|[After 3 games, the match has acquired a pace of its own -and Kramnik is in the driver’s seat. Not only is he leading 1-0 with two draws, but he also appears to be in full control. He has committed no serious mistakes and been in very little danger. He has managed to steer the games in types of positions that do not suit Kasparov at all, without active piece play or serious pawn breaks. It is probably significant that with Black he goes for Queenless middlegames, that do not allow his illustrious opponent to display his imagination and combinative skills.
But Kramnik has not really shown anything new in chess yet. Kasparov has extensive experience in World Championship matches and lots of stamina, so he cannot be beaten simply by avoiding his strong points. To become Champion, at some point Kramnik will have to play like a Champion and defeat Kasparov in complicated battles. He must show that he is at least equal in tactical situations repeatedly, as he did in the second game between them in Linares this year.
Meanwhile, he is White again and both players have to choose an appropriate opening. I believe Kramnik will stick to 1.d4, probably the strongest move against Kasparov anyway. He was successful in his first White game and cannot complain at all about the position he got. Since he demonstrated readiness to engage in a theoretical discussion with Black, especially by using the rather unorthodox Berlin Defence to the Ruy Lopez, there is no reason why he should avoid a similar discussion with White.
Kasparov has a bigger problem. Perhaps for the first time in his life, he finds himself in a special situation early in an event: he cannot afford to lose. Losing his second Black game would create a serious crisis in his camp and make the score almost untenable after only four games! His choice lies among three major alternatives: playing again the Grunfeld Defence, switching over to the Queen’s Gambit, and introducing a third surprise opening. I do not quite believe he will go for the King’s Indian, because he has not been successful with it against Kramnik in the past. Besides, if he had included the King’s Indian in his match repertoire, he should have probably tried it in his very first Black game. The risk now would be too great.
Let us examine then each choice in turn, but in reverse order. Introducing a new opening, like the Benko Gambit (which suits his style perfectly but has limited potential) or the Leningrad Dutch, could create as a big surprise for Kramnik as the Dragon did for Anand in 1995. The problem with such an approach is twofold. First, Kasparov needs to believe in the correctness of his openings and not simply their practical advantages, a consideration that severely limits his choices (for example, his old favourite the Benoni, is most probably out). Second, also important, Kasparov has shown a tendency to introduce some serious surprise after the middle point of a match, at the psychologically most crucial turning point. To «burn» it so early would be undesirable.
The Queen’s Gambit is a safe good option, but it indicates limited aspirations with Black. It would mean that the pace of the match remains slow, with Kasparov and his team preparing to hit back with a vengeance in Game 5. This approach has the added advantage that tension would be built gradually, something that should favour Kasparov. Its problem? Kasparov is not familiar with a safety first approach, having limited himself to asymmetrical openings in recent years.
The Grunfeld Defence remains Kasparov’s primary alternative. It entails some risks, but is consistent with his principles and maintains the theoretical debate mode of the match. If Kramnik needs to play like a Champion to become one, so does Kasparov to retain his title! The problem here is that historically the Grunfeld has been better for Kasparov in tournaments than in matches.
What will Kramnik do? The advantage for White these days lies not only in the first move, but also in the ability to choose more freely the type of position that will occur in the middlegame. In most cases, Black decides about the opening that is used in a game, but White chooses the specific variation.
In the case of a surprise opening or the Queens Gambit, Kramnik will simply go for the variation he has considered before the match. The Grunfeld Defence is much trickier, exactly because he has won the first game with it. It is not so easy to decide exactly which variation to play, since psychological considerations become paramount.
Fischer said that the best psychology is the best moves, a theory fully accepted by Lautier, who is one of Kramnik’s official seconds. This means that the challenger should continue with what has been successful for him, until proven wrong, and wait for Kasparov to vary the game at a point of his choice. A strategy like this is more risky with Black, but now Kramnik can afford to use it with both colours.
On the other hand, Fischer did not always practice what he preached: after winning convincingly with the Benoni in Game 3 against Spassky in 1972, in his very next Black game he chose the NimzoIndian Defence, while later he also used the semi-Tarrasch in the Queen’s Gambit. Karpov has used the same approach when leading a match, that is after apparently winning a theoretical battle. A «moving target» attitude may destabilise the opponent with minimal risk, especially since Kramnik must have prepared seriously at least two variations against the Grunfeld. Then he can revert to his first choice at a later game.
