Candidates, Quarterfinal Tiebreaks: Grischuk, Kramnik Advance (With Games)
The fans definitely got their money's worth today, as both the Grischuk-Aronian and Kramnik-Radjabov playoff matches were full of drama.
To recap, the tiebreaks had the following format: a four game mini-match with a time control of 25 minutes for the whole game, plus 10 second increments per move, followed, if necessary, by up to pairs of 5' + 3" blitz games followed by an Armaggedon game. Grischuk-Aronian was settled in the rapids, while Kramnik and Rajdabov needed two sets of blitz games to decide a winner.
Grischuk got off to a great start against Aronian, convincingly outplaying him on the black side of an English. The success didn't last for long, though. In game two his handling of the normally solid Exchange QGD was anything but, and soon Black was better thanks to his bishop pair. The nature of Black's advantage changed a few times, eventually transmogrifying into a pawn up rook ending that Aronian won. In game three having White finally made a positive difference and Aronian stood better. In fact, Aronian was probably winning at (at least) one moment (31.Qc3!), but Grischuk escaped and drew. Finally, Grischuk outplayed Aronian in game 4 on the white side of a Bf4 QGD. Aronian tried to create something of a fortress, but without any time on the clock it was inevitable that something would give, and it did.
With that, the rating favorites had gone three-for-three...in losing their matches - and it was nearly four-for-four! Kramnik and Radjabov had played four short, dull draws in classical chess, and they played four more forgettable games in the rapid. The first game had some play, and Kramnik pressed for a while with White in the Moscow Semi-Slav, but in games 2-4 the games were balanced, uneventful and quickly drawn.
It was time for the blitz rounds, and now it got exciting in a hurry. Radjabov had White in game 1 (game 9 overall), and produced a match novelty: 1.e4! Kramnik chose a Berlin, which Radjabov "declined" with 4.d3. The game turned into a sort of slow Giuoco, and surprisingly Radjabov handled the maneuvering phase better than 1...e5 professional Kramnik. Black was under heavy pressure, and after missing a nice chance to escape on move 37 (37...Ngf4! rather than 37...Ne7?), Radjabov hit hard with the power shot 38.Nh6! Black's position collapsed after that, forcing Kramnik to win the next game to keep the match going.
Kramnik did manage to obtain an edge in an ending with rooks and opposite-colored bishops, but although he managed to win a pawn Radjabov achieved a clearly drawn position. Desperate to keep the game going, Kramnik wound up blundering the extra pawn back, though it has to be said that it was the least important of his pawns. Anyway, Kramnik continued trying without getting anywhere, until of all things there was a clock malfunction. Play stopped for something like 10-15 minutes while everything got sorted out. When they came back, Kramnik at least managed to find an idea, and it worked. Radjabov fell into a fatal zugzwang on his 68th move, and the game was quickly over. Time for another two-game blitz mini-match.
In game 3 Kramnik again had White, and again Radjabov sharpened things up a bit by finally playing his favorite King's Indian. As usual in this match, Kramnik didn't get anything out of the opening, and the game was still equal once it reached a heavy piece ending with both sides having a queen and both rooks. It remained equal even as a rook ending, but Radjabov made the cardinal error of defending passively. Kramnik had enough positive ideas to make progress, and finally managed to win a second straight "drawn" endgame. (Carlsen would be proud.)
Finally, game 4 was another Declined Berlin-cum-Giuoco Piano, and while Radjabov might have briefly had an edge in the opening Kramnik was soon doing well and even stood better. 42...Qd6 was an error, though, and Radjabov briefly had some hopes. Kramnik defended well and found a fine idea on move 59 that wound up forcing a draw, but his implementation was inaccurate. Kramnik should have played 59...Rxd6 first rather than doing so after 59...Qe3+ 60.Kh1, because if Radjabov had chosen 60.Kf1! instead Black would have had some difficulties. After the exchange of errors, Black forced perpetual check and won the match.
Our semi-finals are set and will begin on Thursday. The pairings are Gelfand-Kamsky and Grischuk-Kramnik. I'll preview these matches tomorrow or Wednesday, and then we'll make a new set of predictions: a 2-G final, 2-K, or one of each?
As for today's games, I haven't annotated all 12 of them, but here they are, seven with notes.
Reader Comments (8)
I'm a Kramnik fan, but I think it'd be more exciting for an underdog to be the challenger. Although I think I'd like to see a Kamsky-Kramnik match over a G-G match. I'm boring!
Shucks, of those advancing, I only had Kramnik right :-(
I'm going with Kamsky-Kramnik as the final round, which I think Kramnik will win. I'm still rooting for Kamsky though.
Two Gs? Two Ks? Nah, one of each is the safe bet: 50% probability that it's GK and only 25% for either KK or GG.
DM it would be great to hear your views on whether the 4 game match is too short to determine with any degree of fairness the stronger chess super GM or do you need 6 or 8 or 12 games for the contest to be meaningful. It seems that clearly 4 games is too little - anything less than 8 games is too little?
[DM: It's a job for the statisticians. I remember reading (actually listening) to a book on probabilities, and while I don't remember the exact figures the gist was that it would take a very large number of games to more or less guarantee that the better player (better team, in the book's context) would win. It's just a question of how much randomness we're willing to tolerate. Anyway, the statisticians are welcome to chime in to let us know how often, and with what confidence level, an n-point rating favorite will win in a 4-game match vs. an 8- or whatever-game match. Clearly there's a high degree of randomness in a 4-game match, and it would be more fun for the fans to see the matches go on a while longer. On the other hand, if all the matches were long the event would either turn into a marathon or one would need additional sponsors, and that's not necessarily feasible these days.]
Technically, if order matters it's 25% GK 25% KG.
Interestingly, there is a Sanghinagar 94 connection with all the players (except Grishchuk). Kamsky beat Anand there, and Gelfand beat Kramnik!
I thin 4 games is clearly too short, but both players know what they have to do. It certainly makes the matches tense, if nothing else. Ideally I'd love to see 12 games. That would make one loss not so catastrophic.
Right, but it's not randomness exactly. There are sporting considerations, e.g. how do you approach a 4-game match situation in terms of opening, psychology, etc., especially given the tie-breaking rapid and blitz rounds. Under these conditions the ELO favorites are not necessarily such favorites. Lower-rated players can snipe at opponents with clever preparation and/or the threat of injecting calculated risk into new positions (esp. early-ish, when the opponent may see ghosts in suspecting home-preparation in complex positions and thereby choose inferior continuations).
I don't know. I just enjoy watching the great players maneuver under pressure (and guessing at the meta-motives behind various moves). Long matches are usually too safe. Seems like White always wants to pose problems without taking much risk. But here you've got to make your stand relatively quickly, whether it's in classical or rapid. It's not exactly 'fair', but for once it's good sport and terrific for spectators.
If chess wants to be anything like a sport, this pressure ought to be welcomed. Maybe one day we'll have a football-style league table w/ tournaments on the side and less obsession with the months-long preparatory mining of opening systems and...just dreaming.
Also, kudos to DM for all of the annotations. There aren't always a ton of comments attached, but those of us who are reading are certainly appreciative!