|[Before this game, it seems incredible that Kasparov can save this match. If my memory serves me well, the last time Kramnik scored 1/2 point out of 3 was in 1994, at the very beginning of hs disastrous F.I.D.E. Candidates' cycle: he lost the first two games in succession, another experience I don't recall him ever having since.
On the other hand, I would bet at least 100 to 1 before the match against Kasparov not winning even one game out of 16 -and who would dare take me on? The World Champion has proved capable of rising in difficult moments, including a memorable last round win against his present Challenger in Linares 1997. Even though his premature draw offer in Game 13 indicated he does not possess enough energy any longer, I refuse to accept the argument that renders his age a significant factor in his performance. This would not only insult Kramnik's excellent psychological match strategy, but also be simply untrue: 37 is not such an advanced age for a 16-game match and the teacher is reputedly in much better physical shape than his former pupil.
All bets are on!]
[As I had already predicted several times in my comments on earlier games, Kramnik switsches over to an even more solid approach for his last two Whites. Reportedly, however, Kasparov now grabbed his head with both hands and started thinking.
Since 1.Nf3 could not have been such an unexpected choice from the Challenger, this sudden physical response from the Champion necessarily betrays his need for concentration. Consequently, one may conclude he feels very determined today to try go where he has never been before: win a game with Black against Kramnik under normal rates of thinking time.]
[This time it is Kramnik who opts for some version of the English/Reti Opening complex. It is a very solid approach, with a lot of tricky move orders that allow or forbid favourite variations of both players. The most important aspect of the battle to come concerns the placement of the d-pawns: where and when will they advance?]
[A smart choice, avoiding the kingside fianchetto and at the same time keeping maximum flexibility in regard to his central Ps.]
[Another obvious possibility is 3...Bb7 , but then White could transpose to the Queen's Indian Defence with 4.d4 . This might not be a totally unwelcome prospect for Kasparov, especially as he also allowed it with his second move, but now the subvariations where Black plays an early ...Bc8-a6 to put pressure on c4 and unbalance the position would no longer be feasible.]
4.Bg2 Bb7 5.0-0 g6
[The double fianchetto was chosen by the very same Kasparov in the critical 23rd game of his 1986 London/Leningrad World Championship match against Karpov. True, there he was more than happy with a draw in what was to be his final Black game, because he was leading by one point, while now he desparetely needs a win. Still, it is nice to use a variation that brings happy memories.]
[Karpov played 6.d4 , commiting himself first and demonstrating that he wanted more space in the centre.;
Another option worth noting is 6.b3 , keeping maximum symmetry and thus more chances that the game will be eventually drawn. It has been well established in practice that the best way to split a point is to play "normal" chess rather than "drawnig" chess, but White can almost always choose among a wider range of alternatives. In this case, I believe it is better to delay the giving of any unnecessary information to the opponent. A likely continuation would be 6...Bg7 7.Bb2 0-0 , when 8.Nc3 is quite OK.]
[A very important decision that leads the game back to the 1986 K-K encounter. Once again 7.b3 is possible, as after 7...Ne4 8.Bb2 Nxc3 9.Bxc3 Bxc3 10.dxc3 the doubling of the pawns is not at all negative for White: a semi-open file and better control of the centre, combined with a relative weakness of the black squares on Black's kingside, would then give White a definite advantage.
Kramnik decides to go for the "hedgehog" formation, even though its asymmetry is what Kasparov is really looking for. Personally, I think it would be better to wait for Black's ...d7-d5 and then exchange everything via a timely d2-d4. If both players keep a waiting stance, it is possible to play e2-e3, Qd1-e2, Ra1-c1, Rf1-d1 and only then d2-d4, in order to be able to choose better whether to recapture on d4 with a P or a piece, according to circumstances.
It is likely Kramnik wants to remove part of the psychological tension by making a significant choice rather than maintaining the present equilibrium, but this gives Kasparov the type of position he has been looking for throughout the match. It is really a strange moment to change match strategy!]
