GM Joel on Bb7 in the Marshall
By GM Joel Benjamin   
September 9, 2008
Hi Joel,
    Is there anything concretely wrong with 11…Bb7 in the Marshall Attack of the Ruy Lopez, instead of the overwhelmingly more popular 11...c6? I have done some analysis myself and Black seems perfectly fine. I even put it into Deep Fritz 10 overnight and it even likes Bb7 better! I understand that it is still pretty early in the game for an accurate evaluation from an engine, but its opinion shouldn't be overlooked after an all-nighter either.

Position after 11...Bb7

In my database there are only a handful instances of 11...Bb7...compared to more than 300 with c6. Why is it so rarely played? The only minus I really see in Bb7 is after a move like 12.d4 Black can't play the normal 12...Bd6 due to 13.Bxd5 Bxd5 14.Rxd5 Bxh2+ 15.Kxh2 Qxd5 and white has two pieces for the rook. White is way underdeveloped but I don't think this should be too hard to overcome.

Position after 15...Qd5

Therefore 12...Bf6 is good but it blocks the queen's normal path to h4. If the queen could get there like in a normal Marshall white would have to be VERY careful playing g3 like he normally does, as the Bb7 is a monster! Maybe black doesn't like this line because Bf6 ruins that entire idea, at least for quite a while. Maybe 12...Nf6 is better because if something like 13.Bg5 Bd6 14.Nd2 (13.Bxf6 Qxf6 14.Rh5 g6 15.Rh3 (15.Qg4?! Rae8) c5 and Black has more than full compensation) 14...Bxe5 15.dxe5 Qd3! 16.exf6 Qg6! and this is probably about equal. Therefore white should try 14.Re1 instead of Nd2 then 14...c5 15.Bc2 h6 16.Bh4 Rc8 with compensation. White's undeveloped queenside could likely cause some problems. None of this is exactly "forced" but it does a great job of showing some of the possibilities.

This is of course some of what may happen after 12.d4. I know there are other moves too but nothing seemed to favor White that much. Does c6 just offer more chances than Bb7? Has a refutation been found to Bb7? Or is developing theory to support this move just something that nobody is really willing to do yet?
Eric J


Your question turned out to be much more complicated than I expected.  I had always thought that 11…Bb7 (a more attractive move than 11…c6) could be dismissed because 12.Qf3 forces the diagonal closed.  It turns out matters are not so simple.  While 11…c6 appears in my database over 2000 times more than 11…Bb7 (get a larger database, my friend), the 100+ examples of 11…Bb7 include quite a few big names.

After 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.0–0 Be7 6.Re1 b5 7.Bb3 0–0 8.c3 d5 9.exd5 Nxd5 10.Nxe5 Nxe5 11.Rxe5 Bb7, 12.Qf3 does not even force 12…c6.

Position after 12.Qf3

 In fact, Black has played 12…Bd6!? in many games.  Now on 13.Rxd5?!:

A)    13...Re8 appears five times, all wins for Black. 
A1) 14.Qxf7+ Kxf7 (14...Kh8 is untested, but looks strong) 15.Rxd6+ Kf8 16.Rxd8 Raxd8 17.Kf1 Be4 18.f4 Bd3+ 19.Kf2 Re2+ 20.Kf3 Re1 21.Na3 Rf1+ 22.Kg3 Rd6 23.Bc2 (after 23.h3 Rg6+ 24.Kh2 Be4 25.g4 Rh6 the computer acknowledges mate is coming soon after 26…Rh1+) and Black can and must take a perpetual.
A2) But I haven’t cracked 14.Kf1: 14...Qe7 15.Qd1 Qh4 16.Rh5 Qe4 17.f3 Qg6 18.g4 Qd3+ 19.Kf2 g6 20.Bc2 Re2+ 21.Qxe2 Qxc2 22.Na3 Bxa3 23.Re5 leaves White well on top.
B)    13…Qe7! looks more reliable.  Now White faces major back-rank issues.  14.Kf1  (14.Qe3 Rae8 leaves White without a good move) 14…Rae8 15.Qd1 Bxd5 16.Bxd5 Qe5 17.Bf3 Qxh2 18.g3 Re6 gives Black a strong attack.  

So that leaves us with 13.Bxd5 c6! regaining the piece.  Check out the big name showdown from 1992 (I was there, but I didn’t remember the game):

Black obtained typical Marshall compensation; Anand felt he had no winning chances against Black’s bishop pair.

So the “standard” 12.d4 may be best after all.  As you pointed out, Black suffers from not being able to chase the rook with 12…Bd6.  Your suggestion of 12…Nf6 has not been played much, but might be worth a try.  13.Bg5 Bd6 14.Re1 c5! is an improvement over 14…Re8, and I don’t see any major advantage for White.

12…Bf6 has been tested several times, including this nervy repeat from Nigel Short.

Anand refuted Short’s 13…c5?! with 15.Na3!  Instead of 19.Qf3?, which essentially equalized the game, Anand missed 19.Bxb7! Rxd1+ 20.Rxd1 and White should win on the strength of the passed c-pawn.  13…Re8 is a better way to seek compensation.

Black has one more option in 12…Qd7.  After several games in the nineties, White seems to come out on top after 13.Nd2 Nf4 14.Ne4.  For instance, 14…Nxg2 15.Kxg2 Bf6 (15…Bd6? 16.Bd5) 16.Qf3 Black doesn’t seem to have enough for the sacrificed material.  14…Bd6 15.Nxd6 cxd6 16.Rg5 Ng6 17.Rg3 is better for White as well.

In the last three years, I found six examples of 11…Bb7, but look at who played two of them.

Same place same day…was Nigel sharing knowledge with his good friend Gata (sarcasm alert)?  Anyway, the results indicate that Black is holding his own in the 12.Qf3 line.

Let me try to sum up.  The under-the-radar 11…Bb7 is actually quite interesting and provides a rich variety of possibilities.  Most of the testing seemed to occur in the early nineties, and for whatever reason, the variation didn’t seem to develop a wider following, both among GMs and with the masses.  However, after a point that which is old in chess becomes new again.  I think 11…Bb7 is sound enough to be a surprise weapon for players on many different levels.  Potential opponents who have not read this column can be in for a rude awakening.

Send your inquiries on anything from openings and endings to the inner workings of the chess world to