These were my predictions before the game. One of the players had a big surprise in store for me and many others.]
1.d4 d5 2.c4 dxc4
[This is it! Discussing possible alternative before the game with some friends who suggested the Queen's Gambit Accepted for Kasparov, I was strongly against it, because I did not consider it suits his style well. Although it has gained in reputation recently, as even Anand has included it in his repertoire (there are games of the Indian GM against both Kasparov and Kramnik), it has never gained a wide acceptance at top level.
My most serious objection against the Queen's Gambit Accepted was not that using it wastes a possible suprise weapon that may be much more useful later in the match (as argued in my introductory remarks), but instead that it would allow Kramnik to pursue the match strategy he has followed consistently up to now: in many subvariations White can force a queenless middlegame with a slight advantage, in which Kasparov would suffer without ever hoping for more that the half point.
In addition, such a choice would admit temporary defeat in the theoretical debate on the Grunfeld.]
3.Nf3 e6 4.e3 c5
[Black cannot do without this freeing advance, which however allows White to engage in multiple exchanges and a symmetrical pawn structure without pawn breaks.]
5.Bxc4 a6 6.0-0 Nf6
[And here it comes! Kramnik once again practically forces the exchange of Qs, while still keeping open the possibility to try for the full point. Even if Kasparov manages to draw, he is likely to have to work for it. Spassky followed a similar approach when he played Fischer in 1992.]
7...Qxd1 8.Rxd1 Bxc5
[A position with symmetrical pawn distribution has occured, which means that a draw is quite likely. Nevertheless, White enjoys a slight lead in development and greater flexibility regarding the pawn structure on the queenside. Many players have used this approach in order to minimize the risk of losing, a consideration surely important to Kramnik at this point of the match.
The traditional centre of the board (squares d4-d5-e5-e4) has been transferred "west", with the open c- and d-files being much more important now. Both players have a strongpoint on the d-file, in order to block the invasion of an enemy R, while the c-file is still up for grabs. As is often the case, the fight for lines on which to activate the Rs shall be determined by the minor pieces, which control important entry points.
In this respect, the pawn structure on the queenside can become very important, because it may tie up some minor pieces in awkward defensive roles. For example, the early advance ...a7-a6 could force Black to post a N on d7, thus blocking the d-file. Then the b-pawn will have to advance one or two squares, in order to allow for the development of the white-squared B, and in this way weaken slightly some important points on the c-file.
Overall, the early simplification of the position may give it a deceptively simple appearance. In reality, there is still a lot of play left.]
[Spassky tried 9.b3 3 times against Fischer in 1992, with mixed results. Other significant options are;
9.Nc3 and; 9.a3 , each with its own special characteristics. Very often though, a transposition may lead from one subvariation to the other.
Kramnik had used 9.Nbd2 in the past, even if only in rapid games. Therefore, by choosing the Queen's Gambit Accepted Kasparov must have been ready for it. Its basic idea is to challenge for some black sqaures on the queenside by advancing to either b3 or c4, depending on Black's reaction, of course.] 9...Nbd7 [The only time Spassky chose 9.Nbd2 against Fischer, the American GM replied 9...0-0 10.a3 b5 . The extended fianchetto seems to gain some time for development and keeps control of c4, but on the other hand makes c5 more vulnerable and subject to occupation by an enemy N in the future.]
[Kramnik sugnals his intention to work on both the c- and d-files, either with Nd2-c4 or Nd2-b3. The B had no future on the a2-g8 diagonal, since Black has no intention of ever advancing ...e6-e5.]
[An important "small" decision, with large consequences. In such situations, both players must constantly choose between apparently equivalent alternatives without some concrete criteria.
I suspect this is also part of a very conscious match plan by Kramnik, with the express aim of making Kasparov feel uncomfortable. The World Champion has demonstrated supreme mastery in complicated positions, in which accurate calculation can lead to definite conclusions. His decisions then have a high degree of confidence, which in turn leads to better play. However, when Kasparov is forced to depend purely on positional intuition about the long-term characteristics of the position, he might feel less sure of himself. Any doubts about his ability to cope with the problems of the position may bring on time trouble and shaky play, or even unforced errors.
The above observations should not lead to the mistaken impression that Kasparov is weal in positions like the one in the game. Naturally he is strong in any type of position (otherwise he wouldn't be the strongest player of our time), but for Kramnik to beat him he must go for situations that do not allow Kasparov to display the full range of his enormous abilities.]