[Now the outlines of the middlegame begin to take shape, since the pawn structure will remain more or less stable for quite some time. White controls more space, at least for the time being, but Black possesses an extra central P.]
[Recapturing with the Q is generally thought to be strongest, because it promotes development and avoids exchanges. However, it is also possible to play 8.Nxd4 Bxg2 9.Kxg2 , "threatening" to establish the so-called Maroczy Bind with 10.e4. The game Karolyi-Horvath, Hungary 1986, continued 9...0-0 (objectively better seems 9...Qc8 , e.g. 10.Qd3 Na6 or 10.b3 Qb7+ and 11...d5 with satisfactory play for Black, although the massive exchanges that will inevitably follow in the latter sub-variation should lead to an easy draw) 10.e4 Qc7 11.b3 Nxe4 (here Fernandez-Psakhis, Novi Sad 1990, diverged with 11...Na6 12.f3 Qb7 13.Be3 Nc5 14.Rc1 Rac8 15.Nd5 , when White's space superiority was undisputable) 12.Nxe4 Qe5 13.Qf3 Qxd4 14.Rb1 Qe5 15.Bf4 Qa5 16.Nf6+ Bxf6 17.Qxa8 Nc6 18.Qb7 Qf5 19.Rbd1 Bd4 20.Rfe1 e5 21.Be3 Bc3 22.Re2 Qh5 23.Rxd7 Qxe2 24.Qxc6 and White stood much better because of the placement of his pieces.]
[This is an indication that the Nb8 will be developed on d7, leaving open the h1-a8 diagonal and the c-file. Attacking the Q with 8...Nc6 does not really gain a tempo, as eventually the Q will want to move away from hidden attacks on the a1-h8 diagonal. The only advantage of hitting it at once is that White is then forced to choose relatively early where to remove her, before more information about other Black pieces is revealed.
Nevertheless, the N may prove to be far away from the kingside and one spectacular example should suffice to demonstrate the dangers of such an approach: 9.Qf4 Rc8 10.Rd1 d6 11.b3 Ne4 12.Nxe4 Bxa1 13.Ba3 Bg7 14.Nfg5 0-0 15.Nxh7 Kxh7 16.Ng5+ Kg8 17.Qh4 Re8 18.Bh3 Kf8 19.Ne6+ Kg8 20.Ng5 Kf8 21.Ne6+ Kg8 22.Nxd8 Rcxd8 23.Bg2 Bf6 24.Qh6 e6 25.h4 d5 26.h5 Bg7 27.Qg5 dxc4 28.Rxd8 Rxd8 29.bxc4 gxh5 30.Bb2 e5 31.Bd5 Rd6 32.Qxh5 Rd7 33.Qf5 and 1-0 in Ribli-Kouatly, Lucerne 1985.]
[Naturally, 9.b3 is still possible and has been essayed on numerous occasions.]
[The 1986 K-K encounter mentioned above reached the same position via a slightly different move order and continued 10.b3 Rc8 (10...0-0 11.Qh4 can be dangerous, as the Bc1 may come to h6) 11.Bb2 0-0 12.Qe3 Re8 13.Rac1 a6 14.Ba1 Rc5 (this R manouvre seems to equalize completely) 15.a4 Qa8 16.Ne1 Rf5 17.Bxb7 Qxb7 18.f3 h5 19.Ng2 Rc5 20.Bb2 Rcc8 21.Ba3 Nc5 22.Rb1 Ne6 23.Qd3 Nc7 24.Nf4 b5 (finally, this typical pawn break moves the battle over to the queenside, even if it has to be made with a temporary P sacrifice) 25.cxb5 axb5 26.Nxb5 Nxb5 27.Qxb5 Qxb5 28.axb5 Rb8 and was eventually drawn, allowing Kasparov to keep his recently acquired title of World Champion.]
10...Rc8 11.Rac1 0-0
[Now that the Bg7 is protected, the enemy Q must move without delay...]
