Author: martin fierz
Date: 22:42:55 10/17/02
Go up one level in this thread
On October 18, 2002 at 00:47:48, Robert Hyatt wrote: >On October 18, 2002 at 00:35:38, Robert Hyatt wrote: > >>On October 17, 2002 at 23:35:58, martin fierz wrote: >> >>>On October 17, 2002 at 04:51:28, Christophe Drieu wrote: >>> >>>>[D] 8/4k3/4p3/pp5p/6p1/2P2rP1/PPK2R1P/8 w - - 0 1 >>>> >>>>Rxf3 ! >>> >>>for all those who find Rxf3 in the first position, how about a slightly modified >>>version: >>> >>>[D] 8/4k3/4p3/1p5p/p5p1/2P2rP1/PPK2R1P/8 w - - 0 1 >>> >>>this time, Rxf3 will not win any more. if your static eval just more or less >>>"randomly" thinks the first position is a win for white after Rxf3, does it also >>>realize that the second one is not a win? >>> >>>i think this also explains quite well why the first position is difficult for >>>engines: since the second, *very* similar position is not winning, you will have >>>some trouble to formulate a rule why the first is and the second is not winning >>>:-) >>> >>>aloha >>> martin >> >>\ >>Crafty for the first position, static eval is +.63, for the second -.30.... It >>likes Rf3 in >>the second but the score is only slightly positive for white. At 11 seconds it >>drops to >>draw score, and at 48 seconds, Rf3 is slightly negative. At that point, it >>still plays the >>move as best, but the score hovers around the draw score, just on either side... > > >I should note that at depth=18,. Rf3 failed high to +1.0... whether that is >significant or >not is a question for another day, it is late here. :) > > 17 5:45 0.00 1. Rxf3 gxf3 2. Kd3 e5 3. b3 axb3 4. > axb3 e4+ 5. Ke3 Kd6 6. h3 Kd5 7. Kf2 > Ke5 8. Ke3 Kd5 > 17-> 11:52 0.00 1. Rxf3 gxf3 2. Kd3 e5 3. b3 axb3 4. > axb3 e4+ 5. Ke3 Kd6 6. h3 Kd5 7. Kf2 > Ke5 8. Ke3 Kd5 > 18 12:24 ++ 1. Rxf3!! > 18 14:55 0.91 1. Rxf3 gxf3 2. Kd3 e5 3. b3 e4+ 4. > Ke3 Kf6 5. bxa4 bxa4 6. Kf2 Ke5 7. > h3 a3 8. Ke3 Kf5 9. c4 Ke5 10. c5 Kd5 either crafty is missing something or i&fritz are - why not 4. ...axb3 and Ke6? aloha martin
This page took 0.42 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.