[In three rapid games, Kramnik has used 11.Nc4 insted. Since however his second Bareev has played 11.Nb3 in the past, Kasparov must have been ready for this eventuality as well.
Kramnik's present choice practically forces the retreat of the enemy black-squared B, but also creates some difficulties for the development of his own.]
[Neither side can afford to exchange one of its Bs for a N in such an open position. Both have kept great flexibility in their pawn structure, which means they are ready to deploy the foot soldiers in any way they need in order to block an enemy B -but not both. Generally, the most important factor as the game progresses will most likely be which piece exchanges are made and when.
Meanwhile, both armies possess somewhat strong squares for their Ns in the centre, but none that is permanently secure from a pawn attack.]
[The N is cetralized not in order to challenge for the c6 square, but to prepare White's next move and thus facilitate the development of the Bc1. Its small defects are that it blocks the d-file and leaves the important square e5 temporarily unguarded.]
[Despite the early exchange of Qs, the position should not be treated as an endgame. All the other pieces are still on the board and there aren't any pawns likely to become passed in the near future. Such a situation should be treated as a more or less balanced queenless middlegame, with empasis on the acquisition of small advantages.
For the time being, White's advantage appears negligible and therefore perfect play shoul lead to an uneventful draw.]
[Kramnik aims to control more space in the centre and at the same time to develop the Bc1 in a productive spot. After 13.Bf3 Bxf3 14.gxf3 Rc8 he would not be on time to take advantage of the relative weakness of c6.]
[Naturally, the function of castling in this position is not to provide a secure shelter for the King, but only to allow the Rh8 to participate in the battle. Kasparov would probably be much happier if he could keep his monarch closer to the centre for the endgame, but unfortunately his B does not have a better alternative than e7.]
[Now White has slightly better squares for his minor pieces. If Kasparov wishes to control equal space with equal material, he must resort at some point to the idea ...g7-g6 and ...e6-e5. Restricting himself purely to piece manouvres means he is likely to need some exchanges, especially a pair of Ns and possibly a pair of Rs.]
[The correct R, so that the King may return closer to the centre.]
[This is technically a novelty, one that Kramnik had most probably prepared (if only for another game) and Kasparov possibly anticipated. The game Bareev-Rublevsky, Elista 1996, saw 16.Kf2 Ne5 17.Nd2 Rc7 18.N4b3 Rc6 19.Rac1 Rac8 20.Rxc6 Rxc6 , with a position quite similar to K-K. It is worth noting the rest of the game, in order to make clear what Kramnik's true innovation is. 21.h3 Ke8 22.Bd4 Nfd7 23.Rc1 Rxc1 24.Nxc1 f6 25.a3 Bd6 26.b4 a5 27.Bc3 axb4 28.axb4 Ke7 29.Ndb3 g5 30.Nd4 Nc6 31.Nc2 h6 32.Nb3 Nce5 33.Nbd4 h5 34.Bb5 h4 35.Ne3 Nf8 36.Bf1 Nfd7 37.Ke1 Bb8 38.Bb5 Nf8 39.Nc4 Nxc4 40.Bxc4 Be5 41.Kd2 Bxd4 42.Bxd4 Nd7 43.Bb5 e5 44.Be3 Kd6 45.Bc4 Kc7 and ˝-˝.]
[At this point, both players had spent about half an hour of their available time, but now Kasparov thought for quite a while. This indicates that he was trying to figure out whether Kramnik was simply going to transpose to his second's game, by simply inversing the move order, or whether the opponent's specific choice has some indepentent significance.
Another serious possibility was 16...Rc7 , preparing the doubling of Rs on the c-file while at the same time keeping the b-pawn securely covered.]
[Apparently Kramnik has decided that moving the King is not such a productive way to use a tempo at this phase of the game and prefers to execute Bareev's manouvres one move earlier.
Furthermore, it soon becomes clear that Kramnik obtains virtually the same position with his King on g1 rather than f2, since Black plays the Rc8 to c6 in one move rather than two. This can mean either that he feels the King is safer on g1 or that he does not know the above mentioned game and is improvising along the way.
The future will bring out the truth in this matter, but for the time being we must restrict ourselves in examining the merits of each idea on its own right, regardless of Kramnik's consciousness.]
[Forced, but good. Blocking the B is not a true problem, because it had limited mobility anyway. It is much more important to activate the Rs, in order to force some exchanges.
