Stefan Buecker wrote on 07/26/11 at 19:09:55:


Stefan, with the greatest respect, could we finish up on 3...a5? We need to take view on it and move on, either that or put it aside for later. Since you are the putative exert on the subject, I suggest you analyze it and come back when you have a pgn showing White's best play against it.

I feel safe in saying that many people would be happy to put aside work on 3...a5 at this point in the project, if no ready demonstration of += is available.

Markovich wrote on 07/26/11 at 11:31:31:

I had looked at 13.c3, 13.Qc4, 13.Rb1, in each case it is fair to write +/-. However, my line has the advantage to exchange queens, reducing Black's chances almost to zero. Looks more pragmatic to me, but of course 13.c3 isn't bad. 

Göran wrote on 07/26/11 at 15:44:33:

Which variation do you mean? Is this (1.e4 e5 2.Nf3 Nc6 3.Bb5 a5 4.d4 exd4 5.Nxd4 Nxd4 6.Qxd4) c6 7.Be2 d5 8.exd5? Looks positionally dubious for Black. I'd rather not take on d4 with the knight. 

fling wrote on 07/26/11 at 11:34:33:

That's why I gave 8...d6 as better. And in the meantime I've found another idea after 5...Bc5 (see below). 

Markovich wrote on 07/26/11 at 12:48:17:

In my files I had in fact the move-order 5...Qf6 6.Nf5 Bc5, while 5...Bc5 6.Nf5 Qf6 only seemed to transpose. The move-order 5...Qf6 6.Nf5 could become "independent" after 6...g6 7.Ne3 Nge7 8.Nc3 Bg7, but this seems worse: 9.Ned5 Nxd5 10.Nxd5 Qd8 11.c3 0-0 12.h4!? and the fianchetto becomes a target for White. It makes more sense to combine a5 with Bc5. 

However, after 5...Bc5! 6.Nf5 Black has an attractive additional idea, overlooked in my first analysis (because of my weaker move-order): 6...d5! 7.Nxg7+ (7.exd5 Bxf5 8.Qe2+ Qe7 9.dxc6 Qxe1+ 10.Kxe2 0-0-0 comp., or 7.Qxd5 Qxd5 8.exd5 Bxf5 9.dxc6 b6 comp.) 7...Kf8, for example:


(a) 8.Nf5 Bxf5 9.exf5 Nd4! and I like Black's attacking chances (so "adding" a5/Bb5 to the well-known Scotch line Nf5 can even be good for Black: Nd4 attacks the Bb5). 

(b) 8.Nh5 Qh4 9.Ng3 Nf6 10.Nc3 Ng4 11.Rf1 Nd4 12.Be3 Nxe3 13.fxe3 Qxh2 14.Nge2 Nxb5 15.Qxd5 Qh4+ 16.Kd2 Qe7 17.Nxb5 Be6 with complications. 

1.e4 e5 2.Nf3 Nc6 3.Bb5 a5 4.d4 exd4 5.Nxd4 Qf6 may be worthy of consideration.

Further after 3...f5 4.Nc3 Nf6 5.exf5, we should include consideration of 5...e4, a move not so simple to deal with otb. Sokolov argues persuasively that 6.Ng5 d5 (6...Nd4 7.d3) 7.d3 is the way to go. Quoting him, 7...h6 (7...Bf5 8.dxe4 dxe4 9.Qe2 Bd6 10. Bxc6 bxc6 11.Qc4 Kd7 (11...Qd7 12.Ngxe4) 12.Be3 Qg8 !3.Qa4 +/- Przepiorka - Tarrasch, Nuremberg 1906) 8.Ne6 Bxe6 9.fxe6 Qd6 (9...Bb4 10.dxe4 dxe4 11.O-O Bxc3 12.bxc3 Qe7 13.Qe7 +/-) 10. dxe4 Qxe6 11.O-O dxe4 12.Nxe4! and he goes on, but it is perhaps sufficient to stop here.

Versus 3...Bc5, I would advocate 4.c3.

Stefan Buecker wrote on 07/26/11 at 11:04:06:


Very interesting rule. However, as you say, I guess the emphasis has to be on "a bit", since not every file is useful.

Anyway, just without analysing the move, I think 4.d4 looks really critical from a "classical" perspective - i.e. meeting a flank move with a counter in the center. After all, White's original idea in the Spanish has to be considered increased control over d4 and e5, no?

Btw, back to the position. I would for sure pick White in this position.

Jupp53 wrote on 07/26/11 at 10:02:01:

If there only were a possibility to ask Tarrasch what he thinks about your main line, after 12...bxc6. He once wrote that he doesn't really care about the pawn formation, what really mattered are the pieces. 


But a while ago I discovered a rule of my own: when you have more open files than rooks, the charm of open files fades a bit. So I don't know. It might be +=, or not. Anyway, I'd prefer to play 8...d6 (instead of your 8...0-0), e.g. 9.Nxe7 Qxe7 10.Nd5 Qd8, about +0.30. 

There is even a third line: 4.c3 f5!, when White's Nb1 can no longer jump to its best square c3.

fling wrote on 07/26/11 at 09:28:10:


Agree to that. 

@ 3. - a5

Maybe there's a second line to explore:

1. e4 e5 2. Nf3 Nc6 3. Bb5 a5 4. d4 exd4 5. Nxd4
a)5. .. Nf6  6. O-O Bc5 7. Be3 Bb6 8. Nc3 O-O 9. h3 Re8 10. Re1 and Rybka 4 offers .41
b) 5. .. Bc5 6. Nf5 Qf6 7. Nc3 Nge7 8. O-O O-O 9. Be3 d6 10. Nxe7+ Qxe7 11. Bxc5 dxc5 and though Houdini offers only .26 after 12. Bxc6 I'd prefer this with white more than an engine would do.