Author: Uri Blass
Date: 04:56:31 09/10/01
Go up one level in this thread
On September 10, 2001 at 06:37:57, Gian-Carlo Pascutto wrote: >On September 09, 2001 at 21:05:23, Dave Gomboc wrote: > >>Is this a won position for Black? DJ6 doesn't seem to think so, but it might >>just be wrong... > >Cray Blitz - Deep Thought II >[D]4k3/7R/3p2p1/4bpB1/8/3R2PK/r6P/r7 b - - 0 1 > >Analysis by Crafty 18.08: > >49...Rb1 50.Rd2 Rxd2 51.Bxd2 Rb3 52.Bg5 f4 53.Re7+ Kf8 54.Re6 f3 55.Rxg6 Kf7 >56.Rh6 f2 57.Kg2 Rb2 > -+ (-2.06) depth: 14/31 00:02:11 84431kN, tb=113 >49...Rb1 50.Rd2 Rxd2 51.Bxd2 Rb3 52.Bg5 f4 53.Re7+ Kf8 54.Re6 f3 55.Rxg6 Rb1 >56.Be3 Kf7 57.Rh6 Rb2 > -+ (-1.89) depth: 15/35 00:02:44 105884kN, tb=245 >49...Rb1 50.Rd2 Rxd2 51.Bxd2 Rb3 52.Bg5 f4 53.Re7+ Kf8 54.Re6 f3 55.Rxg6 Kf7 >56.Rh6 f2 57.Kg2 Rb2 58.Be3 Kg7 > -+ (-2.00) depth: 16/37 00:05:46 217826kN, tb=1256 >49...Rb1 50.Rd2 Rxd2 51.Bxd2 Rb3 52.Bg5 f4 53.Re7+ Kf8 54.Re6 f3 55.Rxg6 Kf7 >56.Rh6 Rb2 57.Rh4 d5 58.Rh7+ Ke6 59.Re7+ Kf5 > -+ (-2.14) depth: 17/39 00:11:33 428732kN, tb=4302 >49...Rb1 50.Rd2 Rxd2 51.Bxd2 Rb3 52.Rh4 Kf7 53.Bg5 Ke6 54.Rc4 Kd5 55.Rh4 Kc5 >56.Ra4 Rb2 57.Be3+ Kb5 58.Ra8 Kc4 59.Rd8 > -+ (-2.29) depth: 18/41 00:31:10 1133497kN, tb=15025 > >I tend to prefer Crafty over Hiarcs in endgames, but I have >no idea whose assesement is better here. But note that Crafty >is already heavy in the tablebases here, and the score is >going up for black... Note that it seems to me as a human that white has good chances for a draw. White only ned to trade the pawns and to sacrifice the bishop for the passed pawn. > >Hiarcs is around -1.55 in all these positions, but doesn't >hit tablebases yet. > >>Back to 47.Kg1 Ra1+ 48.Kf2 R6a2+ 49.Rd2 Bf6 50.Bh6 g5 > >Instead of 50. .. g5, 50. .. d5! and I get -2.26 in >favor of black. > >>I suspect that 46.Re7+ Kf8 47.Rb3 Ra8 48.Rxh7 Rxa2 49.g3 is about the same. > >Hiarcs d5 with -1.9, Crafty -2.2 > >>42.Bh6 Bg7 43.Bxg7 Kxg7 might just be a draw, though. Hmm, 42...Rcxa4, I guess, then it's similar to before. > >Yes > >>I don't know, GCP, I don't think these pawn up endings deserve +2.25. But in >>order for DT2 to settle on that kind of score, if it got to these variations it >>must have assessed them as such. > >Crafty obviously disagrees with you, as does Hiarcs in some of the positions >(although its score is usually more around -1.6, it drops down the lines). > >The question is of course whose assesement is more correct. > >I do think this shows that DT's assesement wasn't bogus or due to >a bug. This shows nothing because there are millions of other lines that you should consider. It is known that most of the lines that program consider are illogical lines and it is clear for everyone who use common sense that deeper thought could not get so deep (if you assume that there is a geometric average of 2 lines that it needed to consider in every ply then seeing 40 plies forward from the root position in the relevant lines is impossible at tournament time control game and even 30 plies is impossible and it is clear that there are a lot of position in the lines that you posted when the side has more than 2 alternatives). This is probably going to be my last post in this subject inspite of the fact that I am not going to change my mind about it. I may argue about the question if the position is a draw or a win for black but not about the question if deep thought could see +2 pawns because it is clear that you do not assume that there was something impossible for Deep thought to do in this discussion when I assume that there are things that are impossible. Deeper blue could not see 60 plies forward against kasparov in a line when almost all the moves were singular or almost singular and you try to convince me that Deep thought that was weaker could see more than it in a clearly more quiet position. I do not buy it. 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.