A typical positional mistake would be the pseudo-active advance 17...b5 , which assumes control of c4 but at the same time relinguishes a5. After 18.Na5 the exchange of N for B is unavoidable, also forcing the R to temporarily abandon the c-file.]
[There is no point in keeping both Rs on the board. Meanwhile, attacking the R with 18.Nd4 would be meaningless, as after 18...Rc7 Black has gained a tempo and White doesn't have anything better than 19.N4b3 , leading to a repetition of the position ( 19...Rc6 ). Kramnik is obviously interested in more than a quick draw -Kasparov may get his half point, but he will have to work for it!;
An interesting alternative is 18.f4 , but it is understandable that the challenger does not wish to provoke an early crisis without first having established better coordination between his pieces.]
18...Rac8 19.Rxc6 Rxc6
[This is the true innovation in the game, not a move, but a plan well deep into the middlegame. Kramnik aims to control more space all over the board, avoiding further unnecessary exchanges.
Especially significant, compared to the Bareev-Rublevsky game, is that Kramnik refuses to exchange the second pair of Rs. Once he manages to push back some of the enemy forces (Ns in particular), he will want to possess as many pieces with different properties as possible. This will allow him to create threats of greater variety and put more pressure on Kasparov.
Last but not least, the advance of the kingside pawns demonstrates once again Kramnik's ability in this match to obtain positions with hidden potential, having first limited Kasparov's counterplay by denying him meaningful pawn breaks.]
[A good defensive idea, as exchanging the h-pawns leaves fewer targets for the enemy pieces.]
[Another useful precaution, providing extra support to the square d7, on which the Nf6 will inevitably land.
Counterattack with 21...Rc2 is rather premature: one R does not bring the summer! White then can continue simply 22.Bxb6 , not fearing 22...Rxb2 23.Bd4 and Black suffers serious material losses. One idea for the defender would be the intermediate move 22...Nc6 , forcing 23.Rb1 and after 23...Nd7 24.Bf2 (the h-pawn was "hanging"), in order to recapture the b-pawn. Unfortunately for Black, there is the continuation 24...Bf6 25.Kf1 (threatening 26.Nc4 and possibly 27.Bd3 or 27.Bd1) 25...Rxb2 (even worse is 25...Bxb2 26.Bd1 , but not 26.Bd3 Rc3) 26.Rxb2 Bxb2 27.Nc4 and the N will land on d6, forcing the win of the a-pawn.]
22.g5 hxg5 23.hxg5 Nfd7
[Reports say that both players had a little less than an hour of thinking time left until move 40. This should not be a problem in such a position, although tension is mounting steadily.]
[The idea of this redeployment is to open the d-file and threaten 26.Bxa6. 25.Nd4 would prevent the invasion of the enemy R, but would also limit White's potential. It seems Kramnik wants to entice his opponent into counterattacking at the wrong moment.]
[More than one commentator before me has pointed out that the World Champion feels uncomfortable when he has to defend passively. This risky attempt to activate a lonely R gets him into serious trouble, while it was still possible to play for a "zoning" defence with 25...Ke8 . Then Kramnik would lack any concrete way to create serious threats and Kasparov would be out of any immediate danger.
It is worth noting how consistent 25...Ke8 is with previous moves like 15...Kf8 and 21...Bc8. Petrosian is still the most difficult to imitate among World Champions!]
[A "wimpy" approach like 26.Rd2 makes no sense after 25.Nf3.]
[Now both Rs are active, but only Kramnik's has real hopes to combine its actions with other pieces. The idea f4-f5 should give Kasparov serious headaches, while the b-pawn seems much weaker than its counterpart on a2 (at some point White may win a crucial tempo by playing Rd7-a7).]