[...and visit the kingside at just the right time. An earlier excursion to h4 could have been met by ...h7-h6 with the idea ...g6-g5, when the Q could get quite embarrassed. It should be kept in mind that Kramnik has not played b2-b3, so the Q also has to perform the duty of defending the c-pawn.]
[The advance ...a7-a6 completes the "necessary" moves that both sides have to make in preparation for the middlegame. Although White enjoys more space, Kasparov has found comfortable spots for all his minor pieces and has kept the Qs on the board in an asymmetrical P structure.
The big question for Black is to find a way for the major pieces to enter the battle. Kasparov's R manouvre on the 5th rank in the game against Karpov was a good idea, but it should not work for more than equality. It is necessary to try for the pawn break ...b6-b5 and create some counterplay against c4, while refusing any massive liquidation. Kramnik, on theother hand, face a much more complex dilemma: he stands slightly better overall, but with no clear plan and a strong subjective wish to equalize the position.]
[This is not as inconsistent as may seem at first sight. The Challenger goes for the exchange of white-squared Bs, which he refrained from earlier, at a point where he has more information about his opponent's pieces. For example, the development of the Nb8 on d7 means that c6 may become weak, especially after the transference of a white N on d5, a subsequent exchange and recapture with the c-pawn.
On the other hand, the very same exchange releases some of the congestion in the Black camp and provides the Q with a most useful diagonal.
It must have not escaped Kasparov in his preparation that the encounter Kramnik-Karpov, Dos Hermanas 1999, had continued 13.b3 Rc7 14.g4 (the very same manouvre of the main game 14.Ne1 Bxg2 15.Nxg2 had been tried by Kramnik's official second in the game Illescas-Gelfand, Pamplona 1999, after which there followed 15...Re8 16.Nf4 Qb8 17.Nfd5 Rcc8 18.Bg5 Qb7 19.a4 with White keeping a distinct space advantage but failing to make it tell: 19...Kh8 20.g4 Nxd5 21.Nxd5 e6 22.Nf6 Nxf6 23.Bxf6 d5 24.Rc3 Bxf6 25.Qxf6+ Kg8 26.Rcd3 Qe7 27.Qxe7 Rxe7 28.cxd5 exd5 29.Rxd5 Rxe2 30.Rd8+ Rxd8 31.Rxd8+ Kg7 32.Rd6 Re4 33.Rxb6 Rxg4+ 34.Kf1 Rh4 35.a5 Rh3 36.b4 Rb3 37.Ke2 g5 38.Kf1 h5 39.Rxa6 Rxb4 40.Rb6 Ra4 41.a6 and ½–½) 14...h6 15.h3 g5 16.Qg3 b5 17.h4 bxc4 18.hxg5 hxg5 19.b4 with an extremely unbalanced position. However, what is good against Karpov in a tournament situation is not necessarily optimum against Kasparov when a draw is required...]
13...Bxg2 14.Nxg2 Re8
[Apparently Kasparov thought for quite some time on this one, but I do not believe it had to do so much with the move itself as with the overall plan for the next few moves. It is evident that 14...Qc7 15.Nd5 favours White, so the e-pawn is protected in order to execute thw Q manouvere without having to exchange Ns on d5. In addition, Black may now avoid the exchange of black-squared Bs after Be3-h6, if he wishes so, with ...Bg7-h8.]
[Sooner or later, this move is practically necessary, to protect the c-pawn in an economical way and release the Q from a duty unworthy of her. However, this casts some doubt on the correctness of the plan of development with Bc1-e3; perhaps the B should have been fianchettoed, after all, or at least developed to h6 in one movement.]
[Now it was Kramnik's turn to think for a while. 16.Nd5 Qb7 would be useless, but it also made sense to try;
16.Nf4 as in the Illescas-Gelfand game mentioned earlier. His black-squared B does not seem to know what it wants.]
[The Q is perfectly placed here: she replaces the exchanged B, supports the desirable advance ...b6-b5, may provide additional support to the e-pawn and in general cooperates with the rest of her forces much better than her counterpart. Kasparov must have been satisfied at this point, especially if one compares the position with his other "Black" games from the match.]
