Warning, I noticed a hole in one of the recommantions from Barsky. Because of that, a CC opponent supposed to be strong lost to me

:
1.e4 e5 2.Nf3 Nc6 3.d4 exd4 4.Nxd4 Bc5 5.Nxc6 Qf6 6.Qf3!? (I bought the book myself mostly because I like the ideas under that move)
6..dxc6 (I think this move is equal with best play but 6..bxc6 is probably equal too but more umbalanced, so this is maybe a better winning try)
7.Bc4 Be6 (Barsky seems to prefer 7..Ne7 to Black but I have not studied that move, he says that Black will have a very long and laborious passive defense after Be6, and I quite agree but not because of his variation) Now he gives
8.Bxe6 Qxe6 9.Be3 Bb4+ 10.c3 Bd6 11.Nd2 0-0-0 12.0-0 Nf6 13.Bg5 "White has an excellent development and begins an operation for destroying his opponent's pawn-structure". I agree that White's moves seems to be very natural and at first sight we might just agree with him. His reference's game is Sulskis (2537) vs Svetushkin (2618) and this is the only game at master level. Black played 13..Kb8 and White managed to convert a slight advantage into a win.
Ok and here is the hole!:
13..Ng4! a nice tactical move with a forced sequence that end with an advantage to Black.
Here is my CC game:
[Corr.]
[Date "2010.04.16"]
[White "Timothy,Chan Wei Xuan"]
[Black "ArKheiN"]
[Result "0-1"]
[WhiteElo "2341"]
[BlackElo "2211"]
1.e4 e5 2.Nf3 Nc6 3.d4 exd4 4.Nxd4 Bc5 5.Nxc6 Qf6 6.Qf3 dxc6 7.Bc4 Be6 8.Bxe6 Qxe6 9.Be3 Bb4+ 10.c3 Bd6 11.Nd2 Nf6 12.O-O O-O-O 13.Bg5 Ng4 14.Bxd8 Nxh2 15.Qe3 Nxf1 16.Kxf1 Rxd8 17.Qxa7 b6 18.Re1 Qh6 19.Nf3 Qh1+ 20.Ke2 Qxg2 21.Rg1 Qh3 22.Rxg7 Qe6 23.Qa8+ Kd7 24.Qa4 h6 25.Rg1 b5 26.Qc2 Re8 27.Nd2 Kc8 28.c4 Bf4 0-1
He lost on time in a clearly worse position.
I don't know where my opponent got his rating (he began in this website with this rating while I began with 1700) but it seems he has followed blindly the game, or the book. I believe most line with 6.Qf3 dxc6 are nice for White in corr game against a computer alone because of the long-terms advantages and the almost endings territory. But in that precise variation, the computer was useful for Black and would have been useful to White.
I would call 8.Bxe6 a "bad" move.
More in the spirit of the opening is 8.Qxf6! Nxf6 9.Bxe6 fxe6 10.f3 with the desired queenless game with a structural advantage without the bishop pair for Black. Practice seems to confirm my thoughts, not much games but nice scores for White. Even GM Ni Hua (2704) has tried that line and won. I got myself that position as Black in another CC game and I think
10..e5 ! followed by a g5 is a good way to bind the kingside and it is the way I have played. The game was drawn and I used more my personnal feeling than computer here. I even think it might be dangerous to blindly play the endings in CC play with computer only. Here is the game:
[Corr.]
[Date "2009.12.31"]
[White "Nita,Catalin"]
[Black "ArKheiN"]
[Result "1/2-1/2"]
[WhiteElo "2093"]
[BlackElo "2023"]
1.e4 e5 2.Nf3 Nc6 3.d4 exd4 4.Nxd4 Bc5 5.Nxc6 Qf6 6.Qf3 dxc6 7.Bc4 Be6 8.Qxf6 Nxf6 9.Bxe6 fxe6 10.f3 e5 11.Nd2 Ke7 12.Nb3 Bb6 13.Ke2 a5 14.a4 Rhg8 15.Bd2 g5 16.g4 h5 17.h3 h4 18.Rhd1 Ke6 19.Be1 Nd7 20.Rd3 Nc5 21.Nxc5+ Bxc5 22.Rad1 b6 23.Rd7 Bd6 24.Rh7 Rh8 25.Rxh8 Rxh8 26.Bd2 Be7 27.Be3 Rd8 28.Rxd8 Bxd8 29.Kd3 c5 30.Kc4 c6 31.b3 Kd7 32.Bd2 Be7 33.Bc1 Ke6 34.Bd2 Kd7 35.Kd3 Ke6 36.Kc4 Kd7 1/2-1/2
It seems that the best strategy for Black in the bishop ending is to put pawns in the same color as the bishop, an opposition of the traditional rule, but here it's a form of fortress. I was a bit worried of the ideas with c3-b4 near the end and he could have try harder. It's ok to play something like this as Black in a CC game if you are happy with a draw. If not, that White idea gives headhaches. In an OTB game I would really not like to play the Black side of it. Technical defense is required even against a lower player just to end with a draw.
I hope my post will be useful to someone