[Tremendous complications arise after 27...Re2 28.Bxb6 , when Black has the option of capturing either on e4 or f4. It is true that then his N escapes from g6 without personal damage, but the idea Rd7-a7 combined with Ra7-a8 will force ...Ba6-b5, when a double attack with a N from d4 is possible. Therefore, a likely continuation is 28...Rxe4 (as explained, it is not possible to play 28...Nxf4 due to the forcing variation 29.Ra7 Bb5 30.Nbd4 Rg2+ 31.Kh1 Be8 32.Bc7 Rg4 33.Nh2 Rh4 34.Ndf3 and Black loses material; , while after 28...Rxb2 29.Ra7 Bb5 30.Nfd4 (better than the immediate 30.Bc7, when Black will not continue 30...Rxb3 31.Ra8+ but 30...Bc6) 30...Be8 31.Bc7 White would also be on the driver's seat) 29.Bc5 Be2 (worse seems 29...Bb5 30.Bxe7+ Ke8 31.Ra7 Nxe7 -or 31...Rxf4 32.Nfd4- 32.Ra8+ Kd7 33.Nc5+) 30.Bxe7+ Ke8 31.Rd4 Bxf3 32.Rxe4 Bxe4 33.Bd6 and White should win the endgame, albeit with some difficulty.
I seriously doubt that both players calculated all the details of such lines here and in the next few moves, but of course they noticed most of them and chose the variations more likely to achieve their aims. The problem is that both must have used much more time than in previous stages of the game, leading inevitably to time trouble.]
[As indicated earlier, Kasparov does not have the time for 28...Rxa2 , because of the simple 29.Ra8+ .]
29.f5 exf5 30.exf5
[One does not really expect the World Champion to retreat 30...Nh8 . Once again Kasparov responds with counterattack, justifiably this time.]
[Practically forced. After 31.Bxb6 Ne5 32.Nfd4 (or 32.Nbd4 Nxf3+ 33.Nxf3 Bc6 with apparently sufficient counterplay) 32...Re1+ Black is out of the woods.
According to reports, at this point both players had fewer than 20 minutes left. The next few moves appear forced because of the multiple attacks at each other's pieces, so that any psychological tension should be a result of the situation at the board rather than the clock.] 31...Re1+ [After 31...Rxe3 32.Nxb5 the twin threats 33.fxg6 and 33.Ra8+ decide the issue.]
32.Kf2 Rf1+ 33.Kg2
[Of course, not 33.Kg3 Bd6+ and suddenly all of Kasparov's pieces participate in a real attack against the King!]
33...Nh4+ 34.Kh3 Rh1+ 35.Kg4 Be8
[The B has no choice on an open board, because of the side threat 36.Ra8+.]
[Although Kasparov has the 2Bs, Kramnik dominates the centre and has much better coordination of forces. In addition, the Nh4 is cornered and faces mortal danger.]
[Worse is 36.Ra8 because of the reply 36...Re1 when White must go for the variation 37.Nc2 Re2 38.Nbd4 Rh2 and nothing concrete is achieved.]
[No choice here either! It is quite impressive to see Kasparov's pieces so dispersed on opposite sides of the board. His 25...Rc2 has provoked a crisis that has turned out very badly for him.]
[Kramnik is over-subtle, when 37.Nf3 would entrap the N and create two new threats: 38.Kg3 and 38.Bxb6 (not only winning a pawn, but also gaining access to the important square c5 -White wishes to exchange one of the enemy Bs, in order to make his King safer for the future).
Kasparov would then have to take care first of the major threat by 37...Bd6 , but after 38.Bxb6 Kramnik would be perfectly justified to expect a full point by the end of the day.]
[37...Bd6 proves quite useless now after 38.Nb5 etc.]
[After such a quiet beginning, it is impressive that this game has so many critical points already. And there is more to come...
For the time being, Kasparov must take care of his N.]
[And he does it the wrong way. What is more impressive is the reports that the World Champion replied almost immediately, despite having almost 10 minutes available to reach move 40.
The correct retreat was 38...Nf4 , severely limiting the enamy King's movements and also recentralizing the galloping horse. If then White continues 39.Nd2 , Black can respond 39...Rd1 and solve all his problems.]
39.Ke2 Rh1 40.Nb5
[At the very last move of the first time control Kramnik creates the lethal threat 41.Nc7 and is very close to his second win in the match.]
[Also losing is 40...Nxf5 41.Nc7 Nd6 , because of the simple retort 42.Bg3 and Black's defensive wall crumbles. Kasparov elects the path of greatest resistance, accepting the loss of a piece and trying to make the best out of it. His basic idea is to eliminate as many enemy pawns as possible and destroy the coordination of Kramnik's pieces, at least temporarily.]
41.Nc7 Ke7 42.Nxe8
[42.Rxe8+ would be quite silly, since after 42...Kd7 43.Bxb6 Rh2+ and 44...Rxa2 the last white pawn must depart for greener pastures.]
