Game 8: Leko wins to take the lead

by ChessBase
10/7/2004 – In a stunning 32-move Marshall Gambit, Peter Leko demolished Vladimir Kramnik's preparation and took the lead in their world championship match 4.5-3.5. Leko accepted the challenge to play the Marshall and then found a flaw in Kramnik's prepared queen sacrifice. Updated with photos and video report.

ChessBase 17 - Mega package - Edition 2024 ChessBase 17 - Mega package - Edition 2024

It is the program of choice for anyone who loves the game and wants to know more about it. Start your personal success story with ChessBase and enjoy the game even more.


Game 8: Brilliant Leko refutes Kramnik homework

Many questions were answered by today's eighth match game. Does Leko have what it takes to beat Kramnik in a match? Is Kramnik's preparation simply unbeatable? Why don't many top players allow the Marshall Gambit with white? (That's "yes", "no", and "because no matter how well prepared you are you may get killed".)

Vladimir Kramnik and Peter Leko at the start of game eight

In the most exciting game of the match so far, Hungary's Peter Leko played a beautiful sacrificial attack to take his first lead of the match. Kramnik allowed Leko's favorite Marshall Gambit for the first time and raced through a rare line in his preparation. Leko was using a lot of time, but he was up to the challenge when both players noticed a serious hole in the world champion's homework. Leko unleashed a brutal attack that scored the full point in 32 moves.

... and in the press conference, which is transmitted live on

Kramnik: "Certainly I'm disappointed, but at the same time I have to say: Peter Leko played a beautiful game."

Leko: "I had quite a mixed feeling during the game. But now I’m very happy that I won it."

The match annotators GMs Artur Jussupow and Helmut Pfleger doing postgame analysis with the public in Bressago. This too is carried live on

Vladimir Kramnik
Peter Leko

Kramnik,V (2770) - Leko,P (2741) [C89]
WCh Brissago SUI (8), 07.10.2004

C89: Closed Ruy Lopez: Marshall Attack 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.0-0 Be7 6.Re1 b5 7.Bb3 0-0 8.c3 d5 9.exd5 Nxd5 10.Nxe5 Nxe5 11.Rxe5 c6 12.d4 Bd6 13.Re1 Qh4 14.g3 Qh3 15.Re4 g5 16.Qf1

[16.Qe2 f5 (16...Nf6 17.Nd2 Bf5 18.f3 c5 19.Qf2 c4 20.Bc2 h6 21.b3 cxb3 22.axb3 Rfc8 1-0 Ponomariov,R-Adams,M/Linares ESP 2002/(45)) 17.Bxd5+ cxd5 18.Re6 f4 19.Rxd6 Bg4 20.Qf1 Qxf1+ 21.Kxf1 Rae8 22.Bd2 Bh3+ 23.Kg1 fxg3 24.hxg3 Re2 25.Be3 Rxe3 26.fxe3 Rf1+ 27.Kh2 g4 28.Rxd5 1/2-1/2 Ponomariov,R-Anand,V/Linares ESP 2002/(28)]

16...Qh5 [16...Qxf1+ 17.Kxf1 Bf5 18.f3 h6 19.Re1 Rfe8 20.Bxd5 cxd5 21.Rxe8+ Rxe8 22.Kf2 a5 23.a3 Bd3 1/2-1/2 Peng Xiaomin-Grischuk,A/Shanghai 2001/CBM 85 (23)]


[17.f3 Bf5 18.Bxd5 cxd5 19.Re1 Rae8 20.Be3 Bh3 21.Qf2 f5 22.Na3 f4 23.gxf4 gxf4 24.Bd2 Kh8 25.Kh1 Rxe1+ 26.Rxe1 Rg8 27.Rg1 Rxg1+ 28.Kxg1 Qg6+ 29.Kh1 Qd3 0-1 Rodriguez,R-Krylov,P/EU-ch M corr 1960/Corr 2000 (29)]

17...Bf5 18.f3

[18.Bxd5 cxd5 19.Re3 Rae8 20.a4 Bh3 21.Qe1 f5 22.f3 Bf4 23.Rxe8 Rxe8 24.Qd1 Be3+ 25.Kh1 Bf2 26.g4 fxg4 27.fxg4 Bxg4 0-1 Megaranto,S-Garcia,E/Oropesa del Mar 2000/EXT 2001; 18.Bd1 Qg6 19.Re1 Rae8 20.Be2 Qe6 21.Nf3 Qd7 22.Bd1 Bh3 23.Qd3 Nf4 24.Bxf4 gxf4 25.Bc2 f6 26.Nh4 fxg3 27.hxg3 f5 28.Rxe8 Qxe8 29.a4 Qd7 30.Kh2 Bg4 31.f3 Bh5 32.axb5 axb5 Asztalos,V-Antal,A/Eger 2002/EXT 2004/1/2-1/2 (37)]

18...Nf6 The first new move of the game, but clearly prepared for by Kramnik, who continued to play quickly.

