The Poisoned Pawn variation in the Najdorf Sicilian is more than half a century old. It attracted players of various styles -- from the Don Quixotes, who charge forward no matter what the cost, to the buccaneers, who grab any pawns and pieces coming their way.
As more and more pawns were shed recently, the controversial opening line became so complicated that, inevitably, the computers got involved. A game from the sixth PAL/CSS Freestyle tournament on the Playchess server in July, where computers were allowed to assist the human players, is a good example why this joint effort makes sense.
1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 a6 6.Bg5 e6 7.f4 Qb6 8.Qd2 Qxb2 (According to several theoretical books and databases, this poisoned pawn was first taken in the Nezhmedtinov-Shcherbakov game, Riga 1954.) 9.Rb1 (Nezhmedtinov played 9.Nb3, a move made famous in two world championship matches. After 9...Qa3 10.Bxf6 gxf6 11.Be2 h5 12.0-0 Nc6 13.Kh1! Bd7 Boris Spassky shocked Bobby Fischer in Reykjavik in 1972 with 14.Nb1!? and won rather quickly. Fischer never repeated that line in the match. For the match against Garry Kasparov in London in 1993, I armed Nigel Short with 14.Nd1!? setting a trap on the previously played 14...Qb4? with 15.c3! Qxe4? 16.Bd3 and black loses either after 16...Qa4 17.Nb2 Qa3 18.Nc4 Qa4 19.Nb6; or after 16...Qd5 17.Ne3 trapping the black queen. However, Kasparov had avoided it with 14...Rc8 and won after Short misplayed the position.) 9...Qa3 10.e5 (We covered 10.f5 in last week's column.) 10...dxe5 11.fxe5 Nfd7 12.Ne4 Qxa2? (Lining up white's light pieces on the 4th rank first with 12...h6! 13.Bh4 Qa1+ 14.Kf2 Qa4! prevents dangerous sacrifices.)
13.Rd1!? (A recent trend. My first experience with the poisoned pawn line half a century ago was pleasant, but it should not have been. In the game Filip-Kavalek, Prague 1957, after 13.Rb3 Qa1+? 14.Kf2 Qa4 my opponent improved on the game Tal-Tolush, Leningrad 1956, where after 15.Bb5!? axb5 16.Nxb5 f6 17.exf6 black missed 17...Nxf6. Tolush later played it against Listengarten in Erevan in the same year and won quickly. Kasparov likes 17...Qxe4 even better. Instead of Tal's 15.Bb5!?, Filip sacrificed the correct piece 15.Nxe6!!, but after 15...fxe6 16.Nd6+ Bxd6 17.Qxd6 Rf8+ gave the game away with 18.Rf3? Rxf3+ 19.gxf3 Qxc2+ 20.Kg3 Qc5 21.Qxe6+ Kf8 22.Bh3 Nc6 23.Qd6+ Qxd6 24.exd6 Nde5 25.f4 Bxh3 26.fxe5 Be6 and resigned. He should have played 18.Kg3! with a decisive attack. ) 13...h6 14.Bh4 Qd5 15.Qe3 Qxe5 (This makes it a triple poisoned pawn variation.) 16.Be2 Bc5 17.Bg3 Bxd4 18.Rxd4 Qa5+ 19.Rd2 0-0 20.Bd6 Nc6!? (Van Wely's promising exchange sacrifice.) 21.0-0 ! (Shirov's idea, rejecting the exchange and keeping one more attacking piece on the board.) 21...Re8 22.Qg3 Qb6+ 23.Kh1 Ne7 24.Bc7 Qc6 (Surprisingly, the machine judged the position as equal, disregarding the following rook sacrifice.) 25.Rxf7! Kxf7 26.Nd6+ Kg8 Black resigned. Why? According to Arno Nickel, the computer suddenly saw the forced win: 27.Nxe8 Nf5 28.Qg6 Qc3 29.Qxe6+ Kh8 30.Qxf5 Qxd2 31.Qf7! Qe1+ 32.Bf1 Qa1 33.Bd6! and black has no defense, e.g. 33...a5 34.c3! Qxc3 35.Bf8 Nxf8 36.Qxf8+ Kh7 37.Bd3+ g6 38.Nf6+ and white wins. This long variation is a humbling and terrifying experience for the humans. The machines are here to stay.
Solution to today's study by M. Schapiro (White: Kc4,Rd5; Black: Ka2,P:g3,h3): 1.Rd2+ Kb1 2.Kc3 Kc1 (2...g2 3.Rd1+ Ka2 4.Rg1! wins; 2...h2 3.Rd1+ Ka2 4.Rh1! wins) 3.Ra2 Kd1 4.Kd3 Kc1 (4...Ke1 5.Ke3! wins) 5.Ke3 h2 (5...Kb1 6.Re2! wins) 6.Ra1+ Kb2 7.Rh1! wins.