[The only way to prevent the following advance is 17.a4 , but then the b-pawn is irreparably weakened and Kramnik would miss the white-squared B or the Q from the queenside. Kasparov would almost certainly continue 17...Nc5 18.Rb1 Nfe4 assuming the initiative, while the ...b6-b5 break would not be long in coming, anyway.]
[The time has come to cross the 6th rank! Just give Kasparov a pawn break and he knows how to improve the mobility of his pieces without fail. Now Black has equalized completely and Kramnik must find a way to neutralize his opponent's plan of further expansion.]
[The exchange 18.cxb5 axb5 is an important alternative, as after 19.Qb4 White can attempt to take advantage of the newly created queenside P majority. The problem is that Black can play 19...Rc5 , protecting the P and attacking the enemy B, while at the same time preparing the doybling or even tripling of major pieces on the c-file. How different all that would be now if instead of 16.Bg5 Kramnik had chosen 16.Nf4 and then Nf4-d3!;
However, I do not quite understand why 18.Ncd5 is not better than his present choice, allowing White to capture on c4 with a piece every time. The only possible objection I can raise is that he would be eventually obligated to exchange his B for a N, e.g. after 18...bxc4 19.Rxc4 Rxc4 20.Qxc4 h6 21.Bxf6 (of course, not 21.Nxf6+ exf6) 21...Nxf6 , but it seems to me that control of the only open file after 22.Rc1 should more than enough compensation.]
[Now Kasparov has an isolated P to bite on, with Kramnik's only cosolation being the proud N on d5. This, however, ca be kicked away with ...e7-e6 at an appropriate moment. Kramnik is much better at nursing small long-term advantages than in transforming a temporary initiative to something more concrete.]
[An interesting choice, possibly revealing more abot Kasparov's way of thinking than he would like to say explicitly. In a tense situation he plays a seemingly neutral move, asking his opponent what he wants to do if left undisturbed. In effect, he tells him there is no real compensation for the damage that was self-inflicted on White's queenside P formation.
The advance ...h7-h5 is not without its purely chessic aspects, but these are neither immediate or apparent to the naked eye: it prevents once for all the idea Bg5-h6 and creates ideas of ...h5-h4 later on, when both the Q and the B will be far away from their present posts. Kasparov must have anticipated a major piece endgame, where the advance of the h-pawn can be of paramount importance.
In any case, 19...Qc6 would not make a great impression after 20.Nxf6+ Nxf6 21.Nd5 etc. Reportedly, at this point Kramnik had 59 minutes left to reach move 40, while Kasparov only 37.]
[Kramnik almost caught up in time with Kasparov after this. He was probably trying to fathom the depths of the Champion's thoughts, but 20.Nxf6+ Nxf6 21.Nd5 was much worse now, anyway, because of 21...Nxd5 22.cxd5 Rxc1 and the black pieces are free to roam the empty queenside avenues, e.g. after 23.Bxc1 Rc8 etc.]
[Now things are quite different, because the Q is much more vulnerable on f4.]
[For example, 21.Nxf6+ can now be answered by 21...exf6 22.Qxd6 (all other replies also lose a piece, for example 22.Bh6 g5; or 22.Rxd6 fxg5 23.Qd2 Qxc4 24.Rxd7 Bxc3) 22...fxg5 (22...Qxd6 23.Rxd6 fxg5 is equally satisfactory) 23.Qxd7 (23.Qxc6 makes no difference) 23...Qxd7 24.Rxd7 Rxc4 25.Rd3 Rec8 and the pinned N is lost.
Therefore, Kramnik is forced to exchange the unfortunate black-square B after all, casting serious doubt on his earlier decision 18.Ned5.]
21...Nxf6 22.Nxf6+ Bxf6 23.Nd5
[Once again White gets a proud N on d5, but once again this strong square is not permanently guaranteed because of the possibility of ...e7-e6. Besides, the weakness of the c-pawn is bound to make itself felt eventually.]
