@Mnb
Quote:12.d3 Bd6 13.Re1 Qh4 14.g3 Qh3 15.Re4 Qf5 16.Nd2 Qg6 17.Nf1 f5 18.Rd4,f4 looks attractive
This was actually played in the game Timman-Huebner, 1985
19.Rxd5+, cxd5
20.Bxd5+,Be6
21.Bxa8 (21.Be4,Qf7!? 22.Bxa8,fxg3 23.Bg2,bxh2+ 24.Kh1,Qxf2!
25.Ne3,Qg3!
And black had more than enough compensation in Balabaev-Schwetlick ICCF corr.,2000)
21......., Rxa8
22.Qf3 22.a4,Bg4 23.Qb3+,Be6= (Nunn)
22.............Rf8
23.Qe4, Bf5
24.Qd5+, Kh8
25.a4
25..........b4! (25....Bxd3 26.axb,Be4 27.Qd4,fxg 28.Nxg3
gave white slightly better chances in Morgado-Brilla-Banfalvi, corr. 1992)
26.cxb4
26.........., Bxd3! (Is I think an improvement)
26....fxg3 27.hxg3,Bxb4 (Timman-Huebner) and now
I think white is for choice after 28.Bd2!
27.Bd2 (what else)
27......., Be4
28.Qd4, Bc7! And in view of the threat Bb6 white hardly has anything better than repeat moves with 29.Qc5,Bd6 30.Qd4,Bc7.
17.Re1 iso Nf1 is a major alternative.
Consensus is that White is probably somewhat better, but nothing major.
17.a4,Nf6 is also OK for Black although he has to know some theory to avoid some major pitfalls.
18.Re1,Bg4! 19.f3,Bh3 is the critical line
(see for instance Goloshcapov-Azarov, Cappella la Grande, 2006)