The 2nd game of the match has become instant chess history. It had everything - brilliant sacrifices, tenacious defence, study-like endgames - and mistakes, plenty of them. Yes, this game is far from ideal - but few will match it for the raw emotions on display.

Slav Defence D19
V.TOPALOV - V.KRAMNIK
Game 2

1.d4. In this analysis I decided to concentrate solely on the period between moves 19 and 42 - you will not find any opening survey here, and I chose to stop the analysis at the point where the pendulum swung, decisively, in Kramnik's favor. What interested me was the phase of play when tension was at its highest - and games don't come much more tense than this one.

1...d5 2.c4 c6 3.Nf3 Nf6 4.Nc3 dxc4 5.a4 Bf5 6.e3 e6 7.Bxc4 Bb4 8.0-0 Nbd7 9.Qe2 Bg6 10.e4 0-0 11.Bd3 Bh5 12.e5 Nd5 13.Nxd5 cxd5 14.Qe3 Bg6 15.Ng5 Re8 16.f4 Bxd3 17.Qxd3 f5 18.Be3 Nf8 19.Kh1.

The first critical point of the game. After meeting 15.Ng5, a relatively rare move, with a novelty of his own, Vladimir got a reasonably comfortable position - his bishop is clearly superior compared to its White counterpart, and the only thing that is standing between Black and mass sterilization along the 'c' file is a potential onslaught on the kingside. One gets the feeling, however, that organizing an attack will prove quite hard.

19...Rc8?! And this is the first indication of just how cosy Black believes his position to be. Playing h7-h6 is wrong structurally - after the eventual g2-g4 Black will certainly regret it - but it was much more prudent to drive the knight away by playing 19...Be7, and after 20.Nf3 Rc8 21.h3 Qc7!? 22.Rfc1 (22.Rf2 Qc4 23.Qd1 Ng6, and White is going nowhere) 22...Qd7 Black has very little to fear.

20.g4! Spurred on by Black's momentary lapse, Veselin immediately goes on the offensive.

20...Qd7. After a long think - the problems are already considerable. After the natural 20...h6 White, of course, will not retreat: 21.Nxe6! Rxe6 22.gxf5, and the attack develops by itself. Black's resources are by no means exhausted, but such a sudden change of the landscape was understandably not to Kramnik's liking.

21.Rg1 Be7 22.Nf3. 22.Nxe6 looks very tempting, but it is no longer needed. A very important detail has changed compared to the position in the notes to 19...Rc8 - Black Queen is now tied to the f5 pawn, and his counterplay on the Queenside is therefore much more limited. He has only one target there - the a4 pawn - and White can certainly dispense with it for the sake of the attack.

22...Rc4 23.Rg2. White is not stopping for anything.

23...fxg4. Vladimir could not have been very happy with this decision - it's obvious that after the opening of the g-file he will need a lot of things to go his way to withstand the direct assault. Since it was possible to postpone the opening of the hostilities by, for instance, 23...Ng6, this indicates the confidence he had in his defensive resources.

24.Rxg4 Rxa4 25.Rag1. For a while now play becomes completely forced.

25...g6. 25...Ng6 26.h4 Bf8 27.Ng5! would lose on the spot.

26.h4! The whole game is now revolving around the g6 square - if Black will be able to hold on to that outpost he has a good chance of survival, while White only needs to break through there to be almost certain of winning, and he throws everything he has at it. Note this single focal point - we will have to revisit it later.

26...Rb4. Black urgently needs to start distracting White's attackers.

27.h5 Qb5 28.Qc2! Not letting g6 out of his sight.

28...Rxb2. It is very difficult not to make a move like that. In fact, Black did have a viable alternative in 28...Rb3 - but after 29.hxg6! h5 30.R4g2! (White will only have a perpetual after 30.g7 hxg4 31.gxf8Q+ Kxf8 32.Qg6 Qd3! 33.Qh6+ Kf7 34.Rxg4 Rg8 35.Qh5+ Kf8 36.Qh6+ Kf7) 30...Qd3 (30...Rxe3 31.g7 Nh7 32.Qg6 loses on the spot) 31.Qxd3 Rxd3 32.g7 Nd7 (32...Nh7 33.Re2 should favor White - the knight on h7 does not inspire confidence) 33.Rh2! Kf7 (33...Rxe3 34.Rxh5 Kf7 35.Nh2!!, and there is no defence against Ng4-h6) 34.Re2 (34.Rxh5 Rg8) 34...Kg8 (otherwise f4-f5 will hurt) 35.Rg6, and White is on top.

