One third of the way through the Candidates finals, I'm already feeling slightly smug about my predicted qualifiers for Mexico. Yes, well, I went with the rating list for Leko, Aronian and Grischuk and opted for previous matchplay success for Kamsky, so I didn't need to be a genius, I suppose. There is still a long way to go, though only one of the eight finalists (Shirov) recovered from being down in the semi-finals to qualify for this stage.
I've not watched any of the final eliminator matches live so far, so feel less qualified to comment on what has gone before. Watching in real-time gives you a strong feel for the ebb and flow of the game, helped by the chatter of the expert commentators and even by the raucous babble of the audience.
The player I feel most sorry for so far is Shirov, who has made the running in both of the games of his match with Aronian but has only half a point to show for it. Maybe he feels he has to make a big impression early before Aronian adapts himself to his style. We've certainly seen some fire on board (will he ever live down this hubristic book title?) but perhaps he is being too 'interesting. for his own good.
Leko pressed hard with White though I thought his 28 Qh2 idea looked weird. But it was only when I consulted Fritz that I found it should have lost outright after the amazing 28...Ne4! when 29 Rxe4 runs into 29...Rxf2!! and White is being done up like a kipper. You need a silicon brain to work it all out, of course. Today the two guys managed a miserly 15 moves. Bareev surely needed to do more with the white pieces.
Grischuk's first-game win against Rublevsky looked pretty wonderful to me, though Mihail Marin on ChessBase seemed very low-key about the whole thing. Grischuk advanced upon his opponent in bold and brassy fashion, laced his attacking play with a couple of well-timed sacrifices and produced a very attractive finish. What more could you ask? If Kasparov had played this game, a hundred chess hacks would already be falling over each other in pursuit of the most effusive superlatives to describe it. Perhaps Rublevsky missed a couple of better defences at critical junctures, but Grischuk would still have had an extra pawn and a dominating position. I'm glad to see that Max Dlugy describes it simply as "a beautiful game" on chessgames.com. Their second game ended in 18 moves. Once again, this is no discredit to Black but Rublevsky as White will only get two more Whites to even the score.
Boris Gelfand has drawn all eight standardplay games he has been involved in since arriving in Elista, but not for the want of trying. Well, in game two, anyway. Game one was a disappointment, with Kamsky failing to make much of an impression with White. The second game seemed worse for Kamsky, with Gelfand building up an impressive position and menacing Kamsky's position with a knight which shuttled between e5 and g5 and back again. The US player grovelled and it looked a bit grim for a while. But then Gelfand allowed Black to break out of the bind. There was a little tactical flurry towards the end but it panned out to a draw. Kamsky is a very effective groveller, something I used to take pride in myself in my younger days. I shall be watching his games very closely from hereon in.
I understand there is also some sort of chess match going on between two bits of software in Elista. Anyone interested? No, me neither. But I did get interested when I read about the Rybka people throwing down the gauntlet to FIDE to get a shot at Fritz or Junior... here is their press release
Rybka $100,000 challenge to FIDE
Dear Mr. Kirsan Ilyumzhinov & members of FIDE,
first, let me start by commending you for your entry into computer chess with the organization of the ‘Ultimate Computer Chess Challenge 2007’ . Computer chess has seen dramatic improvements in the past few years. Some chess engines have progressed dramatically from the primitive beancounters of yesterday and I believe that our games too now qualify as art. Chess at this level inevitably attracts the attention of chess players all over the world.
Unfortunately, the lack of an open, formal qualification procedure for your event was disappointing, and your choice of the two opponents was downright bizarre. You have snubbed my program, Rybka, which leads every single computer chess rating list by a considerable margin at all time controls from blitz games to long tournament games . In many cases the gap between Rybka and her nearest competitor is well over 100 Elo. None of this is anything particularly new - Rybka was released on December 4, 2005, and since then her smallest lead ever in any major rating list at any time control and on any hardware was 60 Elo. In addition to this, she has competed in all eight major international tournaments held since her first release and taken clear (unshared) first in seven of them.  Rybka has also displayed her superiority in competitions against human players. It's no wonder that Rybka is generally considered "the undisputed strongest chess program in the world."
Some of the other aspects of the match also raise questions. Chessbase exclusively markets three of the world's top ten engines, so it's a curious coincidence that two of them will participate. Also curious is the involvement of the ICGA - after all, their own self-titled “World Computer Chess Championship” is being held on overlapping dates. This type of apparent division between insiders and outsiders runs counter to all principles of sport and fairness, and I call on you to uphold democratic FIDE norms in the organization of such events.
In the spirit of open competition, I am formally offering a $100,000 computer chess challenge from Rybka to FIDE, who will be represented by the winner of the Ultimate Computer Chess Challenge 2007. My challenge consists of a 24 game match, at classical time controls, on unlimited hardware and with unlimited opening books, held at 2 games per day over twelve days, with Rybka giving a handicap of one point plus draw odds and thus requiring a score of 13/24 or better to win the match. The prize fund of $100,000 should be a winner-takes-all, loser-pays-all proposition. The remaining details can be worked out in private.
As the Ultimate Computer Chess Challenge 2007 takes place during the Candidates Matches in Elista, it is appropriate that the winner’s match vs Rybka be played in Mexico between September 12 and October 1, 2007, during the FIDE World Chess Championship.
Gens una Sumus,
author of Rybka
FIDE International Master
So there we are. A piece of chess software that is so strong it has started arguing with its rivals. The sign of a true champion...