The Queen sac looks good indeed. But I think Black can avoid it without too much damage:
10...h6 11.Bxf6 gxf6 12.Na4 Bb6 followed by Qe7 and castling queenside;
9...Na5 10.Bd3 0-0 is also reasonable.

GMTonyKosten wrote on 04/08/08 at 10:16:29:


After 1.e4 e5 2.Nf3 Nc6 3.Bb5 f5 4.d3 fxe4 5.dxe4 Nf6 6.0-0 Bc5 7.Qd3 d6 8.Qc4 Qe7 9.Nc3 Bd7 10.Nd5 Nxd5 11.exd5 Nd4 12.Bxd7+ Qxd7 13.Nxe5 Qf5 14.Nd3 0-0-0 15.Kh1 (Kosten), 15...b5 16.Qc3 Qxd5 should be equal according to Harika Dronavalli, commenting on Mamedyarov-Radjabov, Baku 2008 in NICYB 98 (that game saw 15.a4 ending in a draw).

GMTonyKosten wrote on 04/07/08 at 13:32:27:


15. Kh1 certainly looks like a good try for a white advantage.

Black might try (after 1.e4 e5 2.Nf3 Nc6 3.Bb5 f5 4.d3 fe4: 5.de4: Nf6 6.0-0 Bc5 7.Qd3 d6 8.Qc4 Qe7 9.Nc3 Bd7 10.Nd5 Nd5: 11.ed5: Nd4 12.Bd7: Qd7: 13. Nxe5 Qf5 14. Nd3 0-0-0 15. Kh1)
15....Rhe8 with the idea of swinging the rook along the fourth rank and  starting and attack against the white king, profiting from the poor position of the white Queen on c4 an the lack of defensive pieces on the white king side. According to my analysis black gets a lot of activity. I am not sure, however, if its enough compensation for the pawn deficit. My analysis runs as follows:

1.   15....Rhe8 16. Be3 Re4 17. Nxc5 (what else?) dxc5 18. Qxc5 Rxd5 and blacks should have enough for the pawn.

2.   15....Rhe8 16. b4 Bb6 17. Bb2 Re4 (another funny idea is 17....Nf3!? 18. Qf4 (only move) Qxf4 19. Nxf4 Nd2 and maybe Black is   near equality, for example 20. Rfe1 Ne4!))18. f3 Rh4 (threatening Rxh2) 19. g3 Rg4 and Black has got a very strong attack, for example 20. Nf4 Rxf4 21. gxf4 Nxc2

Maybe there are improvements for white, but so far I didnt come across them. Do you see anything?

So far I still believe, the 4. Nc3 main line is critical instead of 4.d3. I really hope we will see some radjabov games there in the near future...

Funnily enough I've been wondering the same thing; during the Radjabov games on ICC I tried to ask this question in the channel and couldn't get a satisfactory answer. I was under the impression that 13...Qf5 offered inadequate compensation, though a quick run past Rybka gives its main line as 14.Nd3 O-O-O 15.Nxc5 dxc5 16.Qxc5 Nxc2 17.Bf4 Qxf4 18.Qxc2 Rxd5 = - maybe this is what the top GMs have found that the old books had missed? Neither 14.b4 or 14.Be3 seem to offer white anything, either.

I think it is time, somebody writes a fresh full book book on this.

ArKheiN wrote on 03/08/08 at 12:08:54:



Well, but that is actually the way Radjabov handled the position. 16....0-0-0 and then Rdf8 instead of your proposed 17....Rhf8.

Its funny to see that Radjabov improved today in his Linares game against Carlsen by employing the idea of a5!, as mentioned above. Carlsen could not achieve more than a drawn ending, despite being a pawn ahead.
The endgame with 4 against 3 pawns on the same side is similar to piket-kasparov, kasparov chess grandprix 2002, where piket won a very instructive game. In that game however, the rook pawn was on a2 (well, actually h2, but you get what I mean). With the pawn on a4 as in Carlsen-Radjabov white simply has no plan to improve his position because of the b3 weakness...