29.hxg6! There is no choice, of course - but it is still a pleasure to make a move like that.

29...h5. The only move: 29...Rxc2 30.gxh7+ Kxh7 31.Rg7+ Kh6 32.f5+ Bg5 33.R7xg5 is an obvious mate, while 29...Nxg6 30.Qxg6+!! (30.Rxg6+ would be a fatal mistake: 30...Kh8!-/+) is easier to miss, but just as deadly: 30...hxg6 31.Rxg6+ Kh7 32.R6g3! Bh4 33.Rh3+-.

30.g7 hxg4 31.gxf8Q+. Up to this moment the only real choice either of the players had was 28...Rb3, but now we enter the slightly surreal passage of play.

31...Bxf8?? Vladimir took a while to make this move. In this time, all those watching and commenting (most of them with engines at full throttle) realized that this move is impossible, and were busy figuring out exactly what is going on after the forced 31...Kxf8! In reply, White will play 32.Qg6 (or 32.Qh7 Qe2 33.Qh6+ Kf7 34.Qh5+ Kf8) 32...Qe2 (importantly, compared with 28...Rb3, Black does not control d3, and has to use the inferior e2 square) 33.Qxg4, and after the spectacular, even if completely forced, 33...Bg5! 34.Re1! Qc2 the position is still quite unclear: 35.fxg5 (35.Qxg5 Re7! 36.Rc1 Rh7+ 37.Kg1 Rg7 38.Rxc2 Rxc2, and the passers on the Queenside are looking ominous) 35...Kg7!? (the king can also run towards the Queenside, but keeping the g5 passer in check makes more sense) 36.Rc1 (36.Kg1 Rf8, planning 37...Qf5) 36...Rh8+ 37.Kg1 Rb1 38.Rxb1 Qxb1+ 39.Kf2 Rf8!, and if Black can somehow get to an ending, his chances will be quite decent.

32.Qg6+?? And now we have to return to my point about the battle of g6. Yes, White had an immediate win here. After 32.Rxg4+ Bg7 33.Qc7!! Qf1+ 34.Ng1

Black can resign with a perfectly clear conscience. How could two of the best players in the world miss such an obvious thing? Of course, I don't know for certain, but for whatever it's worth, here is my theory. In two important ways, this line is very counter-intuitive. Not only is White relinquishing control on the f1 square by playing 32.Rxg4, on the next move he has to make a move along the c-file and away from the b1-h7 diagonal, along which all the action up to this point has been concentrated. Of course, once you spot 33.Qc7, you immediately realize that Black only has one check, and the bishop on g7 is defenceless - but even looking in that direction seems completely unnatural. Another important thing to note is that the position in front of you does not come with a caption at the bottom: 'White to play and win in two moves' - I am quite certain Veselin would have found the solution if he was sure it was there. This is where the computer has an overwhelming advantage over humans - the silicon is of course completely impervious to the kind of psychobabble I have just offered as explanation. Upon reflection, I was able to remember a similar story I (and millions of chess lovers who've read the book on Zurich 1953) read about as a child.

L.SZABO - S.RESHEVSKY
Zurich 1953

Black has just played Bg7xf6. Of course his position is bad even after Kh8, but now White has a obvious mate in two - which he did not spot, presumably because he was not expecting his opponent to blunder in such a monstrous fashion. Mutual chess blindness is rare - but not unheard of.

32...Bg7. Now the game starts afresh.

33.f5! 33.Ng5 Re7 is only good for a draw, and Veselin rightly continues to press for more.

33...Re7 34.f6 Qe2 35.Qxg4 Rf7. Once again a series of only moves - and despite his miss White is still one move away from almost certain victory. However, this time the solution is much less obvious.