[A useful intermediate move, forcing the R to declare its intentions.]
[24.Rc2 looks more sensible, keeping an eye on White's greatest weakness.]
24...Bg7 25.Qg5 Kf8
[Another good defensive move, demonstrating how the compactness of Black's position allows an easier allocation of active and passive roles among the pieces.
By now both opponents had about 25 minutes each, which promised an exciting time scramble.]
[This was Kramnik's idea when he played 24.Rb1, in effect moving the Rd1 over to the b-file. This pleasure, however, has the cost of releasing potential pressure down the d-file, thus prompting Kasparov to play...]
[The right move at the right time. 26...Qc5 would create a much stronger threat of 27...e6, but after 27.h4 all Black would have ahieved is to leave the entry square b7 unprotected.]
[27...Re7 is simply wrong, because there would follow 28.Nh7+ Ke8 29.Nf6+ and the King must return to f8.]
[Now 28.Nh7+ Kg8 29.Nf6+ Kh8 is worse than useless. At least, the King is more exposed to tactical counterchances on f8. Defending the Q is quite important, as will soon become clear.]
[An impressive attacking retreat, it creates the nasty threat 29...Rc5 30.Qf4 Rf5. Kramnik cannot afford to sit by idly.]
[The Challenger believes he must get rid of his primary weakness, but I am not at all sure this is necessary. 29.Qf4 would have maintained material equality, albeit at considerable discomfort. Still, White's counterpressure against the d-pawn is not to be underestimated.]
[Perhaps it was this specific move order that Kramnik had overlooked when he played 28.h4. After 29...dxc5 30.Rxc5 Bxf6 White escapes with the intermediate captures 31.Rxc8 Bxg5 32.Rxa8 , as 32...Rxa8 33.hxg5 justifies the advance of the h-pawn completely.]
[Of course, not 30...dxc5 as the capture 31.Qxc5+ comes with check.]
[All of a sudden the World Champion has an extra passed P and visions of his title staying "at home" must dominate the thoughts of many heads. Both sides have active pieces and the only serious problem for Black is his somewhat exposed King, for example there is the idea Rb1-b6xe6, regardless of whether the King stands on e8, f8 or g8.]
[Naturally, not 32.Rb6 Rd1+ and mate next move. Removing the King from potential danger is very important in similar positions...]
[...which is probably why Kasparov follows suit, but this is not necessarily good.
Immediately after a transitional phase there comes another of reassessment and consolidation, in which a player must adapt to the new circumstances and order his priorities straight. Maybe this is why Kramnik provoked a sudden change in the position, since the player that provoked the transformation of the game is psychologically better prepared to switch to a new mental approach. Here Kasparov is on the receiving end, otherwise he would most certainly remember that his idol of earlier days Alekhine, who had specialized in battles betwen major pieces, always stressed activity before King safety.
The important thing here is to understand that a Q ending is technically (if not easily) winning for Black, because then the passed P can be escorted to the queening square unhindered. On the other hand, the result of a R ending will depent first and foremost on the specific position of the Rs.
What this means here is that the indicated approach for Black consists in first improving the position of the pieces, possibly advancing the P a little in order to provoke some retreat by the opponent, and only later playing it safe with the King. White can create serious counterchances if he can coordinate both his pieces, but never enough with only one of them.
