Someone should send this thread to Aagaard...he blogged recently about how tricky it is to prepare for opponents using the GM Repertoires! 

Marin has a chapter about what he calls "The Inferior Queen's Indian". One line he analyses is 1. c4 c5 2. g3 Nf6 3. Bg2 e6 4. Nf3 b6 5. d4 Bb7 6. d5 exd5 7. Nh4 g6 8. Nc3 Bg7 9. O-O O-O 10. Bg5 (This is in his vol. 3, not vol. 2). – Hope this explains how he arrives in this QI position.

BPaulsen wrote on 12/10/10 at 21:36:57:


Marin covers this line by transposition in his English vol 2.  He doesn't consider 9 ...d6, however his mainline is 9...0-0 10 Bg5 Qc8 11 cxd5 d6 and now 12 Qd2 followed by Bh6, with some Qf4 ideas.  If white exchanges the bishop he has a nice modern benoni - the Nh4 is slightly misplaced but black has also misplaced the Bb7 and lost a tempo with the Queen - so he thinks it's favourable.  If the bishops are exchanged white can play a3/b4/Rac1 happily.

I can't help but wonder if 9...d6 10 Bg5 Qd7 11 cxd5 0-0 12 Qd2 is similar - the only difference is the Qd7 vs Qc8.  In favour of Qd7, b5 might come faster (but black is underdeveloped currently), but otherwise Nbd7 isn't possible without losing another tempo.  So white might even be able to force through e5 faster, since the Nh4 supports this as d5 has extra defence.  I think there are too many possibilities to look at concrete analysis, but I wouldn't be so keen to claim black has a superior benoni just yet.

One thing is 12 Qd2 Re8 13 Bh6 Bh8 14 e4 Na6 15 Nf5.  Marin covers this with Qc8 and claims a clear advantage to white.  This is a better version since d6 is covered, but it seems the Nh4 is back in play already.

I was interested to note that 9 0-0 0-0 10 Bf4 was the course of Alekhine-Capablanca, New York 1927, which Capa won easily playing a good benoni.

As an aside, I was looking through Grivas survey on these lines in CBM and was suprised to see that numerous sidelines after 4 g3 are not that bad.  Tiviakov has been hacking away with 4/5...c6 and d5 transposing to the Catalan for instance.

9...d6 is not a new move, having been played in three correspondence games in 2004-05. But meeting 10.Bg5 with 10...Qd7 is new, and quite clever.

Maybe White can play a 21st Century Gambit a tempo down with 7.cd5 Bd5 (7...Nd5 8.Nh4) 8.Nc3, but White doesn't get in e4 as quickly as in the 4...Ba6 5.Qc2 line so I suspect this isn't a serious try for an advantage.

I also looked at 9.Bf4 0-0 10.Bd6 Re8 11.cd5 which the Informant likes for White, but 11...Nh5! 12.0-0 Qf6 13.Bb8 Rab8 is better for Black, not White.

After 10.cd5 0-0, the most recent game of relevance, Chadaev-Lubczynski, continued 11.Nf3 Na6 12.Nd2 Nc7 13.Qb3 b5 14.e4, but Black is doing pretty well here.