36.Rc1. A very human decision - this move forces Black to trade his active rook, and White will feel much safer for it. But in fact the game becomes very double-edged - with every exchange Black passers on the Queenside are becoming more and more restless. 36.Bh6 looks promising until you notice 36...Rb3! 37.Nh2 Qxg4 38.Rxg4 a5! 39.Bxg7 a4, and Black is bossing the game. But 36.Qh5!! would have won. The problem with this move is that it is not enough to see it - you also have to figure out what the threat is, and that is much more difficult. At the moment White is not ready to force the matters just yet, but Black is unable to stop the killer rook lift: 36...a5 (36...Qxe3 loses to 37.Ng5; 36...Rb3 allows 37.Rxg7+! Rxg7 38.fxg7 Rb1+ 39.Bg1, and Black is lost: 39...Kxg7 40.Qg5+ Kh7 41.Qe7+ Kh8 42.Qf6+ Kg8 43.Qxe6+ Kg7 44.Qf6+ Kg8 45.e6+-) 37.Rg3!!,

and now White is threatening 38.fxg7 Rxg7 39.Qe8+ Kh7 40.Rh3#, forcing Black to play 37...Qxe3 38.fxg7 Rb1+ 39.Kh2 Rb2+ 40.Kh3 Rxg7 41.Ng5! Qf4 42.Qe8+ Qf8 43.Qxf8+ Kxf8 44.Nxe6+ Kf7 45.Nxg7 a4 46.Nf5, with a relatively simple win for White. There is another option as well: 37.Bh6 Rb3 38. Qxf7 Kxf7 39.Rxg7 Ke8 40.f7 Kf8 41.Rg2 Kxf7 42.Rxe2 Rxf3, which is objectively weaker, but easier to find.

36...Rc2 37.Rxc2 Qd1+. Winning a tempo but also improving the position of the White King. 37...Qxc2 was a decent alternative.

38.Kg2 Qxc2+ 39.Kg3.

Despite the positive dynamic, Black position is still very dangerous. However, both players were running short of time.

39...Qe4!? Black needs to trade the Queens, but both ways of doing it have obvious drawbacks. After spending almost all of his 10 minutes, Kramnik chooses the more provocative of the two options. During the game I felt that 39...Qf5 was safer, but now I am not so sure: 40.Qxf5 exf5 41.Ng5!

41.fxg7 is weaker: 41...a5! 42.Kf4 a4 43.Bc1 (43.e6 Rc7!) 43...Rc7 44.Ba3 Kxg7 (44...Rc3 45.Bf8! a3 46.Nh4! is an illustration of the dangers of underestimation) 45.e6 Rc3 (45...Kf6 46.e7 Rxe7 47.Bxe7+ Kxe7 48.Kxf5 a3 49.Nd2 a2 50.Nb3 Kd6 51.Na1+-) 46.e7 Kf7 47.Bb4 Rb3 48.Ng5+ Ke8 49.Ne6! Rxb4 50.Nc7+ Kxe7 51.Nxd5+ Kd6 52.Nxb4 a3 53.Ke3 Kc7 54.Kd3 Kb6, and Black escapes;

a) 41...Bxf6 seems to lose: 42.Nxf7 Kxf7 43.exf6 Kxf6 44.Kf4 a5 45.Bd2 a4 46.Bb4 Ke6 47.Ba3! (White needs to get the bishop away from the king march: 47.Kg5? Kd7! 48.Kxf5 Kc6 49.Ke5 Kb5, and Black wins a crucial tempo) 47...Kf6 48.Bb2! b6 (48...Kg6 49.Ke5 Kg5 50.Kxd5 f4 51.Ke4 Kg4 52.d5 f3 53.Bd4 Kg5 54.d6 f2 55.Bxf2 Kf6 56.Kd5+-) 49.Bc1 b5 50.Ba3, and Black has run out of moves. 50...Kg6 51.Ke5 Kg5 52.Kxd5 f4 (52...Kf6 53.Bc5! with zugzwang) 53.Ke4 Kg4 54.d5 f3, and now the final finesse:

55.Kd3!! (55.Bc5 a3 would ruin all the previous work) 55...Kg3 (55...Kf5 56.Ke3 Ke5 57.d6+-) 56.Bc5 a3 57.Kc3+-;

b) 41...Bf8 42.Kf4 does not look very good either;

c) 41...Rc7! 42.Kf4! (42.fxg7 a5!) 42...Rc3! Black desperately needs to stop the king from getting to f5. The position remains very sharp, but I still prefer White's chances:

c1) 43.fxg7 Kxg7 44.Ne6+ Kg6! (44...Kg8 45.Bd2 Rd3 46.Bb4 b6 47.Kxf5, and it would seem that White is faster: 47...a5 48.Bd6 a4 49.Nc7 a3 50.e6+-) 45.Kf3 Rb3 46.Nf4+ Kf7 47.Nxd5 a5 48.Kf4 Ke6 49.Nf6 Rb4, and the a-pawn gives Black hope;

c2) 43.Bd2!? Rd3

c21) 44.Bc1 Rc3! (44...Rxd4+ 45.Kxf5 Rd1 46.Bb2+/-) 45.Bb2 Rb3 46.Bc1 Rc3=;

c22) 44.Be1!

