(1) Anderssen,A - Dufresne,J [C52]
Game 4, p.26: Berlin 'Evergreen', 1852
[Kasparov, Garry]

1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.b4 Bxb4 5.c3 Ba5 6.d4 exd4 7.0-0 d3 8.Qb3 Qf6 9.e5 Qg6 10.Re1 Nge7 11.Ba3 b5 12.Qxb5 Rb8 13.Qa4 Bb6 14.Nbd2 Bb7 15.Ne4 Qf5 16.Bxd3 Qh5 17.Nf6+ gxf6 18.exf6 Rg8! 19.Rad1!
[ Lasker's recommendation 19.Be4!? In Volume 1 I gave the variation 19...Qh3 20.g3 Rxg3+ 21.hxg3 Qxg3+ 22.Kh1 Bxf2 23.Bxe7! ( 23.Re2? Nd4!! ) 23...Qh3+! 24.Nh2 Bxe1 25.Rxe1 Qh4! 26.Qd1! Nxe7 27.Bxb7 Qxf6 28.Qg4 "with the initiative". It turns out that this position had been studied some time ago by Murey and Fridstein (in 64, 1975 No.38) and after them also by Zaitsev, who thought that after 28...Kd8! White has nothing real. ( Better is, according to Zaitsev, 28.Bd5!? Qxc3 ( 28...Rb6? 29.Qh5! ; less clear is 28...Rb2 29.Qh5 c6 30.Ng4 Qf4 ) 29.Qe2 ( 29.Bxf7+? Kxf7 30.Qh5+ Kg8 with a draw.) 29...Qb4 ( 29...Qc5?! 30.Qe5! ) 30.Ng4 ( 30.a3?! Qd6! ) 30...Kd8 ( Apparently 30...Kf8! is more accurate. In short, a vicious circle arises, where it is hard to find a draw, but, in view of the limited material, even harder to demonstrate a win!) 31.Qe5! c6 ( 31...Qxg4? 32.Qxe7+ ) 32.a3! Qxg4 ( 32...f6? 33.Nxf6 Qh4+ 34.Kg1 and wins) 33.Qxb8+ Nc8 34.Qe5! Qh4+ 35.Kg1 cxd5 36.Qe8+ Kc7 37.Rc1+ Kd6 38.Rxc8 ) 28...Kd8! ]

[ 19...Rxg2+? 20.Kxg2 Ne5 21.Qxd7+! would also have lost; but 19...Rg4 (Lipke) was much stronger. The main line is 20.Bc4 ( More consideration should be given to 20.Re4!? Rxe4 21.Qxe4 d6 22.Re1! Qg6! and Black's defences hold.; Now after 20.c4? Bd4!? with the threat of ...Rxg2+ (Zaitsev) is not bad ( instead of 20...Rf4? 21.Bg6! (Hoppe, Heckner) winning ( or 21.Qb5 (Neishtadt) winning) ; but in my opinion, the immediate 20...Rxg2+! is even better: 21.Kxg2 Qg4+ 22.Kf1 Qxf3 23.c5 ( Zaitsev's move 23.Rxe7+ Nxe7 24.Qxd7+ Kxd7 25.Bf5+ Ke8 26.Bd7+ Kf8 and wins) 23...Qh3+ 24.Kg1 ( 24.Ke2 Ba5 25.Bb5 Nd4+! ) 24...Ne5 with a decisive attack.) ) 20...Qf5! 21.Rxd7 and here, fearing the variation ( Then he tried to improve White's play with 21.Rd2 but he himself parried it with the thematic move 21...Bd4! ) 21...Kxd7 ( Zaitsev suggested 21...Rxg2+!? 22.Kxg2 Qg4+ 23.Kf1 Qh3+ 24.Ke2 Qxd7 25.fxe7 or ( 25.Bxe7 Nd4+! ) 25...Nd4+ 26.Nxd4 Qxa4 27.Bb5+ Qxb5+ 28.Nxb5 Ba6 29.c4 c6 30.Rg1 Kd7 31.Rd1+ Ke8 32.Nd6+ Kxe7 with a draw.) 22.Ne5+ Kc8 23.Nxg4 23...Nd5 24.Qd1 Nxf6 ( 24...Nd8? 25.Bd3! ) 25.Bd3 Qxg4 26.Qxg4+ Nxg4 27.Bf5+ Kd8 28.Rd1+ Nd4 29.Bxg4 Bd5 30.cxd4 Bxa2 with equality); In addition Zaitsev found that after 19 Rad1 good was 19...Bd4! 20.cxd4 ( 20.Nxd4 Rxg2+ 21.Kxg2 Nxd4+ ) 20...Qxf3 21.Be4 Rxg2+ 22.Kh1 Rxh2+ 23.Kxh2 Qxf2+ 24.Kh3 Qxf6 with a draw. It only remained to check once again Lasker's recommendation 19 Be4!?.; Tim Shackel: Hi. This is in regard to the position Garry gave the follow-up for. Is it just me or can black just play in response to Anderssen's 19. Rad1?! 19...Qh3!? threatening mate, and 20.Bf1 is the only move, then 20...Qf5 threatening to pick off the f6 pawn and much of the e7 threat dangers, and then 21.Bd3 ( 21.Kh1 Bxf2 22.Re2 Bb6 23.h3 Rg3-/+ ) 21...Qxf6 22.Bxh7 Kf8 23.Bxg8 Kxg8 24.Rxd7 Ng6 looks very playable for black! I would love that nice attacking position.]

20.Rxe7+! Nxe7!?
Another proof that chess masterpieces require the generous cooperation of the loser! Nowadays a professional player and, of course, a computer, would have without hesitation resorted to [ 20...Kd8 21.Rxd7+! Kc8! 22.Rd8+! Kxd8 23.Be2+! etc. is more prosaic]

21.Qxd7+!! Kxd7 22.Bf5+ Ke8
[ 22...Kc6 23.Bd7# ]

23.Bd7+ Kf8 24.Bxe7# 1-0