Therefore, 32...Qd5 is the correct approach, when the idea 33...Qd4 more or less forces White's hand. 33.Rb6 is then followed by 33...Ra8 (33...c4 34.Rxe6 Qxe6 35.Qxd8+ Kg7 36.Qd4+ Kh7 and 37...Qc6 might also offer winning chances, but is unnecessarily more complicated) 34.e3 (prevents 34...Qd4, while a typical variation is 34.Rxe6 Qxe6 35.Qh8+ Ke7 36.Qxa8 Qxe2 37.Qa7+ Kf6 38.Qxc5 Qxa2 39.Qd6+ Kg7 40.Qe5+ Kh7 etc.; and 34.e4 Qd4 35.e5 Kg8 36.Rb7 Rf8 should also lead eventually to a win for Black) 34...Kg8 (immediately 34...c4 is also possible, e.g. 35.e4 Qxe4 36.Rxe6 Qxe6 37.Qh8+ Ke7 38.Qxa8 , but the King retreat looks safer now) and the plan of advancing the passed P with the support of both Q and R looks realistic now, for example 35.a4 c4 36.Qc3 Qf3 with the twin threat 37...Qxf2+ and 37...Rd8 with possibly ...Rd8-d1 to follow at the appropriate moment.]
[If Kasparov eventually loses the match, one critical factor might prove to be his relative weakness in the transitional phase between the middlegame and the ending. Games 2, 4 and 6 were particularly painful in that respect, even if in the end he managed to lose only the first among them: the others seem to have taken their toll in energy, an energy that was much needed in Games 7 and 13.]
[Just in time to prevent the cooperation of Black's pieces with the passed P.]
[Now 33...Qd5 proves useless after 34.Rxa6 c4 35.Ra7 , as 35...e5 36.Rxf7 is unplayable and 35...Rf8 36.Rc7 undesirable. However, now the R is too passive and White gets an extra opportunity.
By this time, Kramnik was down to 7 minutes and Kasparov to about 9.]
[Theoretically and practically the best chance. If Black is allowed to consolidate, he will eventually repulse the enemy pieces and begin coordinating his own. Transposing to the R ending is Kramnik's ideal option, since his own R is active and can inflict maximum damage on the enemy queenside Ps.] 34...Qxf3 [Because of the threats 35.Rb7 and 35.Rc6, Kasparov has no real choice.]
[Of course, not 35...a5 36.Rb5 (the trap 36.Rc6 Rb8 -or 36...Rd8- 37.Rxc5 Rb2 38.Rxa5 -theoretically superior is 38.Kg2, when White will suffer- 38...Rxf2+ is too transparent, even for zeitnot) 36...Rd8 37.Rxa5 with an immediate draw.]
[Material is equal again, but the black R is better posted and thus Kasparov will once more win a P. However, the situation has been greatly simplified...]
[Exactly the same position will occur after 37.Rb6 c3 etc.]
37...c3 38.Rd1 Ra8
[The immediate advance 38...c2 is completely silly, since after 39.Rc1 the P is going nowhere, so that the enemy King can simply approach and take it.]
39.Rc1 Rxa2 40.Rxc3
[The last move before the time control is often wrong, but here Kramnik correctly refrains from 40.Kg2 Ra3 (the immediate 40...c2 leads to exactly the same position after 41.Kf1 Kg7 42.Ke2 Kf6 43.Kd2 g5) 41.Kf1 Kg7 42.Ke2 (42.Rc2 loses an important tempo and allows the much easier win 42...Kf6 43.Ke2 Kf5 44.Kd3 e5 45.Rxc3 -or 45.Ke3 Rb3 46.Kd3 Rb2 etc.- 45...Rxc3+ 46.Kxc3 e4 47.fxe4+ Kxe4 48.Kd2 Kf3 49.Ke1 f5 50.Kf1 f4 , when White will soon lose the h-pawn) 42...Kf6 43.Kd3 c2+ 44.Ke2 (a grave mistake would be 44.Kxc2 Rxf3 45.Rf1 Kf5 46.Kd2 Kg4 47.Ke2 e5 48.Ra1 e4 49.Ra4 f5 50.Ra6 Kh3 51.Rxg6 Kg2 etc.)