c22a) 44...Rxd4+ 45.Kxf5 is too dangerous, as illustrated by 45...Rd1 46.Bb4 Bf8 47.f7+ Kg7 48.Bxf8+ Kxf8 49.Ke6! Rf1 (49...Kg7 50.Kd7!+-) 50.Kd7!+-;

c22b) 44...Bh6 45.e6 Rf3+ 46.Kxf3 Bxg5 47.f7+ Kf8 48.Bb4+ Be7 49.Bxe7+ Kxe7 50.Kf4 a5 51.Kxf5 a4 52.Kg6+-

c22c) 44...Rd1 45.Nf3!, and nothing will stop the king now. The following line shows just how dangerous White pawns can become when supported by the king: 45…Bh6+ 46.Kxf5 Rd3 47.Nh4 Rd1 48.Ng2! Bf8 49.Bh4! Rxd4 (49...a5 50.Kg6!) 50.Bg5, and White is well and truly on top.

40.Bf4?! With 5 minutes left and an important choice to make, Veselin decides not to commit before the time control - and lets perhaps the last winning chance go by. After 40.Qxe4! dxe4 41.Ng5 Black would face an unpleasant choice: perhaps his best chance is the ugly-looking 41...Rd7!?

a) 41...Bxf6 loses: 42.Nxf7 Kxf7 43.exf6 Kxf6 44.Kf4+-;

b) perhaps Veselin was worried about 41...Bh6 42.Nxf7 Bxe3 43.Nd8 a5! (43...Bxd4 44.f7+ Kf8 45.Nxe6+ Kxf7 46.Nxd4+-), but in fact after 44.d5! Bb6 (44...exd5 45.e6 Kf8 46.e7+ Ke8 47.Ne6+-) 45.Nxe6 a4 (45...e3 46.Kf3 a4 47.Ng5+-) White has an amazing resource: 46.Kf4!! a3 (or 47...e3) 47.Kf5, and White is first;

c) 41...Bf8 is trickier, but probably not enough: 42.Nxe6

c1) 42...a5 is slower, and even gives White a choice:

c11) 43.Ng5!? a4 (43...Bh6 44.Nxf7 Bxe3 45.Nd8+-; 43...Rc7 44.e6 a4 45.Bf4+-) 44.Nxf7 Kxf7 45.d5 a3 46.Bd4+-;

c12) 43.Nxf8!? Rxf8 (43...Kxf8 44.Kf4 a4 45.d5 Ke8 46.Bd4 Rc7 47.e6+/-) 44.Kf4 a4 45.d5 a3 46.Bd4 a2 (46...b5 47.e6 b4 48.d6+-) 47.e6+-;

c2) 42...b5

c21) 43.Ng5 b4 44.Nxf7 Kxf7 45.d5 b3 46.Bd4 (46.Bc1 a5 47.Bb2 Bc5 48.e6+ Kg6, and Black seems to be hanging on: 49.Be5 a4 50.f7 Bf8 51.d6 b2 52.Bxb2 Bxd6+ 53.Kg4=) 46...Ba3 47.e6+ Kf8 48.f7 b2 49.Be3 Kg7 50.Bd4+ Kf8 is a beautiful draw found by GM Alexei Korotylev;

c22) 43.Kf4! b4 44.Kxe4 a5 (44...b3 45.Bc1 a5 46.Nxf8 Rxf8 47.d5 also looks very strong) 45.d5 a4 46.Nxf8 Rxf8 47.Bc5! Rc8 (47...a3 48.d6!+-; 47...b3 48.Ba3+-) 48.Bd6! a3 (48...b3 49.Ba3) 49.e6, and once again White rams his pawn fist down Black's throat.

d) 41...Rd7!?