[Once again, Kramnik has a dangerous passed a-pawn! (see Game 2) At least, compared to Game 3, the total number of passed Ps is reduced to half as many...]
[It is time for both sides to take stock and prepare for the next phase of the battle. Theoretically White is winning, but the situation has not been made clear yet. Kramnik's R and Ne8 are not conveniently placed, while Kasparov managed to create connected passed pawns, even if they are somewhat difficult to advance.
The challenger's goal is to transform the position to one with a clear-cut plan, probably through appropriate simplification. Meanwhile Kasparov will do anything to cloud the issue, from advancing his pawns to involving both Kings in tactical adventures.]
[The King goes west not in search for gold, but for salvation. One of Kasparov's secret hopes must be to exchange everything but his opponent's pawn and "wrong" B (one that cannot the queening square of the P).
Apparently Kramnik thought for quite some time before playing his next move, usually a smart approach after a crisis and when entering a novel stage of a game.]
[Speed is of the outmost importance in similar situations. A different idea is 44.Nd4 , but after 44...Nxd4+ 45.Bxd4 Rh8 46.Nf6+ (naturally not 46.Nc7 Rxa8 47.Nxa8 Bd8 etc.) 46...Bxf6 47.Rxh8 Bxd4 White's task becomes extremely difficult.]
[It makes a lot of sense to cut off the enemy King, but it was worth examining 44...Bf4 as well, in order to restrict the Ne8 and prepare the advance of the g-pawn. This idea could be used in the next few moves as well, but Kasparov obviously decided it was not worth it.]
45.Nc5+ Kc6 46.a5
[Tension is mounting and Kasparov added to it by thinking for almost 40 minutes here and leaving himself with just about 15 more to reach move 60. Was he looking deep in the position the whole time, or did he also wish to make Kramnik more nervous because of the approaching mutual zeitnot? In games like this one wishes to have been spectating live, so as to reach conclusions "in the flesch".]
[The big alternative is 46...Nd4+ , when an amusing variation is 47.Kd1 Rh1# ! Since however Kramnik's King can escape danger with relative ease by 47.Kf1 or 47.Kf2, Kasparov would then have to play ...Bg5-f4 at some point and exchange his B for the Ne8 when it comes to c7, in order to prevent further advance of the a-pawn. Unfortunately for the World Champion, in such a case ha may have to allow the capture Ne8xg7.] 47.Kd1 [Naturally the King runs away from the enemy forces.]
[Maybe this was the last time to play 47...Bf4 .]
[The basic reason my comments to this game are somewhat late is that Sunday was Election Day for the Greek Chess Federation. During the meeting I showed my brief notes to Master Antonis Bogdanos, having chosen to print a diagram at this very point.In a matter of minutes he suggested that 49.Nc7+ Kc4 and only now 50.Ne4 was much stronger, since the Black King is forced to steer away from the passed pawn. A sample variation is 50...Rxe4 51.Ne6+ Kb5 52.Nxg5 Rf4 53.Nxf7 Ne3+ 54.Bxe3 Rxf7 55.Rc5+ and White should have no trouble at all coordinating the remaining pieces. Without the Rs the game would be a theoretical draw, but in this case it constitutes an easy win.
Now events develop by force for thenext few moves.]
[49...Bf4 would be worse, since after 50.Rc5+ Ka6 51.Rxf5 Rxe4 there is a small but useful tactical detail: 52.Rxf7 and the rest is a matter of simple technique.]
50.Rc5+ Ka6 51.Nc7+ Kb7 52.Rxf5 Be3
[If the exchange of Bs is not offered, Kramnik shall play 53.Rb5 with the ustopable threat 54.a6+. Kasparov wisely tries to reduce the number of pieces that may support the enemy passed pawn.]
[Now things are not so clear after 53.Rb5 Bxb6 54.axb6 g5 , while; 53.Rxf7 Bxb6 54.Nd5+ Ka6 leads to a draw straight away.;
Another possibility with reasonable chances of success is 53.Nb5 , very similar to Kramnik's actual choice.]
[At this point both players must have approached serious time trouble.]
[It is a smart idea to entice the enemy pawn to advance as early as possible. The N will be tied up to its defence, while the R will have in turn to defend the N.]
