Fortunatly Black can also try the complicated line 1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 g6 6.Be3 Bg7 7.f3 0-0 8.Qd2 Nc6 9.0-0-0 d5 10.Kb1 Rb8!?

The idea is for White to try and swap a pair of rooks, or else stop Black dominating the d-file; if he manages this then Black's winning chances decrease dramatically. Eddie Dearing talks about it a bit in his book; in particular, White doesn't have to play 20. b3 as 20. c3 is available to reduce the scope of the dark-squared bishop. One way I play against it is to play (I think it's Golubev's idea) 17...Rac8, avoiding the d-file entirely and putting pressure on the c-pawn; in particular, 17...Rac8 18. Qc1?! Bf5! is strong, and after 18. g4 (to prevent ...Bf5) then 18...Rfd8 19. Qc1 Rd6 20. Rd1 Rb6 is fine for Black as 21. c3? just drops a pawn - as such 21. b3 is forced, when Black is doing well: he has an attack (...a5-a4 is coming) and the White bishop is a poor piece on e2.

I think all these lines are playing around with a position that's virtually equal; all of them just mess around with lines that very slightly reduce or increase winning chances for one side or the other. Here, I think 17...Rac8 solves most of the problems for Black.

Adam

The line in question is:

1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 g6 6.Be3 Bg7 7.f3 0-0 8.Qd2 Nc6 9.0-0-0 d5 10.Kb1 Nxd4 11.e5 Nf5 12.exf6 Bxf6 13.Nxd5 Qxd5 14.Qxd5 Nxe3 15.Qd3 Nxd1 16.Qxd1 Be6 17.Be2

I haven't seen the analysis (and if you could give the complete game score that would be great, since I can't seem to find it online), but I don't immediately see why this move would be such a big improvement. 17...Rfd8 18.Qc1 Rd6 19.Rd1 Rb6 20.b3 Rc8 and I don't really see what White has accomplished with Be2 especially.