44...Ra2 45.Kd2 and now the simplest is 45...Kf5 , e.g. 46.Rxc2 Rxc2+ 47.Kxc2 f6 48.Kd3 g5 49.Ke3 gxh4 50.gxh4 e5 and the P ending is always lost because of the reserve tempo with the f-pawn: 51.f4 (equally hopeless is 51.Kd3 Kf4 52.Ke2 f5) 51...exf4+ 52.Kf3 Ke5 53.Ke2 Ke4 54.f3+ (or 54.Ke1 Kf3 55.Kf1 f5 56.Kg1 Ke2 57.Kg2 f3+ 58.Kg1 f4) 54...Kd4 55.Kd2 f5 etc.]
[Now "all is quiet in the western front" and the battle will be decided in the east.]
[Perhaps luckily for the Challenger, he must make anothe important choice immediately after the time control. The best practical approach in such situations is to take a small break, relax physically from the tension and only then try to concentrate again.
The truth is that Kramnik should be able to draw the R ending with eyes xlosed, if he wants to become World Champion. One can only imagine Kasparov's frustration to have reached only this theoretical ending after such a magnificent effort in the middlegame, but life is really tough at moments like this. At least to one of the contestants...]
[It might seem strange to allow Black the "absolute 7th rank", but the most important aspect of the position is to prevent the Black King from joining the attack. This means that the R must remain active and not become tied down to the defence of the f-pawn after 41.Kh3 .]
41...Ra2 42.Rc7 Kf8 43.Rb7 Ke8
[The only way to make progress is to advance the King, but then the Ps will be unprotected. Therefore, they have to advance too, but then some of them will be exchanged. All Kramnik needs to do is to take care that Kasparov does not create a passed e-pawn under favourable circumstances.]
44.Rb8+ Ke7 45.Rb7+ Kf6 46.Kf1
[I do not quite understand this. After 46.f4 Black cannot make progress without exchanging the e-pawn and therefore a draw should be easy, for example 46...Kf5 47.Rxf7+ Kg4 48.Rf6 and Black must agree to a draw.]
[Now Kasparov has kept some tension in the position and can still hope...]
47.Rb6+ Kf5 48.Rb7 Ke6 49.Rb6+ Kf5 50.Rb7 f6
[At last, but it may not be enough...]
51.Rg7 g5 52.hxg5 fxg5
[This typical move creates a much needed greater distance between the defending R and the attacking King, so that the latter cannot easily respond to checks with counterthreats to the R. Now the position is likely to be drawn, as Kasparov aknowledges with his next few moves.
A horrible mistake would be 53.Ke1 Rg2;
53.Kg1 allows Black to keep on fighting with 53...g4 , although even then with best defence ( 54.Rf7+ ) a draw is assured.]
[There is no better way to attempt to make any progress. Even in the endgame, even for a World Champion, pawn breaks are the most fundamental way to improve one's position!] 54.Rf8+ Ke6 55.Re8+ Kf5 [It is clear that this leads nowhere, since the R can keep on ckecking. Much more sensible is 55...Kf6 , trying to squeeze the R in the narrow space of one of the e-, f- or g-files.
I do not know exactly what happens in that case and I do not have the time to analyse it (sorry, I have to go back to work!). Nevertheless, what I cannot understand at all is why Kasaprov did not even try this idea. He had nothing to lose and everything to gain!]]
56.Rf8+ Kg6 57.Rg8+ Kf5 ½-½
[Now that Kramnik only needs 1/2 a point from the remaining 2 games, the match is nearly over. It seems very unlikely that Kasparov will manage twice in concecutive attempts what he has not yet managed once over 14 tries. No matter how close he got today to his intermediate goal or how close he might still get if he wins on Thursday, the fact remains that Kramnik has proven too solid to lose 2 games in a row. Really, it seems much more probable that he will lose neither than that he will allow Kasparov to tie the match.
The Challenger's main goal in the next two days is to find the correct balance between too much tension and too much relaxation, now that he is close to his life dream. The Champion's prospects can stay alive only if he manages to mobilise his full reserves of energy, something he proved incapable of doing on Sunday. Still, both players must be getting very tired by now, a situation quite conducive to a higher rate of mistakes and thus unpredictability.
If the chess is not top level in the next two games, at least it is likely to be exciting. Some more agony around for everybody!]