Author: martin fierz
Date: 13:34:08 10/05/04
Go up one level in this thread
On October 05, 2004 at 12:27:10, Uri Blass wrote: >On October 05, 2004 at 07:07:06, martin fierz wrote: > >>On October 04, 2004 at 16:51:50, Eelco de Groot wrote: >> >>unfortunately this is not a fantastic test position - white appears to be in >>serious trouble, i.e. if he doesn't do something drastic like e5 and Rxf7 he >>will simply lose. >> >>you can see that all posts which say "my engine finds it" are showing clear >>negative scores for white. > >The target in games is to find the best moves and not to find the theoretic >result. right. >programs that understand that e5 is better for white relative to the >alternatives have something positive. right again. >Note only that singular extensions are not needed and better evaluation may be >enough to find 43.e5 right again, probably speculative king safety eval makes a big difference here. but it may just be "by accident" that a program solves this position, simply because it thinks that e5 is -3.1 while Bxb4 is -3.2. of course it is good to realize that e5 is better than Bxb4, but still: the good test positions are those where a program can only make the right move for the right reason. i'm not saying this is a useless test position, but i think other positions are better test positions. besides, after looking at this again, i am not at all convinced that white has anything in this position! after 43. e5 Qxa3 44. Rxf7 what exactly happens after 45...Qc1? black will exchange queens and i don't know how white is going to stop that a-pawn. i would conclude that this is *really* not a good test position, and so i'm not that unhappy that muse has serious problems finding the "best" moves. i admit, e5 and Rxf7 are a good "desperado" attempt to save the game, but i don't really believe it's good enough. >Latest Movei does not find 43.e5 in a few minutes inspite of the fact that Bxb4 >fails low again and again. >Hopefully it can find it at depth 15 after searching to depth 15 but more than >10 minutes in a fast hardware(A3000) is not something that I can define as a >reasonable time same here: muse takes very long to decide to play e5 on an Athlon64 3000+ Muse UCI 58 MB: 10 00:06 -0.99 Ba3xb4 Nb3d4 Qd1e1 Qa1b2 Bb4xd6 Nd4xf3+ g2xf3 Qb2xb1 Ne3d1 Rb8b3 Qe1e2 a4a3 11 00:19 -0.91 Ba3xb4 Nb3d4 Qd1e1 Qa1b2 Ne3c2 Nd4xc2 Bb1xc2 Rb8xb4 Rf3c3 a4a3 Bc2d3 12- 01:15 -1.41 Ba3xb4 12 01:39 -1.55 Ba3xb4 Nb3d4 Qd1e1 Nd4xf3+ g2xf3 Qa1b2 Ne3c2 Rb8xb4 Nc2xb4 Ra8b8 Bb1a2 Qb2xb4 13- 09:46 -2.05 Ba3xb4 13 15:01 -2.08 Ba3xb4 Nb3d4 Qd1e1 a4a3 Ne3c2 Nd4xc2 Bb1xc2 Qa1xe1+ Bb4xe1 a3a2 Be1c3 Rb8b1+ 14+ 34:52 -1.84 e4e5 14 38:29 -1.84 e4e5 d6xe5 and after e5 Qxa3 it takes very long to see Rxf7: Muse UCI 58 MB: 10 00:37 -1.84 e5e6 f7xe6 d5xe6 Qa3c1 Kg1h1 Qc1xd1+ Ne3xd1 Kg8h7 Rf3f6 Rb8g8 e6e7 Nb3d2 11 01:33 -1.77 e5e6 f7xe6 d5xe6 Qa3c1 Kg1h1 Qc1xd1+ Ne3xd1 Kg8h7 Rf3f6 Rb8g8 e6e7 Nb3d2 Bb1d3 12+ 08:21 -1.27 Rf3xf7 12 09:25 -1.15 Rf3xf7 Kg8xf7 Qd1f3+ Kf7g8 Qf3e4 Qa3c1+ Kg1h2 Kg8h8 Qe4xg6 Ra8a7 13 18:18 -1.52 Rf3xf7 Kg8xf7 Qd1f3+ Kf7g8 Qf3e4 Qa3a1 Qe4f3 Qa1xb1+ cheers martin
This page took 0.01 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.