42.fxg7 (42.Nxe6 Bxf6! 43.exf6 Rd6 44.Nc5 Rxf6 45.Nxe4 is good for White, but with the king decentralized Black has decent chances to survive) 42...a5! (42...Rxg7 43.Kf4 Re7 44.Kxe4 a5 45.d5 exd5+ 46.Kxd5 should be quite hopeless) 43.Nxe6 a4 44.Kf4 a3, and precision is required: 45.Bd2! (45.Bc1 a2 46.Bb2 Rf7+! 47.Kxe4 Rf2 48.Ba1 Rf1 49.Bb2 Rb1 50.Bc3 Rc1=) 45...a2 46.Bc3 b5 47.Kxe4, and once again White seems to win: 47...b4 48.Ba1 (48.Bb2 Ra7) 48...Rf7 49.Nc5 Rf1 50.Nb3 Rb1 51.e6!+-. Of course, these lines do not pretend to be the final verdict on this endgame, but it is clear Black would be struggling.

40...Qf5. Computer suggests that White was not threatening anything special, and it was possible to play 40...a5, and it is most probably correct - but why allow positions like 41.Ng5 Qxd4! 42.Qxe6! (42.Nxf7 Qg1+ 43.Kh3 Qxg4+ 44.Kxg4 Kxf7 45.fxg7 b5!, and Black has nothing to worry about) 42...Qg1+ 43.Kh4 Qe1+ (weaker is 43...Qh1+ 44.Kg4 Qd1+ 45.Nf3 - the bishop is better placed on f4) 44.Bg3 (44.Kg4 Qe2+ 45.Nf3? Bxf6-+) 44...Qh1+ 45.Kg4 Qd1+ 46.Nf3. What's more - in these lines all Black can hope for is a perpetual, while in the endgames one mistake from White could land him in trouble.

41.Qxf5?! White should have started with 41.Ng5, and after 41...Rc7 he should probably have taken the draw:

a) 42.Nxe6 Rc3+ 43.Kh4 Qh7+ 44.Kg5 Bh6+ 45.Kh5+ Bg7+=;

b) after 42.Qxf5 exf5 play becomes double-edged, even if draw is still the most likely result:

b1) 43.Ne6 Rc3+ 44.Kh4 Bxf6+! (44...Bf8 45.Kg5 looks very dangerous to me) 45.exf6 Kf7 is not good for White;

b2) 43.fxg7 a5 44.e6 Rxg7 (or 44...Rc8 45.Bd6 Kxg7 46.Kf4 Kg6 47.e7 Kf6 with a draw, as shown by GM Korotylev) 45.Bc1 b5! 46.Kf4 Rc7, and White has to sacrifice his bishop to make a draw: 47.Kxf5! Rxc1 48.e7

b21) 48...Re1 49.Kf6 Re3! (49...b4? 50.Nf3!!+-) 50.Ne6 Rf3+ 51.Ke5 Re3+ 52.Kf6 Rf3+=;

b22) 48...Rc8 49.Ke6 a4, and now White escapes with a study-like 50.Kd7! Ra8 51.Ne6 a3 52.Nc7! Rb8 53.Na6! Rb7+ 54.Nc7 Rb8=.

But as we've seen already in Game 1, Veselin continues pressing for a win even when the position no longer affords it.

41...exf5.

42.Bg5? With the same do-or-die attitude. After 42.Ng5 Black has a wider choice than a move before: 42...Rc7 transposes to the previous note, while he can also consider 42...Rxf6 43.exf6 Bxf6 with a probable draw, but not 42...Bf8!? 43.Kh4! b5 44.Kh5 b4 45.Kg6 b3 46.Nxf7 b2 47.e6!! b1Q 48.e7 Qg1+ 49.Bg5, and White wins. It is hard to say what Veselin could have missed - after all, Black's reply to 42.Bg5 is fairly obvious. Perhaps he simply got carried away.

42...a5. And the rest is history.

43.Kf4 a4 44.Kxf5 a3 45.Bc1 Bf8 46.e6 Rc7 47.Bxa3 Bxa3 48.Ke5 Rc1 49.Ng5 Rf1 50.e7 Re1+ 51.Kxd5 Bxe7 52.fxe7 Rxe7 53.Kd6 Re1 54.d5 Kf8 55.Ne6+ Ke8 56.Nc7+ Kd8 57.Ne6+ Kc8 58.Ke7 Rh1 59.Ng5 b5 60.d6 Rd1 61.Ne6 b4 62.Nc5 Re1+ 63.Kf6 Re3 0-1