[18...Rae8 19.Rxe8 Rxe8 20.Ne4 Bxe4 21.fxe4 Rxe4 22.Bd1 g4 23.Qf2 Re6 24.a4 b4 25.Bd2 bxc3 26.bxc3 c5 27.Rb1 cxd4 28.cxd4 Bc7 29.Rb7 Rf6 30.Qe2 Bb6 31.Rd7 Bxd4+ 32.Kh1 Be3 33.Be1 Elyakim,D-Krempel,T/W-ch M500 corr 1990/Corr Nr. 1/0-1 (33)]

19.Re1 Rae8 20.Rxe8 Rxe8 21.a4 Qg6 22.axb5 Bd3 23.Qf2 [23.Qd1!? axb5 24.Ra7]

23...Re2 24.Qxe2? [better (but still losing) is 24.bxa6 Rxf2 25.Kxf2 Qh6 26.Ke3]

24...Bxe2 25.bxa6 [25.Bc4 Bxc4 26.Nxc4-+ cxb5! 27.Nxd6]

25...Qd3-+ Kramnik's preparation went this far, believing his passed pawn gave him compensation. But White is lost here thanks to some nice sacrificial lines that Leko finds over the board.

26.Kf2 (diagram) [26.a7 Qe3+ 27.Kg2 Bxf3+! 28.Nxf3 Qe2+ 29.Kg1 Ng4! 30.a8Q+ Kg7-+]

26...Bxf3!! 27.Nxf3 Ne4+ 28.Ke1 Nxc3!

[28...Qxf3? This leads only to a draw. 29.a7 Qf2+ 30.Kd1 Qf1+ 31.Kc2 Qe2+ 32.Kb1 Qd3+ 33.Bc2 Nxc3+ 34.bxc3 Qb5+ 35.Bb2 Qf1+ 36.Ka2 Qa6+ 37.Kb1 Qf1+ 38.Bc1 Qb5+ 39.Bb3 Qxb3+ 40.Bb2 Qd1+ 41.Ka2 Qa4+ 42.Kb1 Qd1+ 43.Ka2 Qa4+ 44.Kb1 Qd1+=]

29.bxc3 Qxc3+ 30.Kf2 Qxa1 31.a7

[31.Bxg5 Qxa6 32.Bd1 c5 33.dxc5 Bxc5+ 34.Ke1 Kg7-+]

31...h6 32.h4 [32.Bc4 Qxa7 33.Be3 c5-+] 32...g4 0-1 [32...g4 33.Ne5 Bxe5 34.dxe5 Qxa7+ 35.Be3 Qa1-+]

After we posted this game analysis, this email exchange about the game took place between commentator Mig Greengard and Fritz co-programmer Mathias Feist.

Mig: [Regarding 24.bxa6] I have similar lines, I just don't like to give too many eval marks on quick analysis. That was just Fritz's html export way of saying "better is", but it's confusing. I clarified it.

The interesting question to me is how Kramnik's preparation missed this when Fritz finds it in 20 seconds. Very strange and maybe even suspicious. The long line with 26.a7, queening with check but losing, confuses Fritz for a little while, but after a few minutes it's all clear. If this weren't a WC match I would suspect the usual GM habit of looking for five seconds with three or four lines on a two-year-old laptop. But I know Vlady must have better hardware for his team. Right?

Mathias: I also thought about it. If you have the game it does look strange indeed. But this is just one line of dozens they must have checked while preparing the Marshall. They may have stopped analysis before bxa6 or even before Qxe2 because the next moves are forced, which would practically hide the discovery of any engine analysis unless you wait for a very long time. The refutation was hidden in the sheer amount of preparation.

Leko found it over the board, so it seems. Kramnik relied on his preparation and didn't really work during the game. I think he may have seen the problem over the board, too, if he also would have worked.

Of course there's another possibility: they analysed and saw it, but Kramnik forgot he had to play Qd1 instead of the line actually played. The nice idea with the queen sacrifice was glued to the brain together with a tag that the line is okay, the dreary ordinary stuff was overridden.

Contacts and further information

Rolf Behovits
Press Officer World Chess Championship
Via Ruggero Leoncavallo
CH-6614 Brissago


Reports about chess: tournaments, championships, portraits, interviews, World Championships, product launches and more.


Rules for reader comments


Not registered yet? Register