The point is that if 10…Bxb4 11.Ncxd5 cxd5 12.Nxd5 seems to give White very dangerous, perhaps winning compensation.
 
For example:

1) 12...Bc5 13.Bf4 and now both 13…Kf8 and 13…Nc6 can be met by 14.Qf3.
 
2) 12...Ba5 13.Qe2+! and now 13...Be6 14.Nf4 and 13...Kf8 14.Ba3+ are instant losers for Black, while 13... Kd8 is met by 14.Bg5+ Nf6 15.Nxf6 gxf6 16.Qe5. That leaves 13...Qe6, but then 14.Qh5+ g6 15.Qg5 is curtains.
 
I imagine Black should decline the gambit, e.g. 10...Bd6, but we can then tickle him with 11.b5. I’ve spent less time on this, and attempted to make 11…Nf6 12.bxc6 bxc6 13.Nexd5!? work, though I think it comes up a little  short. Nonetheless, I feel 10.b4 is at least an interesting, and possibly promising idea, and knowing your hatred of the Latvian gambit, I thought I’d send it to you and see what you think.

Jeremy Silman Replies:

Steve, your idea of 1.e4 e5 2.Nf3 f5 3.Nxe5 Qf6 4.Nc4 fxe4 5.Nc3 Qf7 6.Ne3 c6 7.d3 exd3 3 8.Bxd3 d5 9.OO Bc5 10.b4!? is quite an interesting one: 10…Bxb4 11.Ncxd5 cxd5 12.Nxd5

CRUSHING ATTACK

12…Ba5 13.Qe2+ (13.Re1+!?) and Black will probably be torn limb from limb (by the way, in your line 13.Qe2+ Kd8 14.Bg5+ Nf6, I think 15.Bc4! is much stronger than 15.Nxf6).

I also agree that 12…Bc5 13.Bf4 Kf8 14.Qf3 gives White a very dangerous attack (for example, 14…Nc6 15.Nc7 Rb8 16.Bc4 Qf5 17.Rfe1 Be7 18.Rad1). Thus it seems to me that you are right about the necessity of Black avoiding the gambit and trying 10…Bd6.

KEY POSITION

After 11.b5 (of course, White has many other possibilities here) 11…Nf6 (11…Ne7 12.bxc6 bxc6 13.Nexd5! is strong: 13…Nxd5 14.Re1+ Kf8 15.Bc4 or 13…cxd5 14.Nb5 Qf6 15.Bb2 Qxb2 16.Nxd6+ Kd7 17.Nxc8) 12.bxc6 bxc6 13.Nexd5 cxd5! (13…Nxd5 14.Re1+ gives White more than enough for the sacrificed piece) I can’t find a satisfactory continuation of the attack (though I don’t have the time to do the position justice – perhaps a reader will find something?). If the piece sacrifice (13.Nexd5) doesn’t work, then it doesn’t seem to me that White’s other choices (after 10…Bd6) lead to the punishment Black deserves for daring to play the Latvian Gambit. Thus 10.b4!?, though fun, is most likely insufficient.

Our site correspondence expert’s following comments are worth checking out – before reading what he wrote, I was beginning to think that all postal players had lost touch with reality.

Manuel Gerardo Monasterio Replies:

I was most amused by Mr. James’ phrase referring to Mr. Silman’s “hatred of the Latvian gambit.” The Latvian does not deserve to be hated by anyone, it can be simply described nowadays, and with total confidence, as absolute Bovine Excreta. From a theoretical point of view, it is a refuted opening variation; that does not mean that some people keep gaining points with it in practical play. But the fact that there are people still smoking and yet alive, does not mean that smoking is good for anybody’s health!

You will not find a single top modern correspondence grandmaster with the Latvian in his repertoire. As far as I am concerned, John Nunn’s analysis, which I am giving below, is the relatively final word for this triviality.

1.e4 e5 2.Nf3 f5 3.Nxe5 Qf6 4.Nc4 fxe4 5.Nc3 Qf7 6.Ne3 c6 7.Nxe4 d5 8.Ng5 Qf6 9.Nf3 Be6 (9...Bd6 10.d4 Ne7 11.c4! 0-0 12.Qb3 Kh8 13.Bd2!? (13.Be2!, ±) 13...dxc4 14.Bxc4 Nd7 15.0-0) 10.d4

NOT ENOUGH COMP FOR THE PAWN

10…Nd7 11.Bd3 0-0-0 12.0-0! (Already clearly better for White) 12...g5 13.c4 Bd6 (13...g4 14.cxd5 cxd5 15.Ne5; 13...dxc4 14.Nxc4 g4 15.Nfe5 Bxc4 16.Nxd7 Rxd7 17.Bxc4 Rxd4 18.Qb3 Ne7 19.Be3 and White wins) 14.cxd5!! cxd5 15.Nf5 Bxf5 (15...Bf4 16.Bxf4 gxf4 17.Qc1+ Kb8 18.Qxf4+ Ka8 19.Qc7 is winning for White) 16.Bxg5 Bxh2+ 17.Kh1 Qf7 18.Rc1+ Kb8 19.Bxd8 Bxd3 20.Qxd3 Bf4 21.Rc3, winning. This is mostly Nunn’s analysis.

I can supply several games confirming this analysis and exemplifying the miserable fate of Black in this variation.

JS: Please check out Mr. Monasterio’s new series, GOING POSTAL, and see just how tricky opening theory in correspondence chess can be.