|
|
r2qk2r/ppp1b1pp/2n1p3/3pP1n1/3P2b1/2PB1NN1/PP4PP/R1BQK2R w KQkq - 0 1 Nolot's analysis as reported by Baudot: Smaguine - Sahovic, Bienne 1990 White wins with a queen sac but black has several ways to defend 12.Nxg5!! Bxd1 13.Nxe6 Qb8 14.Nxg7!! Kf8 15.Bh6! Bg4 16.0-0+ 17.Kg8 17.Rf4 +- It should take between a few months and a few years for a program to find 12.Nxg5!! Comments by Feng-Hsiung Hsu: Solution move is 1. Ng5!. Verified that the move is sound by following the published analysis, but could not play it within one hour time. This is more a positional sac than what would be normally called tactics. Later comments by Feng-Hsiung Hsu: We took a closer look at this position. 12. Ng5 is a sound positional sac, but depending on the temperament of the player, it might not be the best move. The published annotation gives 12. Bg5 Bg5 13. o-o as +=, but white could play 13. h3 instead and white appears to have a simple positional squeeze. The critical line in the 12. Ng5 variation turns out to be 12. Ng5 Bd1 13. Ne6 Qb8 14. Ng7 Kd8 15. Kd1. Black's queen and rooks are temporarily out of play, black is up a pawn, but white has a protected passed pawn, and lots of pressure. (The annotator gave one line that ended "with the attack":). From DT-2's point of view, The Bg5 line was gaining 0.20 pawn after each iteration, and so was the Ng5 line. Except that the Bg5 line has about a 0.20 pawn lead at the same depth. There appears to be no kill in the Ng5 line when black king goes to d8 instead of the f file. Black would have to give up the extra pawn to activate the queen and the rooks, and while white is definitely better, black is not without counter play. On the deepest search that we checked out, black's evaluation stopped dropping at around -1.4 pawns, and black's pieces were becoming active. My own experience: In approximately 1999, I ran this for five or six days on a quad 450 mhz Xeon machine, and didn't find this one. 3-Sep-01: I ran this for a couple of days on a quad 450 mhz Xeon. Bxg5 (not the key)was selected with a score of about +0.65, which isn't big news -- probably everyone does this. Simultaneously, I ran this on a single-processor 550 mhz P3, with the moves 12. Nxg5 Bxd1 forced. In this second case, the score was -0.06 when I stopped it. That's pretty good but there's still a pretty big spread to overcome. I think it would be possible to solve this positionally if a program had the right terms. Mine doesn't have the right terms, because it values a queen when attacking, and won't be very happy about the idea of attacking with multiple light pieces. A tactical solution is going to have to wait, at least in my case, because finding a concrete win after the first two plies are forced seems to be difficult. |
Send mail to
brucemo@seanet.com with
questions or comments about this web site.
|