Maroczy Bind theory question

1. Nf3 Nf6 c5 cxd4 4. Nxd4 Nc6 5.Nc3 g6 Bg77 .Be3 0-0 8. Be2 d6 Bd7 10. Qd2 Nxd4 11. Bxd4 Bc6 12. f3 a5 13. b3 Nd7 14. Be3 Nc5 15.Rab1 Qb6 

What is the idea behind Qb6 in this queenside attack line? Especially after the standard 16.Rfc1 Rfc8 17.Rc2 black plays Qc8 back anyway? Why do ~700 master games go 15. ... Qb6 instead of Qc7, Qd7 if the only idea is to transfer the rook anyway?

Thanks

The reason is very computerish. The idea of Qb6 is to play Qb4 and push through a4. So then the question is, why does Rfc1 prevent this? If you play a null move instead after Qb4, the problem is that the c3 knight is effectively pinned, because if it moves, after the queen trade on d2, Black plans to bring a rook to a2 and skewer the bishops on the second rank. So one point of Rfc1 is that if Qb4, there's Bf1, a convenient square for the bishop where it hides from this skewer. Now if Black proceeds with a4, there's Nd5 threatening Nxe7, so Black has to deal with that and then White has time for b4. I hope you followed all this!

The question then is "OK fine, but Qb4 never happens in practice, so what was the point?". I think the point is getting the rook to c8, on the other side of the queen. The action is happening on the Q-side, so Black wants the rooks there.

Yes thank you for the explanation! And Bf1 also allows for the knight to retreat to e2 after b4 axb4 axb4 Na4 instead of having to trade.

idfk but i hope you are having a good time <3

GM's Bent Larsen and Margeir Petursson employed this idea some 40 years ago, when chess computers did not exist.

The idea is quite human, and twofold: 1.prepare Black to fight white's queenside expansion by placing both rooks there, and second to pile up on the long a1-h8 diagonal with ...Qd8-f8, Be5 and Qg7, or alternatively ...h5, Kh7, Qd8-h8 (Larsen's refinement).