the weak >>players doubted it's soundness... >> >>cheers >> martin > >As far as I see movei of today has no chance to find it. >It can see enough compensation but it evaluates alternatives as better for black >and even after Rxf4 exf4 Bd7 it can see only +0.27 for black at depth 15 when >the score before Rxf4 was near +0.4 for black. > >I looked now at older posts about this position and there is something that you >said that I disagree about it. > >You said: >"if even the computer sees enough compensation, it >probably is more than enough." > >My response: >If this is the case then a simple improvement for the computer is to increase >the positional scores. no,no, that is not a simple improvement. probably they would go and sac material in lots of positions where it shouldn't be sacced. my statement is absolutely fine; i have analyzed a lot with computers, and the general observation is true: if the computer sees enough compensation for the material, then the sac will turn out to be good. usually you then only have to let it search a bit deeper until it sees that it's good rather than equal. i'm not the first to make this observation, a lot of my friends who use computers to analyze have said the same... cheers martin
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.