[I wonder whether Kramnik saw that he could play 55.Nd5+ , because after 55...Ka6 56.Nb4+ the pawn is immune (56...Kxa5 57.Nc6+). Black would have to play 56...Kb5 and after 57.Rf4 there would soon follow a5-a6, when all of White's pieces defend each other in a much more harmonious way than in the game. The important practical difference is that in this schema Black can never capture the N (at the same time exchanging Rs), because then the King will be out of the "square" of the a-pawn.
Naturally, there would still be some technical difficulties, most involving the movements of both Kings, but the win would not be in doubt. Now, however, some strange developments occur.]
55...Kb6 56.Rxg7 Ra5
[A completely new position has appeared on the board, one that must have created headaches for the players. I expect both adversaries to have felt very tired by this time, but only one of them wishing adjournments had not been abolished...]
[In comparison to an earlier comment, it is clear White cannot play 57.Kc2 Rc5+ 58.Kb3 Rxc7 when the K+P ending is immediately drawn. Nevertheless, because of zugzwang Kasparov cannot keep the c-file barrier forever.]
[It seems to me Black should stand still and prevent the enemy King from crossing the 5th rank. Instead, Kasparov prepares himself for a series of distant checks.]
[This is probably not the line of most resistance. Kasparov could not hold both the a-file and the 5th rank, but now he offers his opponent a serious chance to end the game swiftly and convincingly.]
[But Kramnik does not want to take advantage of such an opportunity and thus Kasparov's perseverance under incredible pressure is rewarded. He could have played 59.Rg8 Ra1 (the N cannot be taken because of 60.a7) and now 60.Nd5+ (the last move of the second time control) Kc5 (or 60...Ka7 61.Nb4 and the N again supports the pawn from behind) 61.Rg5 when Black cannot avoid the appropriate restucturing of the enemy pieces (obviously 61...Rxa6 is unplayable because of 62.Nc7+). Meanwhile, an interesting side variation resulting favourably for Black is 60.Rc8 (instead of 60.Nd5+) 60...Ra2+ 61.Kb3 Ra5 62.Kb4 Ra1 63.Nd5+ Ka7 (of course, not 63...Kxa6 64.Ra8+ Kb7 65.Rxa1) 64.Kb5 Ra5+ (the surprise, because 64...Rxa6 fails to 65.Rc7+) 65.Kc4 (65.Kxa5 is stalemate!) 65...Rxa6 and the undesirable pawnless ending appears on the board.]
[Suddenly Kasparov's R finds itself at the right place at the right time for the first time in the game.]
[According to the match rules, at this point both players receive 30 extra minutes immediately after the 60th move. They also receive a 10 second increment from now on, but it is too little too late for Kramnik.]
[I had this endgame back in 1996 (my last really active year), but failed to win it against a player of near-master strength. Even though I knew Kasparov himself had won it earlier in the year against J.Polgar in Spain, my opponent found it quite easy to defend accurately. There are very few winning positions, most of which involve a badly cornered King and a defending R unable to check from a distance. If such a position does not arrive by force after simplification, the weaker side can avoid it without much trouble.
I suspect Kramnik continued more because he needed to adjust his thoughts to the fact that he had missed a great opportunity to assume an almost decisive lead, than in the hope that he may win the position.]
62.Kb4 Kb6 63.Nd5+ Ka6 64.Rg6+ Kb7 65.Kb5 Rc1 66.Rg2 Kc8 67.Rg7 Kd8 68.Nf6
[The defending King has been pushed to the edge of the board, but nothing more can be achieved. With just a little accuracy by Kasparov, all that Kramnik can achieve is rotation of the position around different sides of the board.
A most important factor is that the attacking King cannot establish fruitful coordination with both his other pieces.]
68...Rc7 69.Rg5 Rf7 70.Nd5 Kd7 71.Rg6 Rf1 72.Kc5 Rc1+ 73.Kd4 Rd1+ 74.Ke5
[This was a game that could make or break a World Champion. Kramnik managed to push Kasparov around, but failed to pass an important test for a title holder: accuracy in the endgame. Although the challenger should be fully credited with achieving a winning ending by subtle means in an unassuming position, Kasparov's resilience is at least equally admirable. On the other hand, it seems almost unthinkable that Karpov in his better days could ever fail to win if he got the position after 56.Rxg7 (or slide to just that from a position earlier in the game, for that matter).
Kramnik's reaction in the next game will be crucial. It is important that Game 5 is scheduled for the following day and therefore he does not have enough time to recover from the psychological shock of having missed such an opportunity. On the other hand, will the real Kasparov